This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
@Okkn: I know only that it isn’t specific to Wikidata; an admin on enwikisource has reported a similar problem. I don’t know whether this is related to the recent Phabricator attack or to some unreported technical change or mishap. Mahir256 (talk) 04:44, 2 July 2018 (UTC)
This page has a wrong image in it. The image is not of a Riley 1½-litre production car but of a "special". I have tried to correct it but my edit does not work, it seems to have created a new record. How do I fix it? Eddaido (talk) 03:12, 6 July 2018 (UTC)
@Eddaido: I'm not sure what you mean; you modified the image property and so there is still only one image. I don't think you need to do anything else. Mahir256 (talk) 03:17, 6 July 2018 (UTC)
Thank you. This page still shows the wrong image. I seem to have created an extra record instead of replacing the image. That's to say I've made a mess - or, is there a delay before the change is actioned? Thanks, Eddaido (talk) 03:21, 6 July 2018 (UTC)
@Eddaido: On the Commons category I see the image that you changed it to. Have you tried clearing your browser cache? That could well be the culprit. Mahir256 (talk) 03:28, 6 July 2018 (UTC)
Coordinates "Imported from Russian Wikipedia" = "Coordinates are likely to be incorrect". Is there any way to put a moratorium on importing data from the Russian Wikipedia? Abductive (talk) 23:11, 9 June 2018 (UTC)
It would be possible to have a bot that removes dta imported from Russian Wikipedia periotically. Before doing that it would however make sense to speak with the people who imported the relevant data. ChristianKl ❪✉❫ 11:20, 11 June 2018 (UTC)
The problem is that ru.wiki users that are doing so think that P143-sourced statement is better than no statement at all and in order to use WD data in their infoboxes... they imported their data here. Wostr (talk) 11:54, 11 June 2018 (UTC)
@Abductive: Sorry, but earlier "imported from Russian Wikipedia" was a sign of quality (your opinion agains mine, uh?). But better consider mass import from Cebuan Wikipedia. If you have some arguments (statistics, research) for imho, you should give it, otherwise your post is abusive. --Infovarius (talk) 15:21, 14 June 2018 (UTC)
Just check my contributions for when I changed coordinates. You'll see that coordinates imported from the Russian Wikipedia are frequently incorrect. Or do your own study, using the random item feature. Actually checking the coordinates against internet maps reveals glaring errors. Abductive (talk) 16:27, 14 June 2018 (UTC)
@Abductive: Well, unless you've made contribution under different accounts and/or anonymously, I can see your corrections of coordinate location (P625) in:
I also looked into a few random items with coordinates taken from ru-wiki: Q522462, Q3214976, Q17388. Frankly speaking I don't see how did you draw conclusion "coordinates imported for ru-wiki are likely to be incorrect". Don't get me wrong, you are doing important job of coordinate verification and your input is greatly appreciated. But before deleting >200K of statements, I'd like to see 10^2-10^3 "glaring errors". --Ghuron (talk) 12:13, 21 June 2018 (UTC)
Abductive, until you can demonstrate with evidence that in respect of coordinates the Russian wikipedia is any more error-ridden than any other langauge wikipedia, I'd greately appreciate it if you could knock the whole trumpian 'error-ridden Russian Wikipedia' thing on the head. thanks. --Tagishsimon (talk) 17:57, 24 June 2018 (UTC)
You're being asked for evidence, which you seem unable to supply. Making stupid ad hominem attacks serves only to undermine your argument and your reputation. --Tagishsimon (talk) 16:10, 26 June 2018 (UTC)
It seems that you don't understand my evidence. I corrected a bunch of coordinates imported from the Russian Wikipedia. But what you can't see is all the times I didn't have to correct coordinates imported from other Wikipedias, because I didn't have to correct them. Get it? Abductive (talk) 03:02, 27 June 2018 (UTC)
The only evidence presented so far is from Ghuron, which does not support your assertion. You have not presented any evidence, just assertions, drama & personal attacks. Speaking as someone who has fixed one or two orders of magnitude more coordinates than you, I cordially decline your explanation, whether or not you italicise it. --Tagishsimon (talk) 10:18, 30 June 2018 (UTC)
Coordinates from zhwiki seem also problemic, as @Artoria2e5: pointed locally, they have GCJ02 (Q29043602) and BD09 (Q29043632) issues which can make shifted coordinates, and I'm still waiting for a resolution on Wikidata to handle all (maybe include left both) non-WGS84 coordinates. I'm not sure if Russian authorities do have such things or not. --Liuxinyu970226 (talk) 04:29, 27 June 2018 (UTC)
In line with Liu's hypothesis, it might be interesting to know how Abductive corrected it, or in other words whether the coordinates are "wrong" in the same technical way. Knowing that can help justify some further actions. BTW, the Chinese stuff is tracked on phab:T162331 I think. --Artoria2e5 (talk) 02:46, 2 July 2018 (UTC)
How to link an item to the standard that describes it?
Over at Wikidata talk:Lexicographical data I made a suggestion to allow the use of lexeme entities as labels for items. As I explain there, this is motivated by the fact that sometimes you want a specific grammatical form of the word that is being used as the label, one that is potentially different from the form used as the label (for example plural versus singular).
Using rain (Q7925) as an example, here is an example of how I envision this could be done in practice:
At rain (Q7925), where you would normally edit the label, you would have the ability to connect the label to a lexeme (e.g. by clicking on a small icon just right of the field where you can enter a label). If you are editing the English label, you can connect this label to the English lexeme rain (L3325). By default, the lemma will be used as the label, but you can also choose any of the forms of the lexeme.
Backwards compatibility could be maintained by the fact that the label still has a string representation (like specified above: either the lemma or a selected form), such that a Lua command like getLabel() can still return a string, and that the lexeme entity would be fetched by a command like getLexemeLabel().
Note that using lexeme entities as labels would not be mandatory; using simple strings would still be possible, and there would be no need for a mass-conversion of existing labels. --Njardarlogar (talk) 11:47, 30 June 2018 (UTC)
I think this is a nice idea, but probably a bit premature. There is a planned property to link Wikidata items to senses of lexemes, but the sense functionality isn't there yet. Though I think what you propose is a distinct approach, it's definitely related and it might be useful to think about what could be automatically done with this new property before putting a lot of effort into this particular approach... ArthurPSmith (talk) 01:18, 2 July 2018 (UTC)
college fraternities: "for men"?
fraternity (Q18811582) gives the English-language description "collegiate social organization for men"; descriptions in several other languages are equivalent. However, quite a few college fraternities are co-ed (for example, Wesleyan University now allows only co-ed fraternities). Do we need a separate item for co-ed fraternities, or should these be described more broadly? - Jmabel (talk) 01:43, 26 June 2018 (UTC)
The Facto Post monthly mass message delivery on English Wikipedia, which I write for ContentMine as their Wikimedian in Residence, has now been around for a year. The June issue is out, with the usual mix of editorial content and links, generally intended to inform Wikimedians about Wikidata-related developments in a broader context.
Hello. At Cyprus (Q229) we speak Greek (Q9129), but not only Modern Greek (Q36510). We have are own dialect: Cypriot Greek (Q245899). We have some different words for some entities. For example, how can I add the word we use for Mespilus germanica (Q146186)? I can add it in "Also known as" for Greek but with this way I can't show that it is a name that came from Cypriot dialect (they word is only use in Cyprus, the other Greek spoken people don't know it). Ι can't use taxon common name (P1843) because you need to add a language, not a dialect.
The list includes "CY" as subtag. I think it's one that can apply. Tags are available for the like of "en-gb" or "en-ca". No need to go into linguistic (or political) debates about its characteristics. --- Jura13:02, 2 July 2018 (UTC)
Celtic Knot Conference, dedicated to Wikimedia projects and minority languages, July 5-6, in Aberystwyth (Wales). A lot of Wikidata workshops will take place during this event.
The Commons template Wikidata infobox has now passed 1.3 million uses on Commons categories, with plenty more still to come. A new help page gives advice on how to add, extend or improve uses of it, and how to fix data problems.
Wikibase’s maxlag now takes dispatch lag in account
This change impacts people running bots and semi-automated tools to edit Wikidata.
Hello all,
Based on the previous discussions that happened around the limitation set up to fix the important dispatch lag on clients, we came with a new solution to try.
The database behind Wikidata is replicated to several other database servers. At each edit, the changes are replicated to these other servers. There is always a short lag, which is usually less than a second. If this lag is too high, the other databases can’t synchronize correctly, which can cause problems for reading and editing Wikidata, or reusing data on other projects.
If the lag is too high on too many servers, the master database stops accepting new edits. When the lag is close to the limit, the system is prioritizing “humans” edits and ignore the edits from bots, sending back an error. This limit is set up by the maxlag option in the API.
People writing bots can set up a number as maxlag for their bot. The default value is 5. This number is used to evaluate two things: the replication lag between master database and replicas, and the size of the job queue.
On Tuesday, July 3rd, maxlag will also evaluate the dispatch lag between Wikidata and clients (eg Wikipedias).
The dispatch lag is the latency between an edit on Wikidata and the moment when it’s shown on clients. Its median value is around 2 minutes.
If you’re running a bot and using a standard configuration (maxlag=5), when the median of dispatch lag is more than 300 seconds, your bot edits won’t be saved and will return an error.
If this change is impacting your work too much, please let us know by letting a comment in this ticket. This is also where you can ask any question. You can also change your configuration in order to increase the maxlag limit.
If we do not have such property than I can propose "horizontal thickness", but I want to make sure I am not missing something. --Jarekt (talk) 12:54, 29 June 2018 (UTC)
Multichill, I was thinking more about this. When we discuss "thickness" of 3D objects it is usually not in context of the horizontal depth. For example thickness of bronze statues, we usually refer to the wall thickness, see [6] or [7]. Same with pottery, [8], Alexander Calder (Q151580) sculptures [9], swords, etc. There are cases where people refer to depth / thickness of an 3D object interchengbly, but it is rare. --Jarekt (talk) 13:07, 3 July 2018 (UTC)
Judea and Samaria Area
Judea and Samaria Area (Q513200): I'm pretty surprised to find nothing here indicating in any way that there is any controversy about including this area as part of Israel. Unless I'm mistaken, not even the U.S., probably Israel's closest ally, recognizes this as Israeli territory. I'd be inclined to mark it with disputed territory (Q15239622), but that says "territory, which status is controversial between two or more existing states" (not very good English) which I take to mean that two or more existing states (emphasis mine) each claim the territory. Since the Palestinian Authority is not a state, that wouldn't exactly apply. Thoughts? Possibly broaden the definition of disputed territory (Q15239622)? - Jmabel (talk) 17:28, 1 July 2018 (UTC)
statement disputed by (P1310) is the property that get's used to mark controversy of claims like that and it would make sense to add it in this case (of course it makes sense to search for authoritative sources for both sides). ChristianKl ❪✉❫ 17:55, 1 July 2018 (UTC)
How do we mark adoptive vs. biological parents. We have stepparent as a field. I tried to find an entry of a person that was adopted to see how it was handled but could not find an example with a parent linked. --RAN (talk) 19:17, 1 July 2018 (UTC)
Interesting question. I vaguely remember a radio programme discussing parenthood questions and which, iirc, posited three classes: biological (i.e. suppliers of gametes), gestational (e.g. surrogate mothers) and legal (inc. adoptors) ... see, e.g. [10]. Clearly many parents are two or more of these. If we're thinking about how we model parents, we might want to factor this sort of diversity into the model. --Tagishsimon (talk) 20:26, 1 July 2018 (UTC)
Now even a third biological parent with a mitochondrial donor. I imagine they would remain anonymous in the records. --RAN (talk) 21:01, 1 July 2018 (UTC)
Help us improving the fight against vandalism on Wikidata
Hello all,
As you may know, we use a tool called ORES is analyzing edits to detect vandalism. It is providing a score per edit. You can see the result in the Recent Changes, by adding the filters "likely have problems" and "very likely have problems" in the "contributions quality predictions" section.
In order to keep improving this tool, we would need feedback about:
Edits that are not detected as vandalism
False positive (edits that are highlighted as vandalism when they’re not)
Any kind of patterns that you notice (for example: "edits on Commons categories have a higher chance to be vandalism")
I was looking at the wheat page on en.wikipedia.org and could not find a link to the italian version. I suppose adding a Local link would not be a great idea. I'm sorry to go meta, but how may I learn more about wikidata? I'd like to understand if the "only one wikidata category for each wikipedia article" is a technical limit or a policy :) Carmelobrianza (talk) 13:48, 2 July 2018 (UTC)
I'd like to understand the parameters under which a scientific research database could be imported into Wikidata with maintenance and updating being crowdsourced with discussions in Wikipedia and Wikiversity. For specificity, consider:
Their Appendix A list 648 different terrorist organizations active between 1968 and 2006. They state that those cases were extracted from the RAND-MIPT database, which began in 1968 (Jones and Libicki, p. # 4, 30/251). I have multiple questions about this:
Might it be worth asking both RAND and the National Consortium for the Study of Terrorism and Responses to Terrorism (START) if they might both contribute their databases (RAND-MIPT for RAND, and the Global Terrorism Database, GTD, for START) to Wikidata, where the maintenance could be crowdsourced with, e.g., discussions of individual data items documented in appropriate articles in Wikipedia and Wikiversity and their associated talk pages?
If we invite RAND and START to contribute their data to Wikidata, what could we offer them in terms of control over the data and recognition for their work?
Per Wikidata:Copyright, "All structured data from the main and property namespace is available under the Creative Commons CC0 License; text in the other namespaces is available under the Creative Commons Attribution-ShareAlike License; additional terms may apply." I have no idea what this means, and I left a question to that effect on Wikidata talk:Copyright.
If you all think this could be a good idea, I could contact RAND and START to see if they might be interested in collaborating in this way. However, before I did that, I'd want to understand this copyright question.
For example, could they have control over certain data fields associated with a purported terrorist incident? Others could add alternative codings as alternative variables, but could RAND and / or START still control the official entry? They could decide to change it if others provided persuasive documentation or analysis? Otherwise, users could decide whether to use the official or some alternative coding for specific incidents?
Similarly, others could add incidents not listed in RAND-MIPT or GTD that could be used by them and still others in alternative analyses.
This could make it easy for other researchers to evaluate the robustness of the conclusions of published studies -- and make it easier for people to update any analysis any time they wanted for any particular purpose.
Similarly, what would you suggest I worry about if I wanted to import the data in Appendix A of Jones and Libicki into Wikidata?
At least in the US, you can copyright expression but not facts. In this context, that suggests to me that we could create a table of terrorist incidents as a direct copy of their Appendix A and have it carry CC0, because this gives us flexibility to add alternative cases and codings with links to articles in Wikipedia and Wikiversity. That flexibility, especially if exercised, would seem to create an "expression" that's different from their Appendix A. However, I'm not an attorney, and I don't know these details -- and I certainly don't know how the Wikidata community has interpreted these questions.
Chenoweth, START and RAND all need some assurance that they would get credit with other people use their data, because their funding depends on the use that others make of their data.
Sounds like this would be useful data if we could get it. Some of this could be linked up with various existing items. There is not currently any way for an organization to have control over a set of items, and given the license, there's no way of enforcing an attribution requirement either. We could, of course, add them as a source, but there would be no requirement for any reusers to mention the source. Regarding legal issues, m:Wikilegal/Database Rights may be useful. --Yair rand (talk) 06:47, 3 July 2018 (UTC)
Not sure if the Wikidata community is really up to moderating edits of this type of data. --- Jura08:28, 3 July 2018 (UTC)
Thanks. Permit me to summarize:
Advantages:
- crowdsourcing,
- potentially easier distribution with higher profile (including via packages for R (programming language) including WikidataR, WikidataQueryServiceR, webchem, wikitaxa, and WikipediR
- conflict resolution via Wikimedia rules of writing from a neutral point of view citing credible sources and documenting questions and concerns on associated "talk / discussion" pages,
- robustness by supporting alternative definitions contributed by anyone in the world -- and available to anyone else to modify, etc.
Concerns: Giving up legal rights (see m:Wikilegal/Database Rights) including copyrights and inability to enforce attribution.
I wonder how can I found out how many labels in Wikidata are for each language, out of the total amount of 50 millions entries.
For example, for Catalán language ("ca") I tried with
SELECT ?lang (COUNT(DISTINCT ?item) AS ?count) WHERE {
?item schema:inLanguage "ca" .
} GROUP BY ?lang
ORDER BY DESC (?count)
and got a result of 703351, but I think it's not correct because I downloaded the Wikidata dump, and I already extracted more than two millions of labels in Catalán (and the extraction is still running)
@User:Tagishsimon, Tagishsimon: I see... There's no way to do these "heavier" queries, right? :/ – The preceding unsigned comment was added by [[User:|?]] ([[User talk:|talk]] • contribs).
Is there a handy tool for cloning items? I run into several items to painting series which need to be split into individual paintings, like Q4451151. The easiest way would be to clone existing item and make changes to clones. --Jarekt (talk) 12:19, 3 July 2018 (UTC)
Is there any standard procedure for disputing a claim in Wikidata?
Is there any standard procedure for disputing a claim in Wikidata? Starting from the Community Portal, I don't find anything. The immediate issue I have is stated at Talk:Q216477#Citizenship? (questioning the claim that Chief Seattle (Q216477) was a United States citizen), and while it's only been a few hours, it struck me that bringing up something like that on a talk page is liable to lead to it just languishing forever. - Jmabel (talk) 00:59, 2 July 2018 (UTC)
For unsourced claims, I'd recommend simply removing it, then adding a note on the talk page explaining why you doubt the claim, pinging the original author of the disputed statement. (Ping @Arbnos:, who added the citizenship claim on the Russian Wikipedia.) --Yair rand (talk) 01:23, 2 July 2018 (UTC)
Yes, if it were unsourced I'd do that; as I say here, the issue is a stated source that I cannot read. - Jmabel (talk) 04:19, 2 July 2018 (UTC)
But pinging the person who added the claim is a good idea. I haven't quite gotten used to thinking of Wikidata items as having contributors in quite the same sense as Wikipedia articles. - Jmabel (talk) 04:21, 2 July 2018 (UTC)
Such statements imported from Wikipedia are usually added en masse by bot or script, by using data from infoboxes or categories. So the editor in question might not have any knowledge about this specific item. And unfortunately, it's highly likely that sooner or later the information will just be re-added by another bot or script, as long as that information is still present in the Wikipedia article. The only reliable way would be to remove the information from the article in question.
Unfortunately the whole topic of citizenship, country of origin, etc. is quite a mess across the board when it comes to people in the past. Too many Wikipedias just use current countries in their infoboxes/categories for historical persons that died before the current state existed. --Kam Solusar (talk) 08:11, 2 July 2018 (UTC)
It's not just people. The whole issue of "conceptual" or "notional" countries and nationalities is a big problem. If a source says a fabric was made in Italy in the 15th century, we can't correctly model it, because there was no country called Italy at that time - yet 15th century people and modern people equally understood what was meant by Italy as a location (even though the borders are probably fuzzy and variable). - PKM (talk) 18:31, 2 July 2018 (UTC)
Part of this is an effort to project backward a notion of national states that is actually relatively recent, with a history of certainly less than 500 years, and which was not universal even in the Napoleonic era (arguably, not until the era of decolonization). The concept of "Germany" or "Italy" in the mid-19th Century was more like the concept of "Southeast Asia" or "Central Europe" today than like a state. - Jmabel (talk) 00:45, 3 July 2018 (UTC)
Thanks, that's very useful, and entirely new to me. Is that possibility at all prominently documented somewhere? I've been active on here about 2 months and never encountered it. - Jmabel (talk) 16:38, 2 July 2018 (UTC)
Someone seems to have added a type constraint without checking the current values. Besides, it seems the constraint reports for that property are currently broken. BTW, deprecated rank is explained on Help:ranking --- Jura08:34, 3 July 2018 (UTC)
I don't think they're accurate. That's more of a class-instance relationship, and the statements are also already implied by location data. So I'd say no, I don't think it's useful. --Yair rand (talk) 18:17, 2 July 2018 (UTC)
Structured Data on Commons - community consultation on basic properties for media files
Hello everyone! This month (July), we are hosting a quite crucial community consultation on Wikimedia Commons: we are listing the Wikidata properties that media files on Commons will need (including ones that might not exist yet, and might need to be created for this purpose). The consultation runs at least till the end of July, maybe longer.
it.wiki pops up a "Proposed EU copyright directive will eat your children" [11] full-screen message when you visit it. en.wiki and this place not so much. Which is a shame, since the proposed EU copyright directive will indeed do that thing. --Tagishsimon (talk) 08:12, 3 July 2018 (UTC)
I'm happy they are trying to settle our disagreements with the Chinese and Turkish governments. We probably don't need our neutrality that much. Thierry Caro (talk) 11:51, 3 July 2018 (UTC)
I also think Wikidata should go down, since the effects on the law on data mining procedures for UE citizens. --93.41.235.1512:34, 3 July 2018 (UTC)
I also support the move to get rid of our neutrality. As I support the directive, I'll get my own full-screen message displayed too! Yeaah! Thierry Caro (talk) 12:37, 3 July 2018 (UTC)
Support For the shutdown, at least one day. It is pretty serious. I called several MEPs' offices today through this website, and I encourage everyone to do the same, so at least they know that there are people who care about this. It is a free call since you are called to your number and then you are put through to the MEP office. If that website doesn't work for you, you can also try this one by the Mozilla foundation, where you even have a script of what to say.--Micru (talk) 14:48, 3 July 2018 (UTC)
Yeah! We've spent half a decade fighting vandalism on individual items one at a time, why not suddenly deface our entire website all of a sudden? Thierry Caro (talk) 15:10, 3 July 2018 (UTC)
Difficult to understand why you would support this, Thierry Caro. It seems as shitty a piece of legislation as I can think of. Would not have happened in Neelie Kroes's day. --Tagishsimon (talk) 17:14, 3 July 2018 (UTC)
Support inaction in the face of the copyright fee industry abuse, will not maintain an "apolitical" movement, but rather a pusillanimous one. Slowking4 (talk) 03:28, 4 July 2018 (UTC)
Update: English Wikipedia is now running a banner for not-logged in users, with an alternative text for EU members vs. outside EU. I didn't change the topic to "Wikidata shutdown" but I think it is an option. I do think at least a similar banner would be appropriate. Jane023 (talk) 16:16, 3 July 2018 (UTC)
I think it would probably be worthwhile to establish a policy of never getting involved with politics, period. --Yair rand (talk) 03:51, 4 July 2018 (UTC)
Yes that's a fine opinion when the political action does not affect the future of this project. However, since the project Wikidata was created and continues to run thanks to lots of effort by people located in Germany who must conform to EU copyright law, then I guess this would be a good time to get off your high horse and vote. Jane023 (talk) 08:57, 4 July 2018 (UTC)
Abortion and the right to freely carry guns also immediately affect our projects by curbing our future numbers of readers and contributors. They are stolen from us. Let's have a banner as quick as possible! Thierry Caro (talk) 11:02, 4 July 2018 (UTC)
straw men and slippery slope are not much of an argument, in the face of an immediate threat. open knowledge is inherently political. look at the array of interests aligned against it - tyrants and rent seekers. a little expression of the wikimedia movement to political authorities is a minimal effort. Slowking4 (talk) 14:04, 4 July 2018 (UTC)
Úpdate: Current projects doing something: Closed their wikipedia : it, es, lv, et. Running a banner : en, pl. Jane023 (talk) 09:01, 4 July 2018 (UTC)
It is a disappointing to see that while mostly everyone was supporting a banner, nothing has been done. Good that the directive has been rejected anyway!--Micru (talk) 11:14, 5 July 2018 (UTC)
Especially the last one is an important one, as the location is an important key to find duplicates.
Given the huge numbers, this is impossible to fix manually, so is there any bot owner who can help with it? Partially, the instance and country can be filled with petscan, but the categories on ceb sometimes merge several related types (e.g. river and wadi), and there are categories like watersheds which make items from different countries show up under the main category of one country. Thus these probably need to be taken from the infoboxes. Importing directly from geonames isn't possible due to the license, but as the GNS database it free, we could use that one once the matching geonames-id to GNS-ID is done. Ahoerstemeier (talk) 21:35, 4 July 2018 (UTC)
That's great, that'll help to be able to clean up the mess created by the uncontrolled import from ceb. BTW - there are also thousands of items where the geonames ID wasn't imported from the ceb-infobox yet, so these would escape your bot. Ahoerstemeier (talk) 22:00, 4 July 2018 (UTC)
change in "imported from" (P143)
Project Chat folks should probably be aware of a recent change in imported from Wikimedia project (P143) - see this discussion: Property talk:P143#Making the usage of this property clearer. I was a little surprised to see the recommendation to only use it with bots; I have often entered it manually particularly to provide the "official website" of a university or other organization, as I don't know what other reference one can use in such a case (the website itself?) if the institution is not in some major database, and use of imported from Wikimedia project (P143) at least indicates agreement between the wikidata value and the value in one of the wikipedias. But I do agree this property shouldn't be used for imports from VIAF etc! ArthurPSmith (talk) 13:48, 5 July 2018 (UTC)
Yes, this sounds like a misunderstanding. You may of course still indicate data provenance with imported from Wikimedia project (P143) if you import manually from a Wikimedia project. The change is about new restrictions that should prevent anything else than Wikimedia projects as import sources. Help:Sources has enough reference models for all kinds of external (non-Wikimedia) sources. --MisterSynergy (talk) 19:12, 5 July 2018 (UTC)
It seems to me the website itself is an adequate source, provided it has a TLS certificate issued by a reputable certification authority. It isn't perfect, but it's better than a wikimedia website as a reference. Jc3s5h (talk) 21:10, 5 July 2018 (UTC)
The description and documentation is set a bit more strict on purpose to have people, who are in doubt, not accidentally use it. The backlog is already quite huge and I hope this will keep it from growing even more. We might want to tweak it a bit in the future when we got the backlog a bit more under control. Multichill (talk) 13:22, 6 July 2018 (UTC)
I would like to localize a web project of me with the help of Wikidata. It shows country profiles, it's like CIA World Fact Book with many charts. Because of this my content consists mostly of single words, which could be translated easily.
Is there any JavaScript library in which I could use wikidata items to localize my strings? I would like to use something like $Q38$en (pseudo code) giving back "Italy" for example.
As far as I can tell, that page only provides graphs for 3244 names. I would only add the URL to those names, rather than to every given name. - Nikki (talk) 12:11, 6 July 2018 (UTC)
Status for a buildings/monuments
Which property should be used to indicate state of conservation / state of use of a building, monument, etc. I'm not talking about its heritage protection (heritage designation (P1435)), but to its present state of conservation or use. Ex.: ruinous, demolished, reformed, correct, abandoned, .... Thanks,Amadalvarez (talk) 07:45, 30 June 2018 (UTC)
Currently such property doesn't exist. Some would recommend to use significant event (P793), but I find this very complicated. I would support a proposal for a "state of conservation" property. — Ayack (talk) 12:06, 30 June 2018 (UTC)
I just though why User:Ezzex called this "religious propaganda" (of me, atheist!). I thought that this was trivial biblical fact about this biblical character. But now with P1310 I see that this can be rejected by some other avramistic religions (though there is another item about Muslim Jesus). --Infovarius (talk) 07:39, 6 July 2018 (UTC)
My opinion is that this is religious propaganda and that we (as a encyclopedia) should not participate in such things. This "Chist"- and "father"-definition do not fit in a infobox, but should rather be placed in the main article.--Ezzex (talk) 14:12, 6 July 2018 (UTC)
Should they merge or whitelist? If you need Hungarian translation, feel free to ask. To be honest, Shenzhen Open is totally weird for me and I have no idea what to do. Bencemac (talk) 13:36, 23 June 2018 (UTC)
The first one shouldn't be merged. It's frequent that we have items for a tournament for men, another for women, and a third for items (articles) combining both. Visit WikiProject Tennis. --- Jura19:47, 3 July 2018 (UTC)
Looks like they are different: different dates, different authors. The one is probably a comment on or correction of the other (there appears to be a set of four publications with this title). - Brya (talk) 07:25, 7 July 2018 (UTC)
What needs to be done for wikia.com (Fandom) to be able to use Wikidata? Familypedia can link to English Wikipedia and use images from Wikimedia Commons, but does not appear to be able to link here. Any ideas what needs to be activated? --RAN (talk) 22:26, 6 July 2018 (UTC)
Re-scheduled deployment of the New Filters on Watchlist
There was a recent announcement about the plans to graduate the New Filters for Edit Review out of beta for this Wiki. The deployment was stalled to fix the performance issue related to the change. The performance of the new interface has been improved significantly as an outcome of the work by the developers [13]. So, the deployment has been re-scheduled. The deployment is scheduled for this wiki on July 16th 2018.
Please let us know of any other issues or special incompatibility that you may face so that we could make sure they are solved before the feature gets deployed. -- Kaartic (talk) 20:28, 7 July 2018 (UTC)
Could we address this with a new property “subcategory of” (or “intersectional category of”) and link to both “parents”? (Thinking out loud - I haven’t spent time trying to figure how this might work). - PKM (talk) 06:31, 5 July 2018 (UTC)
I would think we would want two properties, so that we could express the concept that it partitions musical groups by genre. - Jmabel (talk) 07:24, 5 July 2018 (UTC)
Proposition: A genre is a class of artwork. It’s an instance of the metaclass « genre », or « musical genre » (a subclass of genre) for example. Then we just use disjoint union of (P2738) who is designed for this, see the talk page of the property for help. author TomT0m / talk page07:29, 5 July 2018 (UTC)
One quite often sees entries analogous to these in taxonomies, ie as entries that would correspond to main items, in P31/P279 trees of main items, not category items. So far I think the people like User:PKM who have really been putting the work in on taxonomies have mostly been ignoring such entries, and leaving them unlinked. But I have been wondering whether that's really the best policy, or whether it might be really quite useful to have these as items, to be able to group and distinguish items that were instances or subclasses of them.
But I think my first suggestion would be: don't try and describe this as a category. Create a corresponding main item, eg "Music groups by genre", and describe that.
Though as a data modelling policy, I suspect that on Wikidata we probably don't want to make Group <X> instance of (P31) "Grunge group" or "Folk rock group" or whatever, that might be subclasses of "Music groups by genre" -- instead we probably prefer to leapfrog all of that and have Group <X> instance of (P31) "music group", and then Group <X> genre (P136) whatever. Jheald (talk) 09:23, 5 July 2018 (UTC)
"Music groups by genre", obviously first it's subclass of (P279)musical group (Q215380) definitely not. First, it’s a Wikimedia category like modelling items, whose purpose is to categorize categories. It should not be thought as a class. The need is to classify music groups, and then to classify the music group classes. If « Nirvana » is a(n instance of) « Rock band », then « Rock band » is not a subclass of « music group by genre ». Proof, by absurd: Any instance of a class is an instance of its superclasses, by definition of « subclass of », see description. Then if "« Nirvana » is a(n instance of) « Rock band »" and " Rock band » were a subclass of « music group by genre »" then we would get "Nirvana » is a(n instance of) music group by genre", which is meaningless. What works is to classify classes such as « Rock band » in a metaclass « music band type » with description « music band type whose classifying criteria is the musical genre ».
Using ideas from {{All instances}} it’s possible to describe the « rock band » item by a statement
@TomT0m: I don't have such a problem with the inferred statement « Nirvana » instance of (P31) "music group by genre", since it was a band with a very distinctive genre.
@Jheald: Actually yes, at least in principle. I don’t know what is the best way to express it, but we could add a statement, in the spiret of the forementioned « has quality » solution, that express that any instance of this metaclass is a class that classifies according to the music genre of the band. That would amount to model the « by genre » part. (something like a statement «
⟨ music band type (by genre) ⟩has qualitySearch⟨ has quality ⟩ qualified by—genreSearch⟨ {{{5}}} ⟩
With reference to User:TomT0m's remark above about disjoint union of (P2738) (I assume this is in the sense of a union of sets that are pairwise disjoint), I don't think these genres are necessarily disjoint, especially because some of the genres are specific. For example, there is certainly a large intersection between experimental music and electronic music; funk and soul certainly intersect; Rock and R&B; etc. But also, we are concerned with a general solution, not just this case of music genres, and I'm sure many cases of Category:A [split up by] B the subcategories will have intersections. (Another good example is anything breaking down people by associated place: plenty of people have more than one associated place in the Wikpedia & Commons category systems.) - Jmabel (talk) 03:30, 6 July 2018 (UTC)
I've set QS to do this; might take a good few hours, as it's timeslicing with another 17 or so QS runs. Note that a small fraction of the report rows were or might be female ... I've not set gender for these; can look at the report tomorrow to see what's left. --Tagishsimon (talk) 23:07, 6 July 2018 (UTC)
User:Tagishsimon, thanks a lot! Re females, a manual search had been made before posting here
Yup. Beer&love went on a spree ... there are quite a few male females now. Probably another 20 or so to find; later. --Tagishsimon (talk) 14:09, 7 July 2018 (UTC)
BBLd - Baltisches Biographisches Lexikon digital - year of birth, year of death from P2580
~660 year of birth/~670 year of death can be harvested from P2580 (P2580). The "-None" in the ID is important, because if an ID only ends in -YYYY it could be yob or yod. Here for yob :
This must of have discussed elsewhere, but I haven't been able to find it by googling. I am hoping someone can point me in the right direction.
Templates such as en:Template:Infobox gene pull their data from wikidata. There are cases where a particular wikidata item may be of minor importance or of questionable reliability and hence not suitable for display in an infobox. In these cases, it might be desirable to provide a mechanism where editors can specify the importance or reliability of a particular data item. The infobox could then have the option to display only high importance and/or high confidence data in wikidata. Are there any examples where this has been implemented (or at least discussed)? Boghog (talk) 08:31, 8 July 2018 (UTC)
Various values of the sourcing circumstances (P1480) qualifier can be used to indicate data may have lower confidence or lower precision. An infobox might choose not to display particular cases.
In general, statements that are referenced but wrong should be indicated by setting their rank to 'deprecated', with reason for deprecated rank (P2241) giving any additional information.
Statements that are the preferred choice out of a range of options can be indicated by setting their rank to 'preferred'. Most infoboxes will then ignore any other values given for that property for that item.
Those are the general mechanisms currently in place, but it might be possible to treat some more specific case in the code for a particular infobox. Jheald (talk) 09:51, 8 July 2018 (UTC)
OK, thanks! Now I know what it is called. Help:Ranking gives more details. I was able to change the rank of a statement form normal to preferred, but there was no change in the display of the data in en:Beta-2 adrenergic receptor within the Gene Ontogeny/Cellular component of en:Template:Infobox gene. It looks like the template code will need to be modified to display only preferred rank statements. I will take this up with the maintainers of the template. Thanks again for your help. Boghog (talk) 12:54, 8 July 2018 (UTC)
SQL question
This SQL query, quarry:query/28058, finds Commons categories with an interwiki link to en-wiki, but no sitelink to Wikidata.
Does anyone know if there's then any easy way to find the corresponding Wikidata item for that en-wiki page? I think it would be given in the page_props table in the enwiki_p database, but is it possible to access this from the a query that accesses the commonswiki_p database?
Is there any way to do this from one joined-up query in SQL? Or having got then list of en-wiki pages, is my best bet to then go to something like petscan to find Q-numbers for them all? (It would also be useful to extract whether the Wikidata items have Commons category (P373) and/or existing Commons sitelinks).
Consultation on the creation of a separate user group for editing sitewide CSS/JS
(Please help translate to your language)
Hi all,
I'm preparing a change in who can edit sitewide CSS/JS pages. (These are pages like MediaWiki:Common.css and MediaWiki:Vector.js which are executed in the browser of all readers and editors.) Currently all administrators are able to edit these pages, which poses a serious and unnecessary security risk. Soon, a dedicated, smaller user group will take over this task. Your community will be able to decide who belongs in this group, so this should mean very little change for you. You can find out more and provide feedback at the consultation page on Meta. If you are involved in maintaining CSS/JS code, or policymaking around adminship requests, please give it a look!
When I add a compound given name for a person, for example, "Mary Margaret" at Mary Margaret O'Hara, should I be deleting the original "Mary" or should the given_name field contain both values of "Mary Margaret" and "Mary"? --RAN (talk) 20:44, 8 July 2018 (UTC)
Surely "Mary Margaret" isn't a compound given name? Wouldn't a compound given name be something like "Mary-Margaret" (with hyphen) ?
In France and Switzerland compound names are hyphenated, not in English speaking countries. See the examples given in the Wikipedia article. --RAN (talk) 19:12, 9 July 2018 (UTC)
Wikidata hasn't made up any rules in this regard yet. It depends on whether you consider the person to have one compound given name which has two inseparable words (in that case create a new item for the compound name), or two separate given names which she always uses together (in that case use series ordinal (P1545)). An analogous situation exists for Chinese and Korean given names, which typically have two words, and there's currently no standard on how they should fit into given name (P735). Deryck Chan (talk) 15:58, 10 July 2018 (UTC)
@LydiaPintscher: Not a lot. I was aware of "applies to part" but don't see how to use it in this case: it seems to require that the parts in question have names (and that the names be items in their own right, at that). Would I use "applies to part" with an explicit "no value"? Because that seems weird, too: it seems to say it applies to no part. - Jmabel (talk) 14:50, 9 July 2018 (UTC)
What is the best way to add multiple periods of existence of an organisation to a Wikidata item. For example, en:2nd Battalion (Australia) has been formed and disbanded several times. What to do with that data while editing its Wikidata item. KCVelaga (talk) 15:59, 9 July 2018 (UTC)
I’m having trouble adding ISBN-13 via QuickStatements (using the Zotero export). Should it have hyphens? Quotation marks? - PKM (talk) 04:19, 10 July 2018 (UTC)
@Lea Lacroix (WMDE): Being pedantic, surely it's at least the second time that Wikidata has less than 2 million empty items - as that would also have been true when Wikidata had less than 2 million item? Thanks. Mike Peel (talk) 12:05, 10 July 2018 (UTC)
This is the focus list (i.e. on focus list of Wikimedia project (P5008)) effort for the ScienceSource (Q55439927) project. The aim is to gather up a useful list of scientific articles, open access and biomedical, to forward the aims of the project for medical referencing. The focus list will be a first pass at a list of articles that would be worth downloading.
The page tries to explain how to add useful statements to the list, by hand or mechanically. It doesn't say so much about how the list will be queried to help with the project. Fairly obviously, for those who know the background, seeing which repository or publishers' site hosts the article will be one of the first things to do. SPARQL can do this and quite a lot more.
Everyone is invited to participate. Articles added that turn out not to be quite suitable don't matter so much. We do want to have good coverage of the literature relating to medicine. We'll also be raising the question of quite what the licensing position of open access papers is, which should be interesting.
Generally, the project will be pushing to have enough metadata here to decide questions on Wikipedia on what is a "reliable source".
@Malore: The sintax to CSV is explained there. You have columns to each property; The first one being the "qid" column. Building the CSV to create items, this column is empty, so when copying the csv to QS it will look like:
The tracking week for sales and streaming begins on Friday and ends on Thursday, while the radio play tracking-week runs from Monday to Sunday. A new chart is compiled and officially released to the public by Billboard on Tuesday. Each chart is post-dated with the "week-ending" issue date four days after the charts are refreshed online (i.e., the following Saturday). For example:
Friday, January 1 – sales tracking-week begins, streaming tracking-week begins
Monday, January 4 – airplay tracking-week begins
Thursday, January 7 – sales tracking-week ends, streaming tracking-week ends
Sunday, January 10 – airplay tracking-week ends
Tuesday, January 12 – new chart released, with issue post-dated Saturday, January 16
How would this be modelled, given that each chart is post-dated and this date (i.e. January 16 in example above) is the canonical point in time (P585) for each edition? (Other charts have different rules, such as being labelled according to the data collection period.) Would an item have to be created for every issue of the chart or publication, or would items like Easter − 2 days (Q14795488) (e.g. "two Fridays before last" or "point in time − 15 days") have to be used in conjunction with statements describing each of the events on the item for the chart? This could also be somewhat applicable for other publications which go to press at some point and are published online beforehand but are labelled with an entirely different date. Jc86035 (talk) 10:30, 11 July 2018 (UTC)
In the proposal at Wikidata:Property proposal/chart date (which is marked "not done") it is stated that publication date (P577) can't be used because the date the website is updated is a few days before the stated publication date. I do not agree that this is a barrier to using "publication date". A great many publications of all types make their publications available to the public before the stated publication date. Jc3s5h (talk) 13:27, 11 July 2018 (UTC)
@Jc3s5h: Yes, this is what I meant: how do we indicate both the actual publication date and the date printed in the work? The date of publication couldn't be indicated with object named as (P1932) or similar, because it's not the correct data type. Jc86035 (talk) 15:32, 11 July 2018 (UTC)
It's not customary, outside the Wikimedia Foundation projects, to state the actual publication date. It's customary to cite the publication date cited within the source. The purpose of a citation is to lead the reader to the source, and once tentatively found, allow the reader to confirm the correct source has been found. So the publication date stated in the source is usually what's important. What is your use case for giving the real publication date? Jc3s5h (talk) 15:48, 11 July 2018 (UTC)
@Jc3s5h: The main reason I started this section was the other dates, although it might also be useful to have the actual publication date for some purposes (it might help for validation that some other publication hasn't written about the information in some entity before the entity was made public). Being able to model the time periods that data is collected would probably be useful, maybe in comparing different charts. Jc86035 (talk) 16:31, 11 July 2018 (UTC)
possibly data about when the information is published on the website; Billboard doesn't wait until Saturday to publish articles about its charts
The weakness of this sort of model is that if any of the qualifiers for a statement change, the statement has to be entirely duplicated to indicate the change, with all but the current statement set to deprecated. I'm also not sure what start time (P580) should be in the example (i.e. July 13, the date that the new data collection period began, or July 25, the date that the chart based on that data is labelled with). Jc86035 (talk) 08:41, 12 July 2018 (UTC)
┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ @Jmabel, Jarekt: Would it be useful to have "relative date" properties (values like "-2 days"), possibly with either a set qualifier to indicate what the thing is relative to, and a matching pair property for each like "relative date qualifier refers to" for situations where "relative date" properties are themselves used as qualifiers? This could be extended e.g. with "relative start date" like in the section above. Jc86035 (talk) 09:08, 12 July 2018 (UTC)
I'm not sure I follow all of that, but, yes, the ability to say "relative to this holiday" (Easter, etc.) and give a number of days before or after would be useful. Similarly, to be able to say something like "three days before the first of any month". - Jmabel (talk) 15:14, 12 July 2018 (UTC)
Does anyone actually know the mass of Kylie Jenner (Q1770624) right now? An IP has changed it from 56 to 65, and changed her eye colour statement. I personally think this is not useful information to have, since the mass of a human is highly variable and the statement does not have a time qualifier. Jc86035 (talk) 08:07, 12 July 2018 (UTC)
Added 1 year semi-protection and removed some unreferenced claims which have been vandalized way too often to evaluate the original value and its provenance. —MisterSynergy (talk) 08:28, 12 July 2018 (UTC)
Festál
Festál at Seattle Center (Q5445891) is an annual series of (mostly) ethnic festivals at Seattle Center. I've tried to model a citeable statement that there were 24 of these festivals as of 2009. I have no idea if there might be a better way to model this than what I did; can someone have a look? - Jmabel (talk) 15:31, 12 July 2018 (UTC)
When I edit modules in the module namespace at Danish Wikipedia, the edit window will change to a code editor with syntax highlighting and auto indent. That doesn't happen here at Wikidata where module editing is done in a normal text editor window like in all other namespaces. Could the code editor for modules also be enabled here please? --Dipsacus fullonum (talk) 20:14, 17 July 2018 (UTC)
Missing wikidata properties with property Id P0- P5
Hi all,
Please can anybody give some information about missing properties P0-P5 in the property list. In some text data, I got a mention of properties within this range but here in the official wiki page, I am not able to find them. If they are deprecated, please provide a link from where I can know about them.
It exists, because in the wikidata dumps we can find many entries with P0 to P5 – The preceding unsigned comment was added by [[User:|?]] ([[User talk:|talk]] • contribs).
On San Gottardo railway station (Q16609440) I would like to add the property 'Railway depot'. This property does not seem to exist and is quite separate from station. In this case it is also used as tram depot. No distiction should be made between tram or train. Te distinction between tram and train is in some cases unclear, not to mention metro's and ligthrail. I think railway/tram workplaces(heavy maintenance) should be put on the same (new?) property.Smiley.toerist (talk) 08:46, 12 July 2018 (UTC)
I would create two items for Vashon Island (Q12834566), one for the smaller physical contiguous island, and a second for the combined area of both. If there's some political/administrative structure associated with the latter maybe that's what its class should be, as it's not strictly an instance of "island". A wikidata item should correspond with a single conceptual entity, so two distinct concepts should have two items. ArthurPSmith (talk) 18:56, 12 July 2018 (UTC)
Use an image as reference
In Wat Nong Khet (Q55225353) I added a image from commons as a reference, because that photo of the temple sign easily (if one is able to read Thai) proofs that location statement. However, this gives a warning that the property image isn't supposed to be used as a reference. Is there any better way to do this? Ahoerstemeier (talk) 08:49, 13 July 2018 (UTC)
Am I allowed to withdraw a property deletion proposal? I proposed number of platform tracks (P1103) for deletion in May 2017 due to the data being polluted due to incorrect usage and mistranslated property labels and descriptions, and the discussion still has not been closed. (I think there should really be guidance for closing discussions as no consensus.)
At this point, since the data is still somewhat useful, if the discussion is closed and the property is kept, I think that there should be a check of all values against values of different Wikipedias, and/or the conversion of values to either has part(s) (P527) or new properties if it is found that the count of "number of platforms" is based not on "number of platform tracks", but something else (i.e. number of usable platform edges; number of platform edges in use; number of distinct platform numbers; number of independently usable boarding areas on platforms; number of continuous floor surfaces which have usable platform faces, …). Jc86035 (talk) 15:59, 13 July 2018 (UTC)
Nobels
Hello. It's a detail but I would rather have P3188 (P3188) split into different properties. The IDs would be shorter and would look more like actual IDs than pieces of URLs. Constraints would be tighter. And it would also be useful for the French Wikipedia. We would have articles about scientific laureates displaying links through Template:Research links (Q54913733) while the ones about winning writers would use Template:Literature links (Q54933000), for instance. Again, it's certainly just a detail but would you mind? Thierry Caro (talk) 16:01, 13 July 2018 (UTC)
These fields simply do not work well enough, and they unnecessarily separate all related items that happen at sea. How is the place for a birth or a death at sea different with relation to "at sea"? I heartily agree that we need this more generic descriptor for the non-specific location I just don't see that it should have an event associated with the location when it is solely a location. How/why should they be different places? The term should be a more generic "at sea" where a more accurate place cannot be provided.
This merge and rename would also allow us to capture the missing requirements where people were baptised, married or buried at sea, which cannot currently be captured, and would be a nonsense to create as additional items. — billinghurstsDrewth13:51, 7 July 2018 (UTC)
Fine, then let us create a true location "at sea" and move the birth and death locations to the new item. We can then delete these weird location items that are not locations. — billinghurstsDrewth06:08, 8 July 2018 (UTC)
now fails a constraint on P119: not located in a administrative territorial entity. Possibly an unwanted constraint, since I think it would be undesirable to add everybody buried at sea as exceptions. Ghouston (talk) 06:06, 10 July 2018 (UTC)
I guess the best approach here would be to have "at sea" as an allowed value to the P119 constraint alongside the existing constraints - is that possible? Andrew Gray (talk) 15:16, 12 July 2018 (UTC)
It would be possible, but all the seas and oceans would have to be allowed too, since "at sea" is only needed when the specific place isn't known. Ghouston (talk) 02:44, 15 July 2018 (UTC)
I'm not a poet and I'm unaware of any big poet databases - most poetry collections are focused around good poems from a genre and era, not poets. Deryck Chan (talk) 15:55, 10 July 2018 (UTC)
You might consider asking the Wikisource projects, instead of posting here on wikidata. Many Wikisourcers have been transcribing poetry collections into Wikisource. --EncycloPetey (talk) 15:57, 10 July 2018 (UTC)
@Jmabel: One purpose of creating (classes of) fictional items was not to find fictional items in queries where they are not expected. If there is a strict wall between the two worlds, you can safely search for the citizens of the US omitting properties like_instance of (P31) and still find only actual items and never fictional one. Second, in the cas of Hulk, it seems that he lives in a country that is depicted in movies or comic books, definitely not the real USA. He’s a citizen of some fictional USA, and he’s a fictional characeter. Assuming there is no such thing as a « fictional citizenship ». author TomT0m / talk page15:16, 9 July 2018 (UTC)
That makes sense (especially in this case). Not the way I'd have modeled it, but reasonable.
Do we have a way to distinguish between the setting of a fiction in a realistic place and the setting of a fiction in a more fictional world? I'm thinking, for example, that the United States of Philip Roth's Goodbye Columbus is very much the actual U.S., whereas the one in The Plot Against America is quite fictionalized. - Jmabel (talk) 15:22, 9 July 2018 (UTC)
@TomT0m: I am a bit uncomfortable with this approach, because it will need us to create a whole parallel set of fictional countries, and perhaps a parallel set of cities to go along with London in fiction (Q6671104). The same argument might lead to duplicating all sorts of position-held values ("fictional President of the US"), occupations ("fictional lawyer"), etc just to avoid them showing up in queries. Then we would need to set up all the links between them so that "find me all the fictional scientists" would get all the subclasses etc.
It seems so much simpler to just make sure that we use an appropriate instance of (P31) and stay in the habit of using that for queries, and use all the routine properties and values for fictional people and places. We seem to manage okay at the moment and this would be a change with really dramatic consequences. Andrew Gray (talk) 19:40, 9 July 2018 (UTC)
I disagree. The real USA citizenship is may not be the same with the same USA citizenship. I mean, by adding that the country of citizenship of a fiction character is the real USA, we may be wrong. In the fiction world that country may be bigger, or smaller, or has a different (fiction) president. By say that Hulk has country of citizenship the real USA, is like we say that he lives in the real USA, with the real president, with the real population etc. And what about is the fiction character is the president of USA in the fiction USA? Then that character will has country of citizenship the real USA, but real USA item will have a real president as a value. And if the fiction character is the president of USA, would we add it to real USA item as the president due to value-requires-statement constraint (Q21510864)??? Xaris333 (talk) 20:06, 9 July 2018 (UTC)
@Xaris333, TomT0m: Please go ahead with your logic: if Hulf is citizen of a fictional USA, and Jack Ryan (Q1068314) is citizen of another fictional USA, then you have to create 2 different items for "fictional USA", in order to be able to perform correct queries for each different fictional USA.
The correct way is not to create correct queries: so if you want to want to have all citizens of real USA, then you have to create a query citizen of USA AND instance of human. If you want to have all US citizens of Avengers world, then combine citizen of USA AND instance of Avenger's character. Snipre (talk) 20:47, 9 July 2018 (UTC)
So if a fictional character is the president of USA, we will add it at the item of USA as president? And if you want to have the real USA president, then combine president of USA AND instance of human. That what you are telling us? Xaris333 (talk) 20:50, 9 July 2018 (UTC)
No, of course we don't add it to the USA item as "president". We seem to manage just fine saying
In general, I really don't agree that this is a problem, and I think requiring fictional country items will just make everything much more complicated and confusing. Yes, the Hulk's "fictional USA" may not be precisely the same as the "real USA", but it makes sense to me to use the same item - it's clear that if the item is about a fictional concept, its claims need to be evaluated in that way. Andrew Gray (talk) 21:01, 9 July 2018 (UTC)
Yes, because your proposal is not symetrical: it offers you the possibility to retrieve real citizens or real residents but not US president of Avenger's world. Instead of creating a junk item like fictional USA which mix everything, you should propose a solution which use correctly the data: there is no reason to treat in a better way real data than fictional data or you have to rediscuss the usefulness of adding fictional data in WD.
And about the fact to add the president names to USA item, this example shows why this habit is a bad one.Snipre (talk) 21:06, 9 July 2018 (UTC)
@Snipre: Actually, on a conceptual level, a fiction country is instantiated every time someone live the experience of reading a book, a comic book or another art type, in one imagination. In a sense, a fiction item is a class of experience, so it’s both totally ok to use a singular item that represents « all the experiences of person who lives (read/watch) a fiction in which the USA are mentioned or depicted ». It’s also perfectly ok to subclass it to represent a more precise experience, for example the US depicted in the Marvell universe if we find a reason for it, but it’s far from a logic we have to follow to be consistent internally. author TomT0m / talk page07:26, 10 July 2018 (UTC)
@TomT0m: Thank you for the ontology lesson, but I am a pragmatic guy: one of your main argument to create fictional countries was "One purpose of creating (classes of) fictional items was not to find fictional items in queries where they are not expected". I don't understand why this statement is true for separating real/fictional items and not to separate fictional/fictional items which aren't part of the same fiction world. If you want only real US citizens, I want only US citizen from Marvel's world, so why your logic real/fictional should be favoured ? Snipre (talk) 20:23, 10 July 2018 (UTC)
@Snipre: I’m a pragmatic guy too, at times. It’s pretty clear why we would want to put a clear reality/fiction barrier: I assume unless specified explicitly every query should be assumed to be about the real world, and a user should not have the surprise to see a fiction item and exclude it from the query one way or another. In fiction, this is a completely different story and its not crystal clear which query should be made or not, so the work to build separate worlds is far more questionable, and it is a lot of work. At the limit, you can create items for every reboot of every comic, or every episode, or you can use the same item for a set of work. It’s both less justified and more complicated, so pragmatically the « ontology solution » is a good one, less work and a clear ontological reason that justify the use of the same items for different fictions. A possibility of subclassing that leaves the door open if needed, so it’s flexible. Other argument: there is fictional or mythical analog of (P1074) that can cross the fiction/real world barrier and allows to retrieve values from statements of the real object. We can assume that the value of the fiction item for the property is the value of the real world one, unless in the fiction it’s different (say the USA population has dropped because of a disease in the fiction). Then there is a justification to subclass the « fictional USA » item: there is a property for which we can make a different statement. author TomT0m / talk page07:29, 11 July 2018 (UTC)
Properties for fictional entities work reflect how the entity is depicted in the context of the work. If the work describes a character as being a citizen of the United States, and it is implied that "United States" refers to the country that exists in the world, the item should reflect that. Of course the character isn't a citizen of the actual United States, it's not a citizen of anywhere because it's a fictional character. The Hulk neither lives in a real US nor a fictional US, because the Hulk does not actually live, nor actually exist. The character is, however, depicted as living in the US. A fictional character can't have a real occupation, nor have family members, nor have a gender, ethnicity, birth date, etc, but items have these properties reflect the depiction in an "in-universe" manner. --Yair rand (talk) 22:13, 9 July 2018 (UTC)
@Ghouston: I'm sure you intended that as a relatively frivolous remark, but it sheds light on how complicated this can be. There is no central registry of U.S. citizens. Among others, anyone born in the U.S. is a citizen. They don't necessarily have papers to prove that they were born here. Until rather recently, many, perhaps most, rural southern African Americans lacked such papers. That didn't mean they weren't citizens (although those on the political right in the South did try to make more or less that case in terms of voting). - Jmabel (talk) 16:44, 10 July 2018 (UTC)
If the Hulk were to access US official citizenship records, I'm sure he would find himself listed there. :P --Yair rand (talk) 19:50, 10 July 2018 (UTC)
Not really frivolous, although the method I suggested is wrong. But if I made a list of citizens of the United States from Wikidata, I wouldn't expect fictional characters to be included, since their citizenship status is purely fictional. Saying that I should restrict my list to only humans is mistaken, since only humans (as far as I know) are eligible for such status, and anything else returned is just an error. My preference, rather than creating one or more items for fictional versions of the United States, would be to simply delete the statement from Hulk (Q188760). Ghouston (talk) 03:36, 11 July 2018 (UTC)
To put it another way, Hulk (Q188760) is not a person, but work. It's like saying
@Ghouston: I agree it's a bit odd to need to say "is a human" as part of a query that feels like it should only be about people (and personally, I'm terrible at remembering to do it), but most of the properties used for people are not (currently) defined as "real people only". Strictly speaking, the way to construct a list of "citizens of the United States" is to include a filter saying "and they have to be people" - much as if we construct a list of "things in a geographic area" we need to include a filter to clarify we mean "real things" and don't pick up, eg, historical battles with coordinates.
If we have fictional items, we're going to need to handle them some way or another. I think there are probably four ways (have I missed any others?) to deal with fictional items turning up unexpectedly in unfiltered searches -
- don't have any fictional items, or if we must have them, don't put any "human" metadata on them. This is probably not an approach we want to go down; we have 50,000+ of them already, and it is useful and interesting to be able to run queries for "fictional Canadians", "fictional Presidents", or "fictional dentists".
- normal properties with new fictional values, as @TomT0m: suggested here (country: fictional USA). This is sort of similar to what we do with sex or gender (P21) and animals. One problem here is that it's not just an issue for "country" - there are quite a lot of properties we might want to apply to fictional people (birthplace, educated at, job, etc), all of which might otherwise show up in searches for those topics. It also doesn't work for all types of properties (eg "show me people born on XX-YY-ZZZZ" can't have a value marked as fictional). The biggest problem here seems to be the massive potential for maintenance work to keep the fictional items aligned with the real ones and police their usage.
- new fictional properties with normal values, eg "fictional place of birth: USA". This would be a bit less of a maintenance headache than #2, but we would still need to create parallel "fictional properties" for a wide range of different things.
- define the item as fictional, but otherwise use normal properties and normal values, and explicitly filter searches to human/fictional by adding P31:Q5 in searches. This is more or less the current status quo, but it does mean a bit of extra work in the queries (and we all sometimes forget). We do not have to create any special fictional parallel items, except for eg places that only exist in fictional universes, which we'd need to do anyway.
It's also worth remembering that this isn't just an issue with fictional people. Non-people can have some of these properties - Larry (Q2262318) is undeniably real, for example, and has a real birthplace and date of birth. But you would be a bit surprised to have him returned in a search for "born in London in 2007" if you presumed that meant "people". So the use of fictional values/properties would break here - an instance filter is the only approach that guarantees real-people-not-cats for your searches as well as real-people-not-fictional-people.
Likewise, fictional things aren't just people; we have fictional companies, fictional books, fictional towns, etc. All of these might have various properties to describe them - so we could use the approach of fictional values or fictional properties, but it extends the problem further and requires us to create yet more parallel fictional items or properties (eg fictional co-ordinates?). I really worry that this opens us up to a huge amount of labour and headache.
On the whole, leaving "real values in real properties" for fictional items, explicitly marking them as fictional, and encouraging filtering in searches, seems the least complicated and most sustainable option. It's not perfect, but it's better than the alternatives. Andrew Gray (talk) 12:07, 11 July 2018 (UTC)
« The biggest problem here seems to be the massive potential for maintenance work to keep the fictional items aligned with the real ones and police their usage. » This is a non issue as there is no reason to « keep them aligned » in a strict sense. fictional or mythical analog of (P1074) is meant to connect fiction items to real world one. If you want a strict alignment, just cross the line and get the real world values. More : fictional items may not by the contruction of the fictional universe involved be aligned with their real world analog, so align them strictly may even be counter productive.
« Likewise, fictional things aren't just people; we have fictional companies, fictional books, fictional towns, etc. All of these might have various properties to describe them - so we could use the approach of fictional values or fictional properties, but it extends the problem further and requires us to create yet more parallel fictional items or properties (eg fictional co-ordinates?). I really worry that this opens us up to a huge amount of labour and headache. » Yes, that’s why we have a « fictional stuff» class tree, rooted in fictional entity (Q14897293). Restricting the world to fictional entity amounts to search only classes/instances of this class tree. I agree some kind of filtering this way is necessary if we use the same properties for real and fictional items, and I think no one wants to double the property as it’s far a bigger headache to create a property compared to create an item. author TomT0m / talk page12:24, 11 July 2018 (UTC)
I suspect things would be simpler if we allowed the redundant use of
Larry the cat born in London in 2007 is fine, since it's a real event that happened. A statement about The Hulk being born any place or any time is not fine, because it's not a real event. It would be no better than saying that a fictional character is an instance of human, just because they represent a human in fiction. Ghouston (talk) 02:32, 12 July 2018 (UTC)
From an purely academic point of view, you're probably right that fictional characters aren't actually citizens of real-life countries, but of fictional versions of those, just like they aren't actually humans/homo sapiens, but members of a fictional version. And Bruce Wayne isn't a billionaire, he's a fictional billionaire. But IMO, trying to actually model that in Wikidata would be a huge ton of work for little to no benefit. You'd have to create separate items for cities, countries, organizations, locations, worlds, professions, species, universes etc. for just about every fictional entitity. And consider, there isn't just one Hulk, there's dozens of versions - the one from the original comics, the ones from each comic book reboot of the universe, the one from the 70s tv show, the one from the 2003 film, the ones from the various animated shows, the one from the Marvel Cinematic Universe, etc. Each one would have its own version of the United States. And then there's stuff like fiction within fiction, dreams, parallel universes, different timelines, etc. After Marty McFly returned from the past in Back to the Future (Q91540), was he still a citizen of the original fictional version of the US, or of a different version because he changed the past and lives in a different timeline now? And you'd have to deal with stuff like canon and shared universes to see whether works take place in the same universe and therefore their characters would be citizens of the same fictional version of the US - which is where you'd have to delve deep into the fandom and canon debates to figure this stuff out. And then there's stuff like the whole rabbit hole that is the Tommy Westphall universe... --Kam Solusar (talk) 13:19, 13 July 2018 (UTC)
@Kam Solusar: Please read my comments above, I think I already answered to this point that was raised. We definitely don’t have to create one item per fiction universe if we avoid that approach. author TomT0m / talk page14:19, 13 July 2018 (UTC)
but the existing value is cited as imported from the German-language Wikipedia. As I understand it, I'm not supposed to override a cited statement without a citation, but of course once this information was migrated onto Wikidata it was removed from all the Wikipedias, which now draw it from Wikidata, so I have nothing to cite. Can I just ignore that and make the (I presume) obvious correction? If so, do we really need to keep around the cited, deprecated, (I think) obviously wrong value? How does this work? - Jmabel (talk) 01:16, 14 July 2018 (UTC)
Personally, I'd just make the change. I doubt anybody would insist on retaining incorrect connections just because they were imported from Wikipedia. Ghouston (talk) 04:14, 14 July 2018 (UTC)
Agree with Ghouston. Though before doing so it is always worth checking the Wikipedia in that cited language, to see whether that can give a clue to how the statement that looks odd in English came about. Sitelinks can sometimes be a bit approximate, so a statement that looks odd in English may seem a lot less odd in the context of the other language. Here though the category Category:Planned communities (Q8765027) in de-wiki includes things like "garden city", so should indeed be identified with a rather broader topic that housing estate (Q12104567). Jheald (talk) 07:08, 14 July 2018 (UTC)
Glad to hear it. Do we have any rule about exactly where the line falls between things where we can just make changes like this and where we have to deprecate? Because about a month ago, in a case that didn't strike me as terribly different from this (though it was a bit different), I got reverted as something like "vandalism or ignorance of process". - Jmabel (talk) 08:32, 14 July 2018 (UTC)
So... I made that change and ran into a warning because (I didn't notice this before)
. We do not have an item for "planned city" distinct from "planned community". I can think of more than one way one could sort this out, but at this point I think I'm going to let go and defer to those of you who have been doing Wikidata longer than I. - Jmabel (talk) 08:38, 14 July 2018 (UTC)
In Welsh some of the items listed on the British Q849811 - disambiguation page item would be split in Welsh between Brythonig (relating to things that were on the islands of Britain before the Anglo-Saxon-Jute migration) and Prydeinig (relating to British in its modern sense), which seems to suit the need of other languages who have a Q849811 page. How do I translate the item? Using either / or for the same number wouldn't be correct in Welsh. AlwynapHuw (talk) 02:53, 14 July 2018 (UTC)
In English, I'd use "Brythonic" aliases "Brittonic", "Early British" and "British" for Welsh Brythonig, and just "British" for the modern sense, making the distinctions clear in the descriptions of the various items. - PKM (talk) 22:17, 14 July 2018 (UTC)
The picture itself says he was born 1555, which probably means this is the Elder (now generally said to be born ~1560). - PKM (talk) 00:06, 15 July 2018 (UTC)
Places along a Tour de France stage
Is there a way to include (major) cities along the route of a Tour de France stage? Right now only start and destination point are documented (example: today).--95.222.168.24813:32, 14 July 2018 (UTC)
Hello i am shiv can you help for serval knowledge about "Noval"
like what is "Noval" ??
which type of Noval
Famous noval writher and about Noval
reply as early as posible.
I've been told that for public artworks, like statues, the unveiling date is recorded in Wikidata as inception, not publication date. Ghouston (talk) 02:05, 15 July 2018 (UTC)
There's also an issue that publication date (P577) is not merely "date of publication", but also bearing the alias "date of first publication". There is no means to distinguish date of first publication for an item from the date of printing for a specific version of that item. Many books have a print date for the particular print run as well as an edition date for release of that edition.
There is also no means to distinguish the date printed on a particular publication from the date an item was actually published. Such differences of dating are critical in biological nomenclature, where the actual date of publication is used to determine which author has priority for naming a species or other taxon. --EncycloPetey (talk) 02:23, 15 July 2018 (UTC)
I think there is a difference between the two. inception (P571) is when work started on it while, publication date (P577) is when it was published to the open. These two dates can differ for month or even weeks. I also think I already saw items where this was the case and modeled like this. -- MichaelSchoenitzer (talk) 23:24, 14 July 2018 (UTC)
@MichaelSchoenitzer: You're right, I hadn't thought about it. However, in the case of software products, most of the time what we mean is the publication date, not when the first or the last line of code was written. And currently the use of "inception" is not restricted to these cases, but it's considered equivalent to "publication date".--Malore (talk) 23:44, 14 July 2018 (UTC)
Is the inception date of a building the date when work started on it, or the date when it was completed? Software can be published before it's completed, e.g., on Github. Saying that software is "completed" at any particular point in time would be difficult in some cases. Ghouston (talk) 01:41, 15 July 2018 (UTC)
I suppose you could use the inception date for when work started on it, and various publication dates for different versions (if the software has version releases.) Ghouston (talk) 01:43, 15 July 2018 (UTC)
WE-Framework 2.0
Hello, Wikidata community. Not-well-known WE-F gadget just released version 2.0. There is not a lot of changes outside, but internally it's a full rewrite. Details are following.
New features and secrets
Aliases editing
Gadget supports aliases editing now.
There are some changes for datavalues editors:
Gadget will not break you data until the first edit of particular claim/qualifier. Old-version bug with quantities (when Wikidata team broken compatibility) will not happen. Thanks to redux library.
new time datavalue editor
Internal time datavalue editor now relies on Wikidata API for parsing and formatting of values.
You are not limited to allowed qualifiers constraint (Q21510851) constrain qualifiers. Use new autocomplete feature to select any property. "Correctness" of your choice will remain on your conscience.
There is no more "columns" modifier for claim editors. Claim editors will automatically switch from per-row view to table-view when some magic condition are met. For example, if you have 2-3 values for population (P1082) property you will see "standard" per-row editor. As soon as you have at least 5 values and at least 20% of them have single qualifier (i.e. point in time (P585)) editor will switch to table view. Magic.
There is no "flag" attribute for property editors. Flag will be automatically obtained from country (P17) reference of property.
There is no "search" attribute for property editors. Gadget will automatically take source website for the property (P1896) value and create a magic button for Google search.
SPARQL-based property tab
There is an additional SPARQL-group of properties. Now some editor tab can hold not a predefined set of properties, but a set of properties from SPARQL query. Very useful for quick-populated property sets like authority file (Q36524).
Same applies to external links editor. All tabs there are now SPARQL-groups.
Properties are sorted for your convenience. First goes properties with sites in your language, on second place is current project language, other are sorted by label.
Saving is now single request operation. You don't need separate request for removing claims (as in 1.x version). Well, you still need additional requests to set tags, but they are faster and safe to skip.
Enabled editors settings for we-framework
It is now possible to select which editors are enabled and which are not. Settings are stored in localStorage and bound to particular project. You can have different settings (enabled editors) on different projects.
Technology changes
Instead of single JavaScript file new webpack+babel+(karma+mocha) build system is used. Non-minified version of gadget is 2.7 MiB!
Published version is minified to 850 KiB (before gzip). But it load faster than 1.0 version (because of webpack optimizations)
There are autotests. Not all scenarious are covered, but some are.
Gadget is rewritten with react+redux+reselect libraries. Some third-party libraries are "react-autosuggest", "react-tagsinput", "semantic-ui-react" (for popup) and... jQuery! Yes, mediawiki still relies on JQuery a lot. Dialogs and tabs are "native" JQuery widgets.
React is a fast rendering library (because of virtual DOM usage). But be aware of long tables. Some tabs (SPARQL-ones) may have >1400 different claims. They may require fast CPU.
How to migrate
You don't need to do anything. Update script will handle it for you automatically.
Thank you a lot Vlsergey! Very nice improvement, the "editors settings" in particular is a very good idea (on Wikisources, I only use FRBR Edition and sometimes FRBR Work but never the others). Cdlt, VIGNERON (talk) 10:41, 15 July 2018 (UTC)
What questions concerning the strategy process do you have?
Hi!
I'm Tar Lócesilion, a Polish Wikipedia admin and a member of Wikimedia Polska. Last year, I worked for Wikimedia Foundation as a liaison between communities and the Movement Strategy core team. My task was to ensure that all online communities were aware of the movement-wide strategy discussion. This year, my task similar. Phase II of the strategy process was launched in April. Currently, future Working Groups members are being selected, and related pages on Meta-Wiki are being designed.
I’d like to learn what questions concerning the strategy process would you like to be answered on the FAQ page? Please answer here, on my talk page, or on a dedicated talk page on Meta-Wiki. Thanks!
If you have any questions or concerns, please, do ask!
One objective is to ensure that we have a proper coverage of significant subjects. At this time Africans are less than 1% of Wikidata humans. We do not know its politicians, we do not know the territorial administrative divisions. Our support became worse thanks to the end of Wikipedia Zero. What are we going to do about subjects that have a massive impact on the audience we want? No, Wikimania 2018 does not count as long as there are no results it is window dressing. Thanks, GerardM (talk) 06:38, 15 July 2018 (UTC)
@Multichill: While I will agree that they may be useful aliases to have, they may be better distributed among other languages (whose labels and aliases you can still search in the search bar no matter what your interface language is); for instance, those aliases containing 'vecchio' could instead be moved to the list of Italian aliases, or those with 'velours' to the list of French aliases, and so on as appropriate. Mahir256 (talk) 20:34, 3 July 2018 (UTC)
ULAN doesn't tag labels with a language so I wouldn't know what language to copy it to.
BTW [15] this isn't very helpful. I don't bother undoing it because I know the bot will restore it anyway (unless ULAN dropped an aliase in the meantime). Multichill (talk) 20:52, 3 July 2018 (UTC)
Looking more closely at ULAN, many of these entries are marked "V" for vernacular or local language, presumably Flemish for Jan Brueghel the Elder. Others are marked U for undetermined language. There doesn't seem any basis to conclude that these are valid aliases in English (but not any other language). Throwing them into an arbitrary language (English) just to increase search recall seems to me like a bit of a hack (and violates alias policy). Discarding the information about language and "historical local use" doesn't give us a faithful representation of this source. Maybe we need a separate property for ULAN labels so we can add appropriate qualifiers. Bovlb (talk) 21:17, 3 July 2018 (UTC)
Interesting proposal, since I have noticed that the ULAN aliases are not even complete and I have added aliases myself. It would be nice to have a property for the aliases - we already have the property "nickname" for the Italian aliases known as "bent-names", but I guess we should distinguish between the nicknames (used mostly in archives) and the "misspellings" which are used in museums (though they don't see these as being misspelled. We clearly need to update the policy though if that is the issue, since these aliases are crucial for findability. Jane023 (talk) 21:33, 3 July 2018 (UTC)
What about having only one alias field for all languages? At least for person names, this should be fine and would reduce maintanance work. Steak (talk) 11:14, 4 July 2018 (UTC)
It would look nicer if it just displayed the first three aliases, and kept the others hidden, but used in searches. It would also be nice of we auto generated the alias for names like "John Simpson Doe" as "J. S. Doe" and "John S. Doe" for people that do not have ULAN entries. It would cut down on duplicates being created and aid researchers. --RAN (talk) 15:27, 4 July 2018 (UTC)
I think aliases that are obviously in a different language should be included only in their proper language. Thus “hans holbein der jungere” should not be an alias in English. Nor should “school of Hans Holbein” be an alias for the artist, since it will lead to false matches. Nor should last-name-first variants (and especially broken things like “Edward Coley Burne, Sir Jones”). Yes, this will reduce automatic matches in some tools. But I think that’s better than having alias lists full of utter garbage. - PKM (talk) 19:45, 4 July 2018 (UTC)
Why is a shorter list better than a list which does not reduce automatic matches. What exactly do we gain for this loss in functionality? --Tagishsimon (talk) 21:54, 4 July 2018 (UTC)
Are aliases finding aids, or are they information for other sites and applications to use? What is a third party application supposed to do with the list of aliases for Hans Holbein the Younger (Q48319)? If the feeling is that every recorded spelling in several languages should be valid “English” aliases, then I’ll defer to the consensus. But I think we have done a disservice by importing name fields from catalogues in one language to another, without cleaning up appellations either before or after. - PKM (talk) 03:07, 5 July 2018 (UTC)
To add to this, I personally always interpreted the aliases to be alternative labels; i.e. terms that could have been used as the label, but aren't since we can only have one label at a time.
To mix alternative labels with generic terms that aid in the search for the item seems like bad practice to me; and it would seem better that the two categories of terms were split into separate sections.
On the other hand, if the aliases were always meant to aid searching and not to list alternative labels, perhaps the alias section should have a different name to better reflect its purpose. --Njardarlogar (talk) 10:23, 5 July 2018 (UTC)
@Multichill: We could exclude at least alternative capitalizations and word orders? That would comply with current Help:Aliases and leave out a lot of aliases that aren't needed for search anyway. What do you all think? --Marsupium (talk) 07:16, 12 July 2018 (UTC)
I think the interface could be improved if there are many aliases. It's a bigger problem when there is no alias and a duplicate is created than if there are several aliases. There are still many lists that have aliases for every entry that could be in such lists. I think these should be cleaned up. --- Jura08:07, 12 July 2018 (UTC)
There are a number of related issues being raised here:
Should aliases be limited to those that comply with Help:Aliases, being other common names for the entity in the specific language, without including variations in spelling and word order, and without including related entities? Do we need to balance this restriction against search recall? How does this restriction apply to import bots?
How do we ensure that Wikidata search has good recall? If we create a new property for ULAN, can we use strings from those claims?
How should the user interface deal with large numbers of aliases?
When we import names from a source like ULAN, should we simply make them English aliases, even if they don't comply with alias policy, and even if that means losing important qualifiers about language and usage?
My currently running QS batch does not speed up. The server response time is still at around 13 to 15 seconds per edit. —MisterSynergy (talk) 12:18, 9 July 2018 (UTC)
It is still slow for me too. I am running a batch of 20k+ statements and just clocked it at 220 edits per hour, so it will take 92 hours, assuming it can mantain that rate. Some of my other batches seemed to have slowed down with time. --Jarekt (talk) 12:34, 9 July 2018 (UTC)
@Lydia Pintscher (WMDE): Maybe we should monitor the number and periods when more can edit at reasonable rates (e.g. 10 with 50/per minute). This would allow to better assess capacity problems with the infrastructure. Could you add a chart to grafana? --- Jura13:16, 12 July 2018 (UTC)
@Lydia Pintscher (WMDE): I think a mere count of the number of users who are able to edit at a rate above >50 per minute. With some bots running all the time, there should be at least one. --- Jura13:54, 12 July 2018 (UTC)
As per everybody else. I used to get 60 edits a minute from QS, then more recently 30/min which was barely tolerable, now it's crawling along at 5.
Working on a dataset of 60,000 items that means that to add 1 statement + 1 qualifier + 1 reference is going to take 600 hours -- ie 25 days, even running at 24 hours a day, whereas previously it could be done in two.
Even a casual 1000-edit task is now going to take over three hours, whereas before it would be less than 15 minutes. Six hours if the edits are to be referenced.
I appreciate that previously WDQS was dropping some QS edits, which was also not good, but throttling back this hard makes working with Wikidata simply not viable.
QSbot is a piece of core user infrastructure. It deserves a shared edit-rate quota equal to multiple individual bot accounts. Jheald (talk) 16:24, 13 July 2018 (UTC)
+1 for Jheald's suggestion. The backlog for QS right now is such that I hesitate to add any more batches to it. @Lydia Pintscher (WMDE): do you agree that the interface/toolset is now at a point at which it ill-serves users who wish to make amendments by the 10s of thousands? Are you content that we must wait days for batches to complete? Finally, why is it that QS, a critical tool, is provided and maintained solely by the sainted Magnus, rather than being provided as a core wikimedia offering by wikimedia developers? It gives the impression that wikimedia has scant understanding of or concern for the needs attaching to bulk edits. --Tagishsimon (talk) 19:39, 15 July 2018 (UTC)
Addendum. The short temporary_batches I ran through QS from my client browser today ... 3 edits per minute, tops. I have maybe 150k edits batched up, roughly 34 days to do if done from the client. So I can wait days for QuickStatementsBot to timeslice itself through my queue, or wait a month or so from my client. Neither is attactive. (And for the avoidance of any doubt, none of this is criticism of Magnus; it is, I'm afraid, a criticism of what seems like an antediluvian offering from Wikimedia.) --Tagishsimon (talk) 00:29, 16 July 2018 (UTC)
I somehow suspect that this is a flaw in the QuickStatements tool or any other component of User:Magnus Manske's framework. The dispatch lag and the maxlag value are totally okay these days, and I can’t find any reason why editing should be throttled all the time… —MisterSynergy (talk) 16:53, 13 July 2018 (UTC)
Maybe it's because almost everybody gets throttled. NikkiBot and Citationgraphbot seem to run at decent speeds. I guess I should change tools. --- Jura16:59, 13 July 2018 (UTC)
You might be right Mr S, but I do wonder whether QSbot is being throttled to 60 edits/min to be shared between everybody. If that's the case, it needs to be changed. Jheald (talk) 17:07, 13 July 2018 (UTC)
User:QuickStatementsBot is subjected to the same limitations as all other users, which currently is 80 edits/min to my knowledge. If QS runs smoothly, you can simply use your own (bot) account to edit directly via the QS tool without having to share the edit rate with other users’ batches. The only drawback is that you cannot close the browser window. —MisterSynergy (talk) 17:15, 13 July 2018 (UTC)
@MisterSynergy: This is what I think may have gone wrong, or may have changed. I'm logged in as JhealdBatch in an alternate browser running QS in a window, but I am still only getting 6 edits/min. Okay, so looking at the stats page that Lydia linked to earlier, it does appear that that is being accounted for separately to the 63/min average rate of QSbot; but I can't help wonder whether a cumulative throttle is being applied to all QS edits, across all active accounts. Jheald (talk) 17:40, 13 July 2018 (UTC)
This afternoon I am now getting two edits a minute from QS. Are we seriously trying to build a database here? Jheald (talk) 15:57, 15 July 2018 (UTC)
There has been a lot of interest among the scholarly community in using Wikidata to store linked open data records from academic projects. At the same time, there has been concern about community policies that might veto items scholars wish to add. (This was the main topic of conversation in a 30-person Wikidata break-out discussion at a Linked Data workshop at Digital Humanities two weeks ago.)
We are one such project and are currently evaluating Wikidata as a platform to store biographical facts about people, places, organizations, and events which historians have researched during the course of creating documentary editions. We would want to import these facts from several edition projects into Wikidata, so that they could be re-used by the community and by historians working on documents from similar time-frames. Such a database (of 19th-century USA, in our case) has attracted some support from other scholars who wish to contribute similar datasets of e.g. all soldiers in the United States Colored Troops. But would these contributions be acceptable to the Wikidata community? Does an item like Nathaniel Oldham (a Kentucky barber in 1860 who is mentioned in official documents) belong in Wikidata, or should we look for a different home for this data? Nineteenth Century Digital Cooperative (talk) 17:07, 11 July 2018 (UTC)
@Nineteenth Century Digital Cooperative: Thanks for asking here! Do you have a rough estimate of the number of items that might be created? If it's less than 1 million I think wikidata would be a fine home for this sort of thing. If we are likely talking about millions, tens of millions or more, then it might be better to look into creating a separate wikibase instance (the same software) and looking into federation. I know User:Addshore has done a lot of work on this - for example there's a simple docker image that you can use to run the software now - also see his notes here. ArthurPSmith (talk) 17:30, 11 July 2018 (UTC)
That's certainly encouraging! At the moment we're only talking about tens of thousands of records, though projects like the USCT could bring that into the hundreds of thousands. I suppose that it's conceivable that our project could grow to encompass a census digitization, bringing it into the tens of millions, but that's highly unlikely.
Another question is whether Wikidata would accept items regarding people with incomplete names, such as enslaved individuals referred to without surnames, or individuals who have been researched by historians but whose names are not fully known, since they are referred to as "X's wife", etc. A good example of this from a different project is how Prosopography of the Byzantine World handles "anonymous" people: unnamed strategos of Thrace. Would entries for these people also be acceptible? Nineteenth Century Digital Cooperative (talk) 17:48, 11 July 2018 (UTC)
If you have a project that adds hundred of thousands of entries, the proper way to go about it, is to create a bot proposal and then the bot proposal is the place where the discussion whether or not the dataset belongs takes place.
Items have to point to clearly identifiable entities. If I create an item called "John Smith" it's not possible for anybody to identity which John Smith the item refers to. If a otherwise notable person has one wife and we don't know the name of the wife, that's however enough to identitfy which human is meant. It always has to be possible for other people to build on the item and enough information need to be provided for that to be possible.
As far as sounces go, I consider any item that's well-enough sourced to be of interest of serious historians to have enough sources for Wikidata. ::When it comes to the example of the barber, please fill floruit (P1317) when you don't know the date of birth or date of death but you do know that the person lived at a particular point in time.
In general it's preferential if the data isn't only hosted on Wikipedia but also on the website of relevant project that submits the data and we can link via an external identifier to the data. ChristianKl ❪✉❫ 18:33, 11 July 2018 (UTC)
@Nineteenth Century Digital Cooperative: The way you describe your data sounds to me like it might be a good fit for the FactGrid wiki (described at Wikidata:FactGrid). This is a Wikibase instance that is currently being set up and filled with data (including incomplete information like hypotheses about the identity of a person, location, document or event) related to the Illuminati (Q133957) as a pilot project, but the intention is to make it useful to research in the humanities more generally. Pinging User:Olaf Simons, who runs the project. --Daniel Mietchen (talk) 20:04, 11 July 2018 (UTC)
Thank you for the invitation! It sounds like your project has some interesting parallels to ours. I'm curious whether you imported properties from Wikidata (and possibly some items?) to simplify data donation/reconciliation later? Nineteenth Century Digital Cooperative (talk) 14:27, 13 July 2018 (UTC)
Nevertheless, I would hate to see an item being deleted by mistake. If an item has an external identifier on it (eg linking to a master-record on your site), or even just a statement on focus list of Wikimedia project (P5008) = <an item for your project>, those can both usefully flag up that an item may be in ongoing development, or may have significance as part of a wider project, otherwise this might not be appreciated. Jheald (talk) 14:43, 13 July 2018 (UTC)
I find the notability policy really unclear. "It refers to an instance of a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references. If there is no item about you yet, you are probably not notable." The first part is clear, but is a pretty low bar. The second is vague, and could use a link to a list of examples of notable and non-notable things. The third is just weird, since it implies you can't make an item unless it already has an item. However, Multichill may be right about the deletion request: it will probably just sit there for six months with no resolution. Ghouston (talk) 00:12, 14 July 2018 (UTC)
if you do not understand it, maybe you should not nominate things. maybe we should keep things that "fulfills some structural need". Slowking4 (talk) 03:05, 16 July 2018 (UTC)
I don't see how "structural need" would be relevant for the item I nominated, since no other items link to it. But perhaps I don't understand that clause either. Ghouston (talk) 06:35, 16 July 2018 (UTC)
Property for "known for"
In infoboxes of a person there is usually the "known for" parameter. What would be the most suitable property to collect this information ?.
I have discarded using notable work (P800), because they are not always "creations or works" and even significant event (P793), because they tend to have a chronology of the individual. In each case it is different: it can be known for a work, but also for a discovery or for a malicious act. Any suggestions will be welcome. Thanks, Amadalvarez (talk) 09:52, 15 July 2018 (UTC)
The hook for the image in the en.wikipedia article being "One of these cloaks was included in a painting of Cook's death by Johann Zoffany." --Tagishsimon (talk) 22:01, 12 July 2018 (UTC)
@PKM: This stunning picture of a featherwork cloak appeared in my twitter feed today [16]. Now wondering whether to go to the talk on Thursday. [17] -- being advertised with the headline image from your newly sorted-out item ʻahuʻula (Q8083959). Jheald (talk) 16:21, 16 July 2018 (UTC)
Property to declare the language of the reference url?
Do we have a property to state which is the language of the article/page/url of a reference? I know one could create an item for the reference, but is not always very straightforward. Could language used (P2936) be a candidate for this? Could you ping me if you answer directly to me on this thread? --35.192.51.12316:01, 16 July 2018 (UTC)
For example, the country of Belfast [19] is United Kingdom.
I thought the right value of the country field is UK, and I had tried to change the country of Derry to NI, but an anonymous user reverted it to NI again and again. I think Nothern Ireland's cities must have identical value because of standardization (and Northern Ireland is a part of the UK). Dhārmikatva (talk) 00:19, 17 July 2018 (UTC)
Last year, I've inserted some economists via quickstatements. While, e.g., in Casey Mulligan (Q41805010) the references look fine, in Thomas Crossley (Q41799970) (inserted in the same batch) the reference for instance-of shows an error near the source id ("Issues - format constraint - The SPARQL query resulted in an error."). Similar the reference for the occupation, with an warning symbol near the title ("Potential issues - format constraint - The SPARQL query resulted in an error.") I cannot recognize a violation of the linked format constraints, and have no idea what triggers the messages, but would be happy to adapt my statements generation if somebody could explain to me whats wrong. Jneubert (talk) 11:27, 17 July 2018 (UTC)
As far as I know, this is a problem with the Query Service or the constraint gadget, and likely a temporary one. There is nothing wrong with your data. @Lucas Werkmeister (WMDE) can probably explain this in more detail. —MisterSynergy (talk) 11:43, 17 July 2018 (UTC)
In any event, just to clarify – yes, “the SPARQL query resulted in an error” is never an error with your data (I think). I suppose we shouldn’t show it like a regular constraint violation :/ (edit: filed phabricator:T199788 for that) --Lucas Werkmeister (WMDE) (talk) 12:20, 17 July 2018 (UTC)
@Lucas Werkmeister (WMDE): A second issue is: why are so many of these errors happening now? I'm getting them all the time. Presumably, they indicate WDQS queries failing in some way. Over the weekend I was also seeing a lot of queries that typically run in 30 to 40 seconds now for some reason timing out. Is this related? Is it possible that eg some of the memory management has got really got really messed up on one or more of the WDQS servers? Is there a dashboard that would make it visible, if one or more of the WDQS servers were suddenly underperforming? Jheald (talk) 13:07, 17 July 2018 (UTC)
Thank you, MisterSynergy and Lucas Werkmeister (WMDE). Perhaps it would make sense to prefix messages like these with "Internal system error: ", or even filter out constraint violation messages completely, when a system error occured. Since the end user can't do anything but puzzle about them, they probably would be better fetched in some log file. Jneubert (talk) 15:46, 17 July 2018 (UTC)
"20. century" ≠ 2000–2099
For those who are affected by this, I have proposed at Wikidata:Bot requests#Normalize dates with low precision for a bot to automatically convert all dates with low precision to the range that Wikibase interprets them as being (i.e. to convert a date which currently says "20. century", but is actually +2000-00-00T00:00:00/7, to +1951-00-00T00:00:00/7). This would fix interpretation of date values by external tools (e.g. people stated as having a date of birth from 2000–2099 and date of death in 1996). I believe it is technically feasible with pywikibot and similar tools, although I would not be able to do it myself and I wouldn't do it without consensus for it. Jc86035 (talk) 11:39, 17 July 2018 (UTC)
Documentation of and history of interactive user interface
If I enter an item number in the search box, like Q1, and push enter, I am taken to an interactive user interface where I can view and edit the item.
Does this user interface have an official name?
Is there documentation for this user interface?
Is there any readable description of the development history, that can be understood by someone who was not intimately involved with the development?
The description at Wikibase Client makes it sound like it is an extension that allows the use of Lua or parser functions to extract data from a repository. It seems nothing like the interactive user interface. Jc3s5h (talk) 19:29, 17 July 2018 (UTC)
Hmm, I guess you're right. Wikidata is sort of both the repository and the client - there's some more documentation here. I'm sure somebody else knows a lot more. ArthurPSmith (talk) 19:49, 17 July 2018 (UTC)
It seems that an unusual sequence of renaming lead to the creation of a duplicate I thankfully caught early: https://www.wikidata.org/w/index.php?title=Q55597385&redirect=no . It’s about a french scholar who had an article prior the Wikidata era. The initial item was created with the sitelink fr:Jean-Paul Delahaye, so far so good. Then it was renamed fr:Jean-Paul Delahaye (mathématicien) because a soccer guy has the same name, the sitelink was moved accordingly. Then trouble starts as it appears the footballer is far less notorious than the mathematician, so the article was renamed back by Thibaut120094 (without leaving a redirect) and the sitelink was not updated. GZWDer then create a duplicate item as the article became orphan. What went wrong ? How can we prevent/detect that ? It seems to me that there is a redirect with an item to this orphaned article. Is it a good heuristic? author TomT0m / talk page20:03, 17 July 2018 (UTC)
P27 and UK citizenship
I've started a thread Multiple UK values - reality check pls on Property talk:P27 asking what the consensus is on people items having multiple Country of Citizenship values, for citizens of the UK who were alive before & after 1927, or before and after 1801 or 1707, when the nature and name of the state changed. Grateful for input. --Tagishsimon (talk) 22:28, 17 July 2018 (UTC)
After a stop of new items with cebwiki-sitelinks, GZWDer started creating more of them. Despite their earlier plans to improve these items (what we were told last time), nothing has happened AFAIK. What do you think? Should this continue or be stopped. In cebwiki and/or svwiki, if I recall correctly, these bot stubs stopped being created as there were too many quality issues. --- Jura17:37, 15 July 2018 (UTC)
Lsjbot's creation of cebwiki article is completed and the articles in cebwiki is still being maintained. Note I can not improve any created items simutaneously with creating new ones until phab:T198396 is fixed - for now on I am planning to complete all creation then improve them. Other users (e.g. NikkiBot) may like to help improving before the creation is completed.--GZWDer (talk) 17:44, 15 July 2018 (UTC)
Please do not import elevation above sea level (P2044) for mountain(hills anymore - in 80% of the cases these are totally bogus! Those few I could edit manually I set the value imported (which often was without reference) to deprecated, see e.g. [21]. There are also still many items imported which don't have the coordinates set, hopefully these are now addressed by NikkiBot by matching them with original the GNS database entry. Ahoerstemeier (talk) 18:27, 15 July 2018 (UTC)
This example is not copyright related, is however how can you give us an example that, if any single wiki has both templates, how can't they be merged? --Liuxinyu970226 (talk) 10:10, 16 July 2018 (UTC)
As you seem unable to fix them now anyways, can you avoid creating them until this can be done and you actually plan to do it within a reasonable timeframe (e.g. 1 month). One could get the impression that you create a bunch of duplicates every 6 months and then expect other people to clean them up. Creating these items also make it difficult to import directly high quality datasets. --- Jura07:26, 16 July 2018 (UTC)
There are already 33.000+ items which are either a 100% duplicate, or at least a semi-duplicate due to the administrative unit/populated place separation in geonames. Nobody seems to care to get [22] back to normal numbers. And these are only the duplicates which can be noticed due to the identifier, there are even more which need to be checked by name and coordinates. And while it may be easier to check for duplicates after creating the items - if this is not done even for the trivial cases then better not import. Ahoerstemeier (talk) 07:57, 16 July 2018 (UTC)
Comment (Personal view.) I have no problem with cebwiki imports (or at least smaller than with research papers). Still, a mass of new items with an only sitelink and an only statement, which is just an external identifier, is too few, though. Matěj Suchánek (talk) 08:48, 16 July 2018 (UTC)
1. leaving duplicates unlinked in cebwiki does not encourage anyone to fix them. 2. I plan to add more statements to these items, but I can not do it currently due to phab:T198396.--GZWDer (talk) 08:49, 16 July 2018 (UTC)
The addition of these low quality item discourages the addition of high quality datasets in the same field. --- Jura08:55, 16 July 2018 (UTC)
Strong oppose continue creating items for cebwiki, GZWDer, you're just polluting Wikidata, just stop that work, you're feel free to use your bot on other areas. --Liuxinyu970226 (talk) 10:05, 16 July 2018 (UTC)
I have to agree that avoiding duplicates is very tricky. How is a bot supposed to work it out? The GeoNames IDs we already had are full of mistakes. The old-style interwiki links Lsjbot added to the pages were also full of mistakes. There are lots of very similarly named things which are near each other. While the duplicate items are annoying, I think merging items which are the same is much easier than splitting items where the sitelinks were added to the wrong item. So in my opinion, creating new items is the best of a bad set of options.
As was pointed out, I'm currently running a bot which tries to improve these items, although my priorities are to reduce the number of missing statement violations and add references for the original source GeoNames used. For me, the easiest items to work with are the items with only a cebwiki sitelink and no statements other than the GeoNames ID, because then I know the sitelink hasn't been added to the wrong item and I don't have to worry about conflicting statements, so I don't have a problem with what GZWDer is currently doing.
Cleaning up the mess needs time, but it's easier to do it at Wikidata than on cebwiki as there're more tools for help (constraint violation report, projectmerge, SPARQL, etc.)--GZWDer (talk) 10:55, 16 July 2018 (UTC)
Proposal A
Here are my two proposals:
Remove Phase 1 and Phase 2 support from cebwiki, unless and until if they actually do something to stop Lsj-isms; and/or
Add exception to WD:N/EC that we don't allow creating places items with only GeoNames property.
@GZWDer: In support of what many people have said above, saying that such duplicates are much easier to deal with once they are on Wikidata is plainly not true, if you have no method of dealing with them.
In my experience, dealing with duplicates is a very slow, tiresome, manual process because even having done a merge the resulting merged item then always seems to need further manual investigation and manual clean-up. I am therefore very strongly of the view that creation of duplicate items should be avoided if at all possible.
At the very least, will you not run them through OpenRefine first, to try to first find an existing matching item, rather than just bulldozing ahead with this item creation? Jheald (talk) 10:56, 16 July 2018 (UTC)
If you have no method of dealing with them on Wikidata, you have no method of dealing with them on cebwiki either. Information in cebwiki is not structured.--GZWDer (talk) 10:59, 16 July 2018 (UTC)
Here's another reason that you should really stop it, that your actions result The SPARQL query resulted in an error. happened on image (P18) of every items, *at least five times per item*. --Liuxinyu970226 (talk) 11:27, 16 July 2018 (UTC)
proposal B Based on the planned activity by Nikki, I suggest we create only new items when they are being referenced with another identifier than Geonames (at item creation). If there is interest in dealing with the current backlog and this can actually be done, we might want to review if more items for the 600,000 bot pages should be created. Items for manual articles at cebwiki can always be created or articles connected to existing items (if there are any contributors there). --- Jura12:23, 16 July 2018 (UTC)
Neutral While this matches my second point above (excluding any creation of items that only have Geonames property), I wouldn't support this kind of solution easier, as there are still no potential good sanctions to cebwiki, which is required to prevent more wikis from such behaviors. --Liuxinyu970226 (talk) 14:38, 16 July 2018 (UTC)
I am getting concerned about your mental health: "Sanctions to cebwiki." There's nothing like sanctions to a wiki, no. They are not our business. If you don't like it, don't care. Matěj Suchánek (talk) 17:23, 16 July 2018 (UTC)
I don't understand how it is related. Maybe because he believes it is a possibly invalid entry requiring further references? Matěj Suchánek (talk) 07:06, 17 July 2018 (UTC)
I'm not sure I used the right solution there, but those were cases where I was trying to find a location in Australia and wikidata had a huge number of them almost all imported from cebwiki; when I looked at (a few?) coordinates they specified in Google maps, there seemed to be no park and nothing matching that label at the given locations. So I marked ones that looked wrong to me, as I recall. ArthurPSmith (talk) 13:53, 17 July 2018 (UTC)
Support Planning that it "is much easier to find such duplicates if the data is stored in a structured way", so deliberately importing duplicates (which won't be merged within short time) is an abuse of Wikidata and our resources. Resources spent on cleaning the mess of some origin are missing to bring high quality data to other wikis and elsewhere. The duplicates are a big problem, they pop up on search and queries etc. Sitelinks might be added after data is cleaned off-Wikidata (if cleaning is feasible at all; no idea perhaps deletion of articles on cebwiki is a better solution than importing sitelinks to them here). --Marsupium (talk) 23:26, 18 July 2018 (UTC)
Items about Wikimedians
I have witnessed two instances of a group of items about Wikimedians being nominated for deletion due to lack of notability. The first was in September, in which Jarek argued for lack of notability but MisterSynergy opted to keep due to the presence of Commons creator templates. (At the time—when I wasn't yet a sysop—I tended to agree with deleting all but those of @Moheen Reeyad, Emijrp: since they seemed to stand alone better than the others.) The second was several days ago, in which Bodhisattwa argued for lack of notability and which Maarten deleted on such a basis—not the only major difference between the two acting admins in these instances. As I had begun taking the liberty of pinging frequent, accustomed editors whose items were up for deletion, I didn't think it appropriate to take any action in this recent set of requests until the affected users were able to opine about them. Now there is a new, isolated instance which is only prompting yet another long notability discussion to occur (@Mazuritz, Ash Crow, Infovarius:, as those involved in this instance, plus @Esh77: as I ran across your item some time ago).
It is clear that there is something unresolved about WD:N as it pertains to Wikimedians making items about themselves. Of course there may be something I don't recall in P4174's proposal, among all which related to privacy issues and whatnot, on the topic of how that property applied to notability criteria, but what logically comes to mind about this is as follows. On Wikipedia we routinely RfD clear self-promo articles by amateur rappers, actors, SEO folks, and the like, all of whom very likely think they are sufficiently notable for Wikidata, about themselves—and often do the same to the appropriate Wikidata items much to their chagrin—so ideally items by Wikimedians who may very likely think they are sufficiently notable for Wikidata about Wikimedians should nearly equally share the same fate. A strict interpretation of WD:N under the second and third points tends to preclude their deletion, however, but I can certainly understand the mindset of those who find it odd that contributors to this knowledge base find it in them to add themselves to it (I seem to recall @Liuxinyu970226: complaining about @Shizhao: on this front). Some may find it unfair that those most familiar with those workings of Wikidata available to them, who otherwise would most likely fail the notability criteria had others tried making items about them, can "weasel" their way to meeting said criteria by the aforementioned strict interpretation, whereas others who try to do the same—and who could very well meet the notability criteria had someone more thoroughly familiar with Wikidata created their item—get blocked and their items nuked. Let me make clear that I am not condoning the presence of self-promo/spam with this last statement.
I think the items should be permitted for consistency, since authors of notable works can have items, and Wikimedians are authors of Wikipedia. Ghouston (talk) 04:30, 17 July 2018 (UTC)
When I notice Wikimedians creating items about themselves I delete them on sight as non-notable. If I notice items being created about other Wikimedians, I check if they're really notable and usually deleting them. Far to many privacy and conflict of interest issues. I deleted items about some people more than once and I even deleted an item about myself. Multichill (talk) 08:43, 17 July 2018 (UTC)
Then it's not so much a notability problem but a biography of living persons issue, with information that isn't properly referenced. There are some items like Jimmy Wales (Q181) and Magnus Manske (Q13520818) for notable Wikimedians, although even these have unreferenced statements. E.g, Jimmy Wales' spouses and Magnus Manske's date of birth and education. Ghouston (talk) 09:37, 17 July 2018 (UTC)
If we don't have an item, we're not going to be able to write a depicts (P180) statement on CommonsData for any of the photos of this person. Is that a problem? Jheald (talk) 10:02, 17 July 2018 (UTC)
There was a proposal to have some hybrid data type for author fields on Commons data that allows us to specify people either be item ID, username or URL. I assume we will use something similar for depicts (P180). For me the issue with items for not-so-notable wikimedians is that in such items most or all the information comes from the subject with no proper references. If there are external identifiers they are self-created. On the ohtr hand I do not want to punish otherwise notable people who might want to edit Wikipedia, so the thershold should be: is there enough non-self-referencing information out there about individual. We should delete non-referenced and self-referenced info from such items and see how much is left. --Jarekt (talk) 12:13, 17 July 2018 (UTC)
In a previous discussion, it was suggested that we can create items for authors even if practically nothing is known about them, e.g., the thousands of authors of the CERN Higgs Boson article. These would fulfil a "structural need" in case there was another article with the same author. The same could just as well apply to Wikimedians. Ghouston (talk) 12:29, 17 July 2018 (UTC)
When asking for this discussion, I was thinking of a loophole in WD:N: many items about French Wikimedians where created last year by @Deansfa: (@Envlh: warned me that one had been created about me) on the ground that there is a Commons category for these contributors. I wanted to delete all those items but found that I could not, as they are neither excluded by the current redaction of the 1st criterion of WD:N, nor listed in WD:Notability/Exclusion criteria. If we are to delete all those items, we should explicitely exclude them in one of those two pages to avoid them recreated after, either by someone acting out of good faith like Deansfa, or by someone with bad intentions (to be clear, I fear that such items can be used for targeted harassment far more than I fear self-promotion from contributors.) -Ash Crow (talk) 15:12, 17 July 2018 (UTC)
It's useful to have the items where there are categories of photos of the specific Wikimedian on Commons - although I agree that they shouldn't be created by the Wikimedian themselves. Thanks. Mike Peel (talk) 15:13, 17 July 2018 (UTC)
Many long-term wikimedians have category on Commons and some even have Creator pages but that does not make them notable. The fact that shch pages exist might make it harder to clean up but should not be a reason to justify creation of items on Wikidata. Maybe we need to clarify this on WD:N. As for authors of articles we have author name string (P2093) for that with 17M uses. Maybe we should replace some items with no other info than name with those. --Jarekt (talk) 17:58, 17 July 2018 (UTC)
@Multichill, Ash Crow, Infovarius:, The element Q40676142 Benoit Soubeyran was deleted today. For me, it's not to make self-promotion, it's just a good way use the tools and templates linked with Wikidata on Commons and on my wikipedia user page. It's not right because there is a lot of elements on wikimedians : Benoît Prieur, Moheen Reyad,... @Agamitsudo, Moheen Reeyad: Of course, I'm not against these and I think they have clearly right to exist. But if you want to delete the wikidata elements of non-notable wikimedians, you must delete all this elements and not only one !!! I prefer a large discussion before to decide if we must keep or delete all these elements and what the wikimedians thelmselves have right to edit or no. --Mazuritz (talk) 20:14, 17 July 2018 (UTC)
I found an interesting issue with Wikimedia username (P4174) in it's proposal, Wikidata:Property proposal/Wikimedia user name. User:Innocent bystander opposed it, and felt if it was created, there should be a requirement that a reliable source should be required. I think I'm qualified for a Wikidata item because I am a low level elected government official, and this can be verified on state and town web sites. But I do not have a famous blog, and none of the handful of reliable sources about me mention any of my Wikimedia user names. Therefore, if an item ever got created about me, no one would be able to add Wikimedia username (P4174) to my item. If I were to make any on-wiki statements about my item, they could only be used to help decide whether to delete it, not to decide that the Wikimedia userid belongs to the person described by the item. Indeed, it is rare for an interview or third-party biographical piece about a person to mention a wikimeda userid, so virtually the only people eligible to have Wikimedia username (P4174) in their item are those who are reliably associated with a website or blog, and self-publish it there. Gerry Ashton (talk) 21:13, 17 July 2018 (UTC)
The purpose of Wikidata is not notability but data, and accuracy of the data. Wikidata is a meta-database: any item should be referenced by at least the "unique key" (or "unique identifier") on an other database (and Github, Wikimedia Global or Twitter have unique identifiers for users and we have properties for those). This way, we can query data across databases and across the world, across any type of knowledge (and it's potentially super powerful). (BTW, Wikidata is massively used on Commons, so deleting the items just remove the data that is exposed there).
For those who delete items, how many of you are making actual queries against Wikidata, and are using the data? I do. I'm creating items about Wikimedians for the same reasons I'm creating items for every building in Manhattan (which, for lots of them, are not notable): to be able to get, for example, all architects that has build something on 56th Street as get every humans that have a female gender, that have wikimedians as occupation and have a Github account. Wikidata is about querying the world.
We are slowly moving to having all humans represented... Either we should establish the threshold, or we should confess that we want to have items about all humans. Millions of Wikimedians is only the step... --Infovarius (talk) 22:45, 18 July 2018 (UTC)
Link to how to merge
Hello! Wikipedia administrator/Wikidata newbie here. When I add a translation to an item (Q55605143) which reveals the item to be a duplicate, I get the message "The link zhwiki:北京首都旅游集团 is already used by item Q55065197. You may remove it from Q55065197 if it does not belong there or merge the items if they are about the exact same topic." I would strongly suggest adding a link to Help:Merge to this message, as finding how to merge two items was not easy. Or, better yet, a link to Special:MergeItems with the two IDs prefilled.
Hello Dan! This would be the original template message, changing which requires modifying this list of English messages. I'd modify the message myself to at least link to Help:Merge, but I have yet to actually start keeping a clone of the Wikibase source code handy. I am sure that if you file a bug report on Phabricator, someone there will gladly oblige. Mahir256 (talk) 02:58, 18 July 2018 (UTC)
Help us teaching ORES how to better detect vandalism
Hello all,
As you may know, ORES is a tool analyzing edits to detect vandalism, providing a score per edit. You can see the result on Recent Changes, you can also let us know when you find something wrong.
But do you know that you can also directly help ORES to improve? We just launched a new labeling campaign: after authorizing your account with OAuth, you will see some real edits, and you will be asked if you find them damaging or not, good faith or bad faith. Completing a set will take you around 10 minutes.
The last time we run this campaign was in 2015. Since then, the way of editing Wikidata changed, some vandalism patterns as well (for example, there are more vandalism on companies). So, if you're familiar with the Wikidata rules and you would be willing to give a bit of time to help fighting against vandalism, please participate :)
@Lea Lacroix (WMDE): I'm trying this out (I've already nominated some spam here and on Commons for deletion). For edits which involve sitelinks being automatically changed or removed after a page move or page deletion (e.g. Special:Diff/520792890; Special:Diff/530855598), am I always supposed to label them good-faith constructive edits (even if I think the deletion or move was wrong)? And for edits which are formatted incorrectly but are otherwise constructive (e.g. Special:Diff/616158518), do I label the edit as damaging? Jc86035 (talk) 12:48, 18 July 2018 (UTC)
Is this addition of a site link to a machine translation that was deleted the next day damaging? I don't know how I'm supposed to classify these, especially without options between "damaging" and "perfect" and between "good faith" and "bad faith". Jc86035 (talk) 13:10, 18 July 2018 (UTC)
User:Jc86035 Thanks for the labels, The rule of thumb here is that even if this edits are damaging on client wikis, fixing them is outside of scope of Wikidata (for example, we should not revert those edits, but we need to fix client, like commons, and it will be automatically reverted here). So please mark them as good-faith and not damaging Amir (talk) 13:15, 18 July 2018 (UTC)
@Ladsgroup: Thanks. If an edit is constructive but needs another editor to fix it because of formatting issues, should it be marked as damaging? Jc86035 (talk) 13:19, 18 July 2018 (UTC)
I'm getting lots of messages that "The SPARQL query resulted in an error" for property values that had no such errors an hour ago. --EncycloPetey (talk) 19:37, 18 July 2018 (UTC)
Polish article is about generic concept of library located in or run by a church, especially in historic times. --Jarekt (talk) 03:47, 23 July 2018 (UTC)
Could somebody please import the en:Template:Graph:Stacked to wikdata so that it can be used here? Currently there is some kind of gap in creating such graphs for articles, which is probably the reason why it is not used so far: For the template, queries are needed, which are typically not totally trivial, but at en:wp there is no help page where queries can be requested. On the other hand, here at wikidata there is a query request page, but the template cannot be used because its not here so far. It would be nice to close this gap and make the template available here. Steak (talk) 13:57, 24 July 2018 (UTC)
I don't know where to put this, but can someone keep an eye on Kamisettyramesh's edits? I don't know what their native language is, because they haven't edited a text field yet, but they seem to have no understanding of the distinctions between items with similar labels and they are not responding to talk page messages. Thanks, Jc86035 (talk) 09:13, 19 July 2018 (UTC)
@Jc86035: The only Indian-language site this person's account has been linked to at present is the Teɶlugu Wikipedia, so I have made the assumption that this person speaks Telugu and pinged three recently active Telugu speakers. Mahir256 (talk) 15:17, 19 July 2018 (UTC)
Starting new page
What are the details of creating musician page? (Q55655133)
@R3adyWrit3s: I suggest that you look at the item for some comparable figure and follow that as a model. By the way, did you really mean to write those aliases all lowercase? Is that how she writes her own name? - Jmabel (talk) 23:10, 19 July 2018 (UTC)
Hello everyone! My bot works on interwiki maintenance and sometimes it finds interwiki conflicts and stores them to a log. Now I decided to reorganise and publish these logs. Here is a first part starting with 'A':
Every conflict contains two transition chains starting from same page and ending with different (local pages on same language or Wikidata items). There are 3 types of transition:
'(wd) ->' - next page is obtained through Wikidata item
'(local) ->' - next page is obtained through interwiki on the local page
'(redir) ->' - the first page is a redirect to next (redirects to sections are excluded)
You can see that conflicts with redirects twice more than without. Also, a lot of redirect conflicts are false positive since they links narrow theme with its generalization (such as a song and an album with this song) and are not exactly the same. That's why I separated redirect and non-redirect conflicts.
I invite everyone to check these conflicts. Hope it will be useful. Later I can publish conflicts on other letters and other languages. --Emaus (talk) 19:38, 19 July 2018 (UTC)
My bot from time to time removes interwiki from local pages but only when there is no interwiki conflict. In case of conflict it is not obvious.
Also I have an idea to select separately conflicts caused by Wikidata when 2 found Wikidata items have links to a different languages/ Probably there should be simple merging. --Emaus (talk) 17:30, 21 July 2018 (UTC)
My version isn't used, I use the "official" but, @Xaris333: what do you want translate? The only text is "Check sitelinks!" in function init() --ValterVB (talk) 08:57, 21 July 2018 (UTC)
In QS v2 all cells must be filled?
@Magnus Manske: In Quick Statements v2, is there a way to jump a certain property/reference/qualifier for a given item (for example, if I have a table with a "P31" column but for one item of the table I don't know its class).
Alternatively, load (or cut-and-paste) the file into an editor Notepad++ in tab-separated form, then use search and replace from the menu to replace all instances of \t\t with \t and \t\n with \nJheald (talk) 13:10, 21 July 2018 (UTC)
Need some help on how to list material designations according to different standards
Hi, I started a discussion on the materials project to find out the best way to list material standards. It is a long topic so that the discussion is on the project page : Wikidata talk:WikiProject Materials/Properties. It would be nice to have some advice from experimented Wikidata users. ;-) --Thibdx (talk) 06:26, 21 July 2018 (UTC)
Tags removal in Watchlist
Hello, is there a way I can remove all the edits with specific tags, such as edits by QuickStatements and a couple of other tools? For quite some time I am facing difficulty with my watchlist filtering (here how the watchlist looks now). I am using old watclist interface. Thanks in advance for your help. --Titodutta (talk) 14:39, 21 July 2018 (UTC)
No. (Sorry this answer is so brief.) However, if you are ok with just hiding (in contrast to filtering), you can use CSS .mw-special-Watchlist .mw-tag-OAuth_CID_699 { display: none }. Matěj Suchánek (talk) 15:00, 21 July 2018 (UTC)
For those who deal with patrolling, I wrote a userscript which enables patrolling interface in histories, user contributions and multi-change diffs. It highlights unpatrolled changes in histories, marks unpatrolled changes in user contributions and adds convenience links to patrol all relevant changes to all three interfaces. You can enable it by adding this code to common.js:
Four months ago, there was a dispute over the validity of community decisions made off-wiki. Some users will occasionally discuss Wikidata issues using off-wiki methods of communication, such as the mailing list, IRC, direct email, social media, and even at IRL events such as Wikimania. I think we need to determine whether decisions made using such off-wiki channels can ever be called "community consensus" and be considered valid. --Yair rand (talk) 23:20, 11 July 2018 (UTC)
The more people, the harder it gets to get any consensus. Besides sometimes it is difficult to take any decision on wiki, because people just drop their "oppose" and disappear from the conversation, which is very uncivil but it happens very often. In general if it is an important consensus for the on-wiki world, I guess it is necessary to inform about it at least here on the project chat. --Micru (talk) 11:08, 12 July 2018 (UTC)
I would say that any off-Wiki discussion like this should be posted to Project Chat, more or less in the format "In a recent conversation at [place], a group of Wikimedians agreed that [whatever was agreed to]. Notes from that conversation are available at [link]. Is there general consensus here that this should be our best practice? Is there something we failed to consider?" In the final analysis, consensus can only be achieved by making the conversation available to all Wikidatans. Not all Wikidatans will have the opportunity to participate in a discussion at an event, or on Facebook or irc. - PKM (talk) 20:06, 12 July 2018 (UTC)
Aren't there projects that have rules banning discussions on Facebook and the like? --- Jura22:19, 12 July 2018 (UTC)
you cannot govern off wiki activity, by fulminating here. if people think their deliberations are appropriate, then they will mention it on chat. if you want to encourage more on wiki discussion, then you need to make it more friendly, and easier than wikicode. Slowking4 (talk) 02:58, 16 July 2018 (UTC)
I don't think that any off-wiki decision should be treated as producing a consensus as it excludes people who aren't part of the venue in question from the discussion. ChristianKl ❪✉❫ 15:58, 22 July 2018 (UTC)
When sourcing is most relevant
When sources differ of opinion, when both cannot be correct, it is a matter of sourcing the right answer that will ensure that Wikidata has the correct information. I have been told that information I added based on en.wp should be removed because it is "Falschinformationen". Obviously the German Wikipedia thinks differtly from en.wp. Based on German information information is completed but we now have a half way house of correct and incorrect information.
In my opinion, this is exactly the situation where sources are vital. How to deal with this and how to reconcile differences. Thanks, GerardM (talk) 06:43, 15 July 2018 (UTC)
What you describe is what I could come up with. What I am doing is adding "administrative territorial entities" for Africa. I am bold; relate information based on available items in Wikidata (yes, I do use Arabic, Cebuan, Italian links).
Adding sources to Wikidata like in this instance does NOT solve the issue I raise. The issue is how to curate multiple sources, particularly Wikipedias that disagree on factually singular information. That is the issue, not what I can do, I am aware what I could do. Thanks, GerardM (talk) 08:17, 15 July 2018 (UTC)
We can look at the question of whether Tumana district exists in Gambia, or not. It's mentioned at [23] as a constituency in the Upper River Region, as of Jan 2018. So it doesn't seem to be fake information. Perhaps there's a difference between electoral constituencies and administrative districts, more information is needed. That article also mentions Foday N. M. Drammeh, who is also named as the winner of Tumana at en:List of NAMs elected in the Gambian parliamentary election, 2017. Ghouston (talk) 10:09, 15 July 2018 (UTC)
There's a table of 2017 National Assembly election results (cached) at [24]; I think items could be created for all those constituencies and winning politicians, if not already present. Ghouston (talk) 10:21, 15 July 2018 (UTC)
"Being bold" is a slogan that means that it's good to do something on a Wiki because it's easy to revert edits. The fact that many bold edits get reverted is part of the idea.
I agree with Ghouston that the fact that Wikipedia is not a reliable source in inside the Wikiverse. It doesn't matter what enwiki says if there an argument about whether a given piece of information is true.
Are there any guidelines for when and for how long pages should be protected? I think it should be at least standardized, and making semi-protection a little more common would make it quite helpful for editors on other wikis (e.g. the English Wikipedia) who get complaints about how this anonymously transcluded Wikidata content has been vandalized and not fixed for two weeks. Personally I would prefer allowing semi-protection for 3 years or more of any pages that normally get vandalized more than about once every two months on average, because that would at least take care of most of the random innocent vandalism by people who don't know any better. (Should there be an RfC for this?) Jc86035 (talk) 08:17, 18 July 2018 (UTC)
+1 If we offer a large flexibility in term of editing, we should have s strong response in case of vandalism. And several years of protection is not too much: often the items which are victim of frequent vandalism are controversial topics like religions or personalities, and this vandalism won't stop rapidly. Snipre (talk) 11:10, 18 July 2018 (UTC)
On one hand, yes, on the other hand, many users from the projects are not autoconformed at Wikidata and will not be able to edit protected items. We need to find a middle ground.--Ymblanter (talk) 22:31, 18 July 2018 (UTC)
Would it be technically feasible for a bot to auto-assign confirmed to users who are autoconfirmed on other wikis? Jc86035 (talk) 04:21, 19 July 2018 (UTC)
Why would we want to autoconfirm others who have tht status on other Wikis? The fact that someone participated on other Wiki's in no way implies that they understand how Wikidata works and what edtis are appropriate for Wikidata and thus shouldn't be subject to patrolling. ChristianKl ❪✉❫ 15:29, 22 July 2018 (UTC)
Erratic enwiki import of infobox values
Is it standard practice to import infobox values without validating any of them? Nvrandow is currently importing station code values from enwiki, some of which duplicate existing values from different properties and some of which are actually multiple values for different lines or different platforms (which should have been split and have had qualifiers added properly). I just reverted one edit which contained br tags but they haven't noticed yet. Jc86035 (talk) 05:20, 21 July 2018 (UTC)
@Jc86035: No, it certainly isn't (or if people have been doing it that way it shouldn't be). I put in a 6-hour block so Nvrandow can discuss these matters with you. (The Wikipedia with the most edits by Nvrandow is German, so you may wish to call on German speakers if the person's English isn't that great.) Let us know on WD:AN if these errors persist after the 6-hour window. Mahir256 (talk) 05:31, 21 July 2018 (UTC)
@Mahir256: I just realized that they have never replied to a single post on their talk page (including the wikitext archive), even those of the "stop why are you doing this" variety (and they have edited some labels in English), so it might be better to just mass-revert the whole import and do it properly some other time. Jc86035 (talk) 05:49, 21 July 2018 (UTC)
@Jc86035: You may wish to call on someone else to help you with that, as I have not previously set up such a mass revert myself. In the meantime I will extend the block to three days so that a course of action can be formulated with regard to future station imports. Any admin may undo this block earlier if he or she desires, but please take note of the degree of error that has been pointed out here and elsewhere. Mahir256 (talk) 06:38, 21 July 2018 (UTC)
@Pasleim: would you consider making Harvest Templates compatible with EditGroups, so that we can undo problematic batches more easily? The instructions are here. I am happy to assist if needed. − Pintoch (talk) 16:54, 21 July 2018 (UTC)
While I haven't received a response, it looks like an IP has gone over most of the duplicate values and formatting errors, so a full revert is probably out of the question. Jc86035 (talk) 14:28, 22 July 2018 (UTC)
Nowadays I'm observing more and more merges like this where a newly created id by a tool is preferred over the older concept QId. I have the impression this related to tools developed by Magnus. I thinks concept ids should be stable and not be changed at will. --Succu (talk) 21:38, 21 July 2018 (UTC)
@Succu: I'm not sure I understand what upset you here. There was a duplicate pair of items for the same run of the same magazine, so I merged them. Is there an issue to resolve here, or can we close this? Jheald (talk) 15:08, 22 July 2018 (UTC)
@Jheald: This was not about your specific edit (witch I misread) but about a general observation I made in the last few weeks. --Succu (talk) 17:47, 22 July 2018 (UTC)
Renaming a city or country vs creating a new project
Hey everyone. I have a question for the community. The town Grahamstown in South Africa is being renamed to Makhanda. The question came up as to whether the wikidata object should be renamed or whether a new wikidata item should be created.
The problem that I have had with the former is by renaming the item I will use the property replaces (P1365) of succeeds. However, in order to use Grahamstown as the value I need an item that it refers to.
The problem with the later is that I now have two items where in the real world only one place exists.
My main aim here is not to start a philosophical argument but to rather ask what is the correct way to create items in such a case.
– The preceding unsigned comment was added byTirab (talk • contribs).
For items like embroidery stitches, pockets, and sewing seams, I would like to use two images: one a photo of a representative item, and one a diagram showing its structure - see example at satin stitch (Q1982711). This isn't technically a cross-section - should I use sectional view (P2713) for the structural image, or should I propose a new property for this? - PKM (talk) 20:02, 19 July 2018 (UTC)
I noticed that we have comic book series (Q14406742) like The Amazing Spider-Man (Q944598), but (at least this item) does not have the individual issues within that series. Is that data that should be here? Should each issue be a new item (I can imagine a lot of properties that would be used, like characters in the issue, writers, artists, etc.). I imagine perhaps we already do this somewhere? Regardless, if this data should be included (I think it should), what property should be used? I was thinking has part(s) (P527), but maybe there is something more specific (or there should be)? Thanks! --U+1F360 (talk) 00:17, 23 July 2018 (UTC)
Yeah it would help if I read the instructions! Thanks for your help! I'm excited to help add this data! --U+1F360 (talk) 00:35, 23 July 2018 (UTC)
@U+1F360: Big job, needs doing. Given the scale, perhaps consider gathering the data together in a spreadsheet and then using QuickStatements to create & populate issue items, and handle the 'has part' business in the series record. Might be the easiest & most robust approach. --Tagishsimon (talk) 01:16, 23 July 2018 (UTC)
Parks to which buildings have been moved
Do we have a particular subclass of park (Q22698) for the sort of park to which a number of historic buildings have been moved? Or do we have a particular pattern for describing them? They are pretty common in medium-sized U.S. towns. An example is Bothell Landing Park (Q49475372). - Jmabel (talk) 19:19, 22 July 2018 (UTC)
It's not unlike that, but it's something more informal. I don't think much anyone things of Bothell Landing as a "museum". There is one building there that is a museum as such. - Jmabel (talk) 21:09, 22 July 2018 (UTC)
In this case, no one lives there. Definitely a park, not a district of the town. One building was an old one-room schoolhouse; another is a house that now functions as the parks dep't HQ; another house is now a museum, but not really a house museum, they just re-used the building; etc. - Jmabel (talk) 03:37, 23 July 2018 (UTC)
I've created historical park (Q55720514). The practice of moving old buildings, which in Europe would probably be seen as an heresy most of the times, is actually pretty common in the United States, where there are a lot of such sites where they are gathered – there is one, for instance, very next the Oconaluftee Visitor Center (Q36000768). This is why it may sound confusing to some users here or why there was no item so far. Thierry Caro (talk) 11:12, 23 July 2018 (UTC)
I tend to think of an open-air museum as something more like Commons:Category:Pioneer Park, Ferndale, Washington (which lacks a Wikidata item). But Bothell Landing is more a place you go to look at the birds on the water, or to access a bike trail, or even to watch a musical performance than to look at the buildings (most of whose interiors are seldom or never open to the public). - Jmabel (talk) 22:21, 23 July 2018 (UTC)
Live comments possibility?
I am not familiar with discussions related to the questions about Wikidata, but it is strange that when I will do any edit in the Wikidata then I cannot comment it. Any reason for this?--Estopedist1 (talk) 08:05, 23 July 2018 (UTC)
Hello, you will get some information about edit summary can be found here Help:FAQ#Editing #3. Edit summaries are sometimes really useful, but mostly we do not have the option (other than project pages etc). Item/property talk page discussions are not also a common on Wikidata. --Titodutta (talk) 09:21, 23 July 2018 (UTC)
How to describe a heat treatment when used as a qualifier for a mechanical property?
Dear,
We need some advice from Wikidata users who have some experience with properties and a basic understanding of heat treatments :
I've had to revert almost all of the recent edits by Madhumitha Velampalli because they just don't make any sense (e.g. statements like religion or worldview (P140) → India (Q668)). I'm really confused by all of this – are the Telugu labels for items terrible, is there some classification system I'm not aware of, is there something that prevents item descriptions from showing up somewhere, or is she just bad at editing Wikidata? Jc86035 (talk) 12:41, 22 July 2018 (UTC)
@Jc86035: In the absence of such a list at the moment, a quick proxy list of such users may be found by going to the page Special:Search/insource:"{{#babel:hi-2|te-N|en-4}}" (the quotes are part of the search term), which may or may not include false positives. Mahir256 (talk) 14:07, 24 July 2018 (UTC)
Presumably we should have a distinct item for facepainting as against body painting (Q620505), as much as anything because we have a pretty massive Commons:Category:Facepainting (thousands of photos including those in subcats). Right now that Commons category is linked from facepainting (Q2783428), to which I've ventured to add the English label "facepainting": this may be the item we need. Its only sitelinked Wikipedia article is nl:Schmink, which seems to be this same topic, though it might be intended to be narrower in scope; my Dutch is pretty weak (basically, I understand some because it is not that far from German), and the article is pretty stubby. Would people agree this is the correct item?
Is it normal to have an item sitelinked to a redirect like this? If so, shouldn't it also be sitelinked to the analogous en:Facepainting, and probably others?
Rather than change the English name of facepainting (Q2783428) to a word that doesn't exist in English, shouldn't we keep the name and qualify it in the description? And what distinguishes schmink from facepainting in general? - Jmabel (talk) 16:37, 24 July 2018 (UTC)
The two differ in the sense that Suffragette's were members of the Women's Social and Political Union, and were the more militant, terrorist end of the Suffragist movement. So we have suffragist (Q27532437) as an occupation, and suffragette (Q322170) as member of a specific group. If you bother to read the history, you'll find that there were yawning chasms between Suffragettes and the wider suffrage movement. --Tagishsimon (talk) 02:48, 24 July 2018 (UTC)
I still disagree, a "suffragette" is singular so it cannot be a "group of humans" which is plural. Your argument above is about "Suffragette" with a capital "S" as if it is the proper name of a group like "Communist Party" vs. "communist party". And again you are attacking me again personally in the previous section for bringing up this issue here. Please stop. No one should be derided for posting a question here, be polite and just make cogent arguments, or better yet, let someone else give an opinion that can answer politely. --RAN (talk) 04:24, 24 July 2018 (UTC)
Then the description needs to be changed to "member of the Women's Social and Political Union", if others agree. --RAN (talk) 04:28, 24 July 2018 (UTC)
What the hell are you doing now, RAN? Look at this diff. Explain why you are adding nonsense into a message signed by me.
I'm happy to agree that the P31 for Suffragettes may be suboptimal. But that's a new issue. The issue you raised was the merge. And on that score, "Your argument above is about "Suffragette" with a capital "S" as if it is the proper name of a group like "Communist Party" vs. "communist party" THAT'S THE ENTIRE FRICKIN POINT. THEY ARE AN IDENTIFIABLE GROUP WITH AN IDENTIFIABLE CAPITAL-S NAME and they differ from suffragist-the-occupation. I'm exasperated because the difference between Suffragettes and Suffragists is a very basic, very well known aspect of suffrage history; yet here you are determined not to understand it.
Please do not merge these two as they are separate items. And kindly use different headers allowing other users to follow the treads. Pmt (talk) 11:31, 24 July 2018 (UTC)
If you want me to be error free, you must be too: "female member of the women's right to vote movement - the term refers in particular to members of the British Women's Social and Political Union (WSPU)who engaged in direct action and civil disobedience" please add the needed space. And aren't descriptions supposed to be as terse as possible? Isn't there a more concise way to distinguish one from the other without writing a dictionary definition. --RAN (talk) 12:59, 24 July 2018 (UTC)
It doesn't actually work that way, RAN. We all of us make errors. That should not stop us expressing concern about users who repeatedly display a lack of understanding of basic linked data concepts. As for the definition, you'll appreciate given the scope for confusion evidenced in this thread, that I sought to make it idiot proof. But by all means, make it more concise if you wish & are able. I'm sorry that you didn't see fit to comment on your amendment of my comments, nor to revert it, nor to apologise. That seems like that bad form you were speaking about earlier. Perhaps it only applies to other people. Meanwhile I'd be more than grateful if you would respect the temporal order of threads, which generally means adding your comments at the bottom; and if not, then honouring and using indentation levels to make clear what's going on. --Tagishsimon (talk) 15:45, 24 July 2018 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. RAN
Capital of prefecture
Hi,
The item capital of prefecture (Q17221353) has only one site-link to a disambiguation page in the Japanese Wikipedia, which in turn points to articles about capitals of prefectures in several countries.
As a value, however, it appears to be used only for capitals of prefectures of Japan.
This situation is confusing.
What seems sensible to me is the following:
Remove the descriptions in all the languages, which describe it as a disambiguation page.
Remove the sitelink.
Add descriptions that say that it's a capital of prefecture in Japan.
In February, @本日晴天: incorrectly applied another P31 instead of moving the sitelink. As items shouldn't be repurposed and this has gotten a bit of a mess, it would probably be good to create three new items and update links/sitelinks. Once done, Q17221353 can be deleted. --- Jura14:32, 24 July 2018 (UTC)
name in native language (P1559) and Lakota
Been poking on data around quite a bit today and tried adding this to Hollow Horn Bear (Q5881745) but it doesn't seem the Lakota language is supported. Suggestions? GMGtalk19:41, 20 July 2018 (UTC)
By the length of that ticket, I take this to mean "all hope is lost, move on to something else for the foreseeable future." GMGtalk21:33, 20 July 2018 (UTC)
Can someone at Wikimania raise this issue in the context of decolonization and inclusion? It's hard to be sensitive to the concerns of marginalized peoples if we can't name things in their languages. - PKM (talk) 23:05, 20 July 2018 (UTC)
It works if you enter the language code (lkt) even though it doesn't appear in the list, e.g. I was able to enter Special:Diff/713841407 just now. The list is based on the interface languages (and Lakota isn't included as an interface language because the Lakota translation at [25] is still too incomplete). A number of other language codes have been added for use in monolingual text statements, but the list doesn't yet support displaying them - that's what the ticket linked above is about. - Nikki (talk) 17:15, 21 July 2018 (UTC)
Hmm...that's interesting. Thanks Nikki. So if there is a language that can't be searched for in English, I need to look up the page for the language itself and use the three letter abbreviation rather than the English name for the language. GMGtalk12:56, 25 July 2018 (UTC)
Tags as identifiers or strings?
Wikidata:Property proposal/Gfycat tag is stalled, like other proposals before it, because of a disagreement over whether to represent tags using the string or external-identifier datatype. We currently have properties using both types, so there is no single precedent. We need to resolve this issue once and for all, rather than rehashing it with every such proposal. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits12:05, 22 July 2018 (UTC)
My vote goes for no tags at all, just don't add them to Wikidata items, neither as strings nor as external identifiers. − Pintoch (talk) 14:58, 25 July 2018 (UTC)
What do we do when a Wikipedia article has been redirected
i would note that wikidata tends to keep articles that were moved to draft space and deleted. you might want to keep items that english for whatever reason has objectified, and merged with or without a consensus, and might still be notable for wikidata. Slowking4 (talk) 21:52, 24 July 2018 (UTC)
We were trying to work out a way to link to items in Wikidata at the English Wikipedia that only appear in lists, but it went nowhere. Maybe the answer is to create a field in Wikidata "appears in Wikimedia list" so at least there is a connection in one direction. What do you think? --RAN (talk) 15:41, 25 July 2018 (UTC)
@Tagishsimon: The page gives a fair amount of information about the Roundup Riders of the Rockies, namely all the titles authored/published by the organisation. I don't think P973 is restricted to "description in prose". Jheald (talk) 16:01, 24 July 2018 (UTC)
Hmm. Per the 973 talk page, "This is to be used to provide links to reliable external resources that are not the item's official website, when no relevant "authority control" property exists". So your advice was in fact absolutely correct, and I learn something new. --Tagishsimon (talk) 16:10, 24 July 2018 (UTC)
So we have two items, one for the article and one for the retraction note? Well, then we need here an item for the retraction also. And I think we need some property to include links to retraction, comments or other reactions to an article. Steak (talk) 15:10, 25 July 2018 (UTC)
Wikidata weekly summary #322
Here's your quick overview of what has been happening around Wikidata over the last week.
Hi, I'm adding Slovak descriptions to Wikidata items and I sometimes find vandalised descriptions, which are made by non registered users (IP address). I think, that it's really important to delete and change this descriptions, because they are also shown on Wikipedia mobile web view and in mobile app. But I find them just accidentaly, for example one item had bad description for half year, so many people saw it.
So, I would like to ask if there's any way to see recent changes just for Slovak descriptions (adding and changing) from non registered users. I was looking in filters in recent changes, but I didn't find anything for this purpose.
Thanks for answer
--Luky001 (talk) 12:59, 25 July 2018 (UTC)
It is the case of 2 of these 3 projects, and the proposition is very important for those projects. It seems rather bad that this way of notifying participants does not work on the more successfull projects.
Is there a way to circumvent this, other than copying the whole list every time and pinging them individually, especially when the user who pings is a member of the project ? --Hsarrazin (talk) 15:48, 25 July 2018 (UTC)
How about filling a phrabrictor ticket for creating the ability to ping >50? While we are at it, it would be good if the notification text would specify that the notification was created via ping project. This also wouldn't be a Wikidata specific issue. ChristianKl ❪✉❫ 15:58, 25 July 2018 (UTC)
Now that I've been reminded that "ping project" doesn't work for large projects ...
There's a recent discussion of possible revised models for books posted at the WD Books WikiProject. The proposals are by Snipre and by me based on Snipre's proposal and the feedback it has received so far. This has been an open issue without clear consensus for some time, and it would be great if more people could join the discussion on these new proposals. - PKM (talk) 18:25, 25 July 2018 (UTC)
Why would those be merged? One is a disambiguation page, the other is a list-article. Those should have separate data items. --EncycloPetey (talk) 14:57, 26 July 2018 (UTC)
Is there a demonym (P1549) property for nouns, or is the current property supposed to be for adjectives only? (There's a question at Talk:Q159.) According to the Wikipedia articles in English and German a demonym should be a noun (e.g. Briton@en, Québécois(e/es)@fr, Deutsche(r/n)@de), but the property descriptions indicate that adjectives (e.g. British@en, québécois(e/es)@fr, deutsch(e/er/em/en)@de) are preferred. Jc86035 (talk) 08:19, 29 July 2018 (UTC)
Never mind; this isn't relevant because the property doesn't require this level of technical correctness. Lexemes would probably help with this. Jc86035 (talk) 04:34, 31 July 2018 (UTC)
What, then, will you use as the P31 for a person who works with cows, in a country which has no ranches? And given, as you say, that ostrich ranches exist, why would you wish to merge rancher with a breed-specific item such as cowman. It makes no sense. Finally, did we just decide to give up on gender inclusivity ... cowman, cattleman. --Tagishsimon (talk) 01:52, 24 July 2018 (UTC)
Let's take the opening paragraph of w:en:Ranch "A ranch is an area of land, including various structures, given primarily to the practice of ranching, the practice of raising grazing livestock such as cattle or sheep for meat or wool. The word most often applies to livestock-raising operations in Mexico, the Western United States and Canada, though there are ranches in other areas. People who own or operate a ranch are called ranchers, cattlemen, or stockgrowers. Ranching is also a method used to raise less common livestock such as elk, American bison or even ostrich, emu, and alpaca.". So. Ranches. A term most often applied in Mexico, the Western United States and Canada, and other than amongst the older population of the UK familiar with Dallas's Southfork Ranch, pretty much unknown in the UK. Next, ranches: places where one migh raise cattle, but might also raise elk, American bison or even ostrich, emu, and alpaca. Why then would you want to merge cowman (e.g. Joe Grundy with Rancher (e.g. w:en:Montie Ritchie. --Tagishsimon (talk) 02:41, 24 July 2018 (UTC)
FFS, @Richard Arthur Norton (1958- ):, you really cannot do this sort of thing. You have changed Cowman to Cattle Rancher, and the description from 'person who works specifically with cattle' to 'person who owns a cattle ranch'. Do you not see the huge gulf between, say, a low-paid & muddy farmhand and a ranch owner? Do you not see that 'rancher' is a region specific concept, whereas cow-person is a general term. Reverted. --Tagishsimon (talk) 03:20, 24 July 2018 (UTC)
@Tagishsimon: You removed "cattleman" and restored "cowman", ranchers raise bulls and cows. Dairy farmers raise cows, the female bovine. I have restored it. --RAN (talk) 03:24, 24 July 2018 (UTC)
I live on a dairy farm. We raise cows & bulls, as do many dairy farms. Where do you think the little cow thingies come from? That said, I don't mind the change from cowman to cattleman - gender issue aside - since this is not the sweeping change of meaning which was the kernal of my revert - as well you know. --Tagishsimon (talk) 03:33, 24 July 2018 (UTC)
Please try and be civil, it isn't hard to do, and most people here manage to disagree and still be civil. There is no need to write "I worry about your understanding of basic wikidata concepts." Resorting to personal attacks so early in a conversation is just bad form all around. You can disagree how best rancher be broken down into the different types of animal ranchers, but tone down the personal attacks please. We are all volunteers. --RAN (talk) 03:47, 24 July 2018 (UTC)
I do worry about your understanding of basic wikidata concepts, RAN. Specifying that a Rancher 'has part' Cowman is evidence that you do not understand the use of has part(s) (P527), and inferentially that you do not understand subclass of (P279). Pointing it out in the edit summary of the revert of that error is not uncivil. A bigger concern is that you seem happy to make semantic changes - changes of meaning - with no concern that associations which meshed with the old meaning are broken once the new meaning is imposed. That's a very quick and powerful way to do profound damage to wikidata, and again evidence of a lack of understanding. A further concern is in your merge suggestion below, where you snark about "postal worker" and "postal workers" without appreciating that the properties of the two items you're discussing point to them being entirely different concepts - even absent a knowledge of the historical area in which you're proposing a merge. And then, finally for this evening, two new threads in a row from you on this page with exactly the same heading, as if that's of help to anyone relying on edit histories to understand which thread is being edited. May I suggest, when in a hole... --Tagishsimon (talk) 03:52, 24 July 2018 (UTC)
Even "You can disagree how best rancher be broken down into the different types of animal ranchers" suggests you labour under the misapprehension that there's some room for debate w.r.t. P279 versus P527. There is not. --Tagishsimon (talk) 04:20, 24 July 2018 (UTC)
If there's a difference between cowman and cattleman renaming cowman into cattleman produces damage as it removes knowledge. Arguing that the difference exists is an argument against renaming.
Given that a cattleman clearly herds catttle and a rancher might also herd different animals, it seems to me that the terms have a separate meaning and there would be information lost if we merge them.ChristianKl ❪✉❫ 15:47, 25 July 2018 (UTC)
I changed it from cowman to cattleman the first day of discussion, and the description was altered for rancher to show that they can raise livestock other than cattle. Sorry for the confusion, things moved quickly. And thank your for being polite. How do you close a thread? --RAN (talk) 23:02, 25 July 2018 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. RAN
@Pmt: Yes, the latter is definitely to be preferred. Then one can end up with a map like this: tinyurl.com/y9rjwerj. Jheald (talk) 08:52, 25 July 2018 (UTC)
I am not sure where the discussion left off on patronymics. We now have Property:P5056 where we add in data like Jóhannesson (Q25935009) for Icelandic names. Norwegian and Swedish patronymics became surnames around 1900 and were adopted by families that migrated to the US earlier. Is the plan to create an entry for "Hanson (patronymic)" to distinguish it from "Hanson (family name)" or is the plan to just add "Hanson patronymic" as an also_known_as at the entry "Hanson (family name)"? You can see what that would look like here: Hanson (Q9145324). --RAN (talk) 23:13, 25 July 2018 (UTC)
When I visit the item Noteroclada (Q7062574), all the reference statement and most of the section labels at the bottom of the page are in Ukrainian. I get English labels at the top, but Ukrainian for the references and labels at the bottom. If I visit older versions of the page in the Edit history, everything is in English (my language preference), but with the latest edit, now much of the page displays in Ukrainian.
Temple De Hirsch Sinai (Q7698513) is a weird amalgam of data that applies only to the mostly demolished old Temple De Hirsch (built 1908), clearly the subject of an NRHP listing in 1984, the present-day merged congregation Temple De Hirsch Sinai (the inception date given here is 1971); it is described as an instance of a synagogue rather than of a congregation, though, and the congregation has two present-day synagogues (one in Seattle, one in nearby Bellevue). There is no separate item for either precursor congregation (Temple De Hirsch, founded 1899, and Temple Sinai, founded 1961). I'm guessing there should be separate items for at least:
The 1908 synagogue
Each of the two extant synagogues
And I just found out that the current Bellevue campus isn't the original Temple Sinai in Bellevue.
The present-day congregation, embracing all of these, plus various schools etc.
Trying to move ahead with this. I've pulled out Temple De Hirsch (Q55540966) for the 1908 synagogue, and will work on the two extant synagogues [and the earlier Bellevue synagogue]. However, from what I can see, we don't currently have a class suitable to a Jewish congregation. This seems remarkable after so many years, so perhaps I am missing something. We have church congregation (Q2638480) that is specifically Christian; congregation (Q2135977), which might be OK judging by the linked Spanish-language article, but some of the other linked articles (e.g. nl:Gemeente (kerk) imply a specifically Protestant usage; also, the current description for congregation (Q2135977) doesn't seem to match the linked articles. Do we have a more general item (perhaps under some other name) that would include a Jewish congregation? - Jmabel (talk) 00:09, 15 July 2018 (UTC)
"Congregation" and Christianity
Since I've gone 5 days without response, I'm asking again:
From what I can see, we don't currently have a class suitable to a Jewish congregation. This seems remarkable after so many years, so perhaps I am missing something. We have church congregation (Q2638480) that is specifically Christian; congregation (Q2135977), which might be OK judging by the linked Spanish-language article, but some of the other linked articles (e.g. nl:Gemeente (kerk) imply a specifically Protestant usage; also, the current description for congregation (Q2135977) doesn't seem to match the linked articles. Do we have a more general item (perhaps under some other name) that would include a Jewish congregation? - Jmabel (talk) 16:36, 20 July 2018 (UTC)
The articles for congregation (Q2135977) isn't about the same. On dawiki and nowiki it is a congregation for any religion. On nnwiki it is a Christian congregation (I think). On eowiki and svwiki the articles explain different meanings of a word. --Dipsacus fullonum (talk) 17:20, 20 July 2018 (UTC)
No objection to creation of a new item "Jewish congregation". But I suppose the thing is that a lot of the time (here in London, at least) one tends to think of the congregation pretty synonymously with the synagogue that is its focus.
Staying with London, we do have items for quite a few synagogues, tinyurl.com/y8kgc4ad, inherited either from Wikipedia or from the heritage building lists; though a long way off 100%, with even several of the 'flagship' ones missing. And the items do seem to be fairly architectural-only, without even eg movement affiliation linked, or notable rabbis that may have been associated with them.
Probably, simply, nobody's ever done much work on this area.
Here, too, when a congregation is 1-1 with a synagogue, we tend to think of them as one thing. Similarly when a museum has had the same building throughout its existence, especially if that building was purpose-built. But when they are not 1-1, obviously that fails in terms of modeling. - Jmabel (talk) 03:33, 23 July 2018 (UTC)
I don't particularly want to create a special "Jewish congregation" class. What I suggest instead is, since church congregation (Q2638480) is already specifically Christian and congregation (Q2135977) is not, those sitelinks from the latter that are specifically about Christian congregations should be moved to the former, and the latter should be unambiguously generic rather than Christian. Are there any objections to that? - Jmabel (talk) 04:18, 23 July 2018 (UTC)
congregation (Q1365916) (as in the "congregations" of the Roman Curia) strikes me as only etymologically related, and quite distinct in meaning, to the point where even though it is mentioned in some of those same articles, I think we should have
Tricky call as to what is the better value for religion or worldview (P140). Reform Judaism (Q1133485) is not wrong -- in the UK both Liberal and Reform would identify themselves as strands of "progressive Judaism", which is one of the aliases that Q1133485 covers as an umbrella term. And rabbis seem to move quite easily between Reform and Liberal pulpits and vice-versa. affiliation (P1416) is also available.
Did not look at the issue and don’t plan to do this today, but any idea why I’ve recieved 5 notifications for this thread? weird. author TomT0m / talk page18:23, 24 July 2018 (UTC)
@TomT0m: Sorry, entirely my fault. I kept trying to get the {{Ping project}} to format correctly, and failing, so removing it and trying again. Eventually the issue I found was that it really doesn't like being indented -- it works fine, so long as there is no initial :. But (as Multichill is apt to remind me from time to time), there is a Preview button, and I should use it. My apologies for all the excess pinging. Jheald (talk) 18:39, 24 July 2018 (UTC)
@Nikki: OK, I will add the "different from" to each other and the "adjacent locations with the same name" so they will not be confused again. Can you check that I added the Cebu version to the correct one. --RAN (talk) 21:03, 26 July 2018 (UTC)
The cebwiki link is correct. They're not adjacent locations though, they're different types of location, and one is almost entirely within the other. - Nikki (talk) 21:59, 26 July 2018 (UTC)
I could not find any explanation and I'm still seeing two options to add range into quantity datatype. First is to use lower/upper bound option within the datatype (which I think was designed for uncertainties rather than ranges), and the the second is to add two statements with criterion used (P1013) = maximum (Q21067467)/minimum (Q21067468). The second option works if there is only one range to add. I'm not quite sure what should I do if I have to add two or more ranges of the same property, e.g. in three different source there are three slightly different values (ranges) of boiling point (P2102)? I can add six statements to an item, but how do I match proper maximum (Q21067467) with corresponding minimum (Q21067468)? Are there any established guidelines for this? Wostr (talk) 01:47, 25 July 2018 (UTC)
Hmm, I didn't think of that (I fought that these qualifiers can be used only with item datatype properties); it's not ideal, but may be helpful. Wostr (talk) 16:26, 27 July 2018 (UTC)
What's more, it's completely impossible to add range to a qualifier, e.g. if property has qualifier 'temperature' or 'pressure' which is a range of values. Wostr (talk) 15:44, 25 July 2018 (UTC)
I'd also like native support of range in quantity datatape, which does not depend on the central value. --Okkn (talk) 09:24, 26 July 2018 (UTC)
Ranges for numerical values?
Still fiddling around trying to update topics I've heavily contributed to in the past, but I'm not seeing how to include ranges for numerical figures like number of deaths (P1120), number of injured (P1339), or number of casualties (P1590). The topics I'm looking at are generally fairly old (100+ years), and in many cases there is no single figure for these. For example, on Scranton General Strike (Q24060790), there simply is no single figure for injuries. The only figure is between 16 and 54 if you take the high and low estimates of the sources. GMGtalk16:37, 26 July 2018 (UTC)
Of course if you have different sources that say different things, those should be entered as two separate claims, for example one with the low value (referencing the source) and one with the high value. ArthurPSmith (talk) 18:13, 26 July 2018 (UTC)
So...bear with me...(I've mostly just added images to WD until recently) but when I read the guidance for citing books at Help:Sources, it tells me to create an item for the book so I can reference it. But when I read the guidance at WD:N it tells me that I should only create items that have at least one page on a non-WD project. So...what am I supposed to do if I'm citing a source that's not in the public domain, and so not suitable for Commons or Source, but also doesn't meet notability criteria for Wikipedia or Quote? GMGtalk21:09, 26 July 2018 (UTC)
@GreenMeansGo: The second or third criteria would apply in this case: there is no requirement that an item has to satisfy all criteria on WD:N, just that it has to satisfy at least one. In your case whatever books you're citing would most likely be notable under the second, provided you give enough identifying information (ISBNs, links elsewhere, etc.) about them. Mahir256 (talk) 21:13, 26 July 2018 (UTC)
Ooooooooooh. So the existence of an OCLC or ISBN makes it auto-notable for WD. Hmm. Ok. Thanks. GMGtalk21:17, 26 July 2018 (UTC)
Wait, does that feed back into Wikipedia and Quote? Can I cite a source there using only a template and the information added to WD? That would be helpful. GMGtalk21:21, 26 July 2018 (UTC)
Well...maybe...unless the template could be tweaked so that it could be subst and produces the local citation template itself, rather than a persistent link to WD, with the WD ID as a hidden (or even visible to the reader) identifier. So that the citation that's rendered in wikitext is the citation at the time that I make it (so I can verify it). So instead of duplicating a citation template in Wikipedia and then Quote (which Quote is pretty bad at supporting inter-functionality with) I make one citation in WD, and then subst it on the WP page and the WQ page. In wikitext, the generation of the citation template could include hidden text saying "to refresh this template based on updated WD information, simply re-invoke {{template-thing}}". GMGtalk22:38, 26 July 2018 (UTC)
A book published by a reputable publisher satisfies 2; it describes itself and it is a serious publicly available reference. "Publicly available" does not mean free, nor does it mean online. A book available in libraries or available for purchase at places that sell new or used books is publicly available. The book also satisfies notability criterion number 3: it satisfies a structural need. Jc3s5h (talk) 10:08, 27 July 2018 (UTC)
Feature request: auto corrector for ID's - links
When introducing a "MusicBrainz artist ID", if the user introduces the full MusicBrainz URL, a short JavaScript can remove the unnecessary part and leave only the ID from the URL. To give an example, it can transform "https://musicbrainz.org/artist/a3cb23fc-acd3-4ce0-8f36-1e5aa6a18432" into "a3cb23fc-acd3-4ce0-8f36-1e5aa6a18432". The same for other websites like for example allmusic - transforming "https://www.allmusic.com/artist/mn0000219203" into "mn0000219203". I think it would be very nice to have such a feature, it would make things go much faster, when you add a lot of such ID's.
MusicBrainz.org already has this feature. If you add the link "https://www.discogs.com/es/artist/6520-U2", it will automatically transform it into "https://www.discogs.com/es/artist/6520", removing the unnecessary part ("-U2" in this case).
Also, when the user is introducing the full URL, the script can automatically detect what kind of property it is, so the user won't have to waste time to chose it. For example, if the user introduces "https://itunes.apple.com/us/artist/78500", the script can automatically detect the property value: "iTunes artist ID". Lots of time saved when introducing many such Identifiers. MusicBrainz.org implemented this feature already.
I think something similar is already possible with Template:Autofix. If you look at Property talk:P973 for example, there are various patterns listed where URLs will be replaced with the respective properties. But I've never used it before, so I don't know how those changes happen exactly. --Kam Solusar (talk) 18:06, 26 July 2018 (UTC)
User:ThieolBot added a lot of French labels for e.g. TV episodes before they had been released in France. Meaning he added the English titles as French labels. Now these programs do have French titles for a long time and the French labels should be changed to the correct French titles. But no bot is doing this, since a (wrong) French label is already present. So please somebody do this. An example would be episodes of Glee (https://www.wikidata.org/wiki/Q2270524). For more look at the contributions of ThieolBot. --92.213.8.5822:53, 27 July 2018 (UTC)
@Pigsonthewing: That's kind of nit-picky. It's a theological belief -- the positive belief in the absence of gods and supernatural forces -- and certainly if I am asked to fill out my religion on a form, I'll say "Atheist." - Jmabel (talk) 15:50, 29 July 2018 (UTC)
Hi, I am working with other local users to improve the area around the city of Pistoia, we are mapping many new sites for Wiki Loves Monuments 2018. We would like our work to be as precise and robust as possibile, because in the end we want to show a very good final result to the town councils that offered us the permission to include their heritage sites on the WLM list.
Many of these items in facts will have a Wiki Loves Monuments ID (P2186) in few weeks or days but we don't know it yet (we are not the manager, just local volunteers) so I am trying to find better IDs to add now, although it's a matter of weeks before the WLM ID is chosen.
I have therefore started to add manually the GeoNames ID. This insertion is also a good revision step, so I am not complaining do it manually. However, it is quite time consuming, so I would like to know if there is some tool to speed up the process in future.--Alexmar983 (talk) 00:05, 28 July 2018 (UTC)
I am also discovering many items about parks based on Geonames and created on cebwiki that look quite poor. We probably need some query to revise them, I am asking that as well. But every tool is welcome to add these ID.--Alexmar983 (talk) 13:23, 28 July 2018 (UTC)
@Alexmar983: here's a report on items that are parks, which have a ceb-wiki link, and a geonameID. I'm afraid I'm no help on geoname tools, but ping me if you want any more reports. ceb-wiki is notorious for questionable quality items, I'm afraid. Any headway we can make into them is very welcome - as is the wider work you've taken on; thank you. --Tagishsimon (talk) 14:09, 28 July 2018 (UTC)
Oh dear. My next contribution is a lot less colourful: a report currently constrained to country=Italy, showing located in administrative territorial entity where it exists. --Tagishsimon (talk) 14:37, 28 July 2018 (UTC)
Tagishsimon we will fix those, for the rest of the world, people should strat to train new users... "I" have an army of volunteers now, it took time to create it but it's quite useful :D--Alexmar983 (talk) 14:56, 28 July 2018 (UTC)
(marginally related) if you just want to list what's available around Pistoia, you could try this. --- Jura14:33, 28 July 2018 (UTC)
Jura I was thinking to use something similar, but at the moment we have dozens of new items from every municipality and most of the users are refining and creating the item one by one. I am using more these tools to refine all the area around (like WSM. I am trying to avoid the "cathedral in the desert" effect (a.k.a. "white elephant" in English, I guess). So we are leaving a robust network a good items. if someone can help in Property_talk:P646#Searching I have another question about another ID to use.
Just discovered the "create new item" based on image feature at wikishootme. Does it work for you? --- Jura16:36, 28 July 2018 (UTC)
Jura it always worked but I never used it besides probably the first time. In general, I did not invest time in explaining that feature, I even hoped it to be removed because we have tons of georeferenced images on commons imported from external archive with poor descriptions, I was worried to be flooded by dozens of poor items here. My goal was to improve the existing items before creating new ones and I must say, people really are too direct in importing here many new items with little analysis of the surrounding ecosystem. With so much backlog that few people seem to care about, it's quite rare in my experience to see the users who I know (and who sometimes I have trained) using it. We have fixed and seen so many poor items over the years that we know it's not just a click, you have to refine it and it's more work in the end.--Alexmar983 (talk) 19:17, 28 July 2018 (UTC)
Yeah, it could open the new item in a separate tab to make it easier to add more, especially P31. Still, it starts items with more than the usual number of statements. Obviously, many images are not a great start. I'm not entirely convinced that improving cebwiki-geonames items that might possibly be about the same is preferable over creating new items with better references. It could help to go through Commons categories and create items for many local topics already covered there. --- Jura21:48, 28 July 2018 (UTC)
Yes, but in recent years coverage of commons categories is getting maybe worse than wikidata items. That's because commons is not investing in metadata as IMHO it should, so the standardization and automation of wikidata is improving compared to commons "manual" output. I found abandoned areas both here and on commons, but more on commons nowadays. On wikidata there tons of bad poor items, but they are quite fresh... "freshly abandoned" at least. So you should do what you say Jura, there are many things to do and scrolling commons category to create related wikidata items when geographical IDs exist is a very good and necessary thing, but in Italy is not the core problem... the core problem are missing categories on commons, and I am creating them because of new wikidata items for example in these days. More in general, it-N users are more and more inclined to the wikidata structure of metadata (we are a community with a strong "middle class" of wikidata users), and they start to find commons architecture a little bit prehistoric and slow. In the end, it's just that we are kinda introducing structured data on commons via wikidata, not on the file but at least on the categories. This way many established commons users don't get upset because they don't see it happening there...--Alexmar983 (talk) 22:39, 28 July 2018 (UTC)
Jura of course it helps, everything helps, whatever you do, i don't have to do so... I have not contacted wikivoyage yet but I will, they are interested on the "banner" images, the elongated images they use for the top of their ns0 page, they can probably take care of this aspect.--Alexmar983 (talk) 23:59, 28 July 2018 (UTC)
I found it mainly to be a source for coordinates. As they don't have that many active contributors, I'm not sure if they can provide with much support. In some languages, they did start using information from Wikidata items. --- Jura00:20, 29 July 2018 (UTC)
Mr Ademola Adeleke, a Nigerian senator did not finish his education at Jacksonville University. He does call himself doctor. My question: how do you indicate that he studied at a university but did not finish? Thanks, GerardM (talk) 06:14, 29 July 2018 (UTC)
How do we note that a Navy person served aboard a particular ship? We have a field if they were the commander, but I do not see how it is handled for other ranks. --RAN (talk) 17:42, 27 July 2018 (UTC)
I agree - "person was associated with X thing" is the sort of thing where it's easy to swamp the target item with inverse links. Seems simplest just to link from the person to the item and not worry about the other way around, unless their involvement was particularly critical (eg a small crew or a permanent association). Andrew Gray (talk) 21:42, 28 July 2018 (UTC)
I always think it is best for inverse properties to be handled by a bot, humans make too many mistakes. Think of mother, father, children, and sibling properties. President Tyler had 15 children with two wives, the permutations are mind boggling. --RAN (talk) 06:05, 30 July 2018 (UTC)
Good database design would try to avoid making duplicates at all. It means you can get different results depending on how you write a query. Ghouston (talk) 07:13, 30 July 2018 (UTC)
I'm reminded of this w:en:Bill Bruford quote: "In Yes, there was an endless debate about 'should it be F natural in the bass with G sharp on top by the organ, or should it be the other way around?'. In King Crimson, almost nothing was said ... you were just supposed to know." --Tagishsimon (talk) 00:06, 29 July 2018 (UTC)
I'm relatively new here, and have at least twice recently been told by presumably more experienced Wikidata users that I'm not yet knowledgeable enough about Wikidata for my thoughts to be taken seriously (Commons:User:Rodrigo.Argenton) or not thorough enough to be up to Wikidata standards (User:Snipre), so I'm continuing to try to not to be overbold without checking first.
As I remarked, there is no stated procedure to close an RFC, so I would not even know what to do.
Putting together those two points, for all I know you have to be an admin here to be allowed to do this, though I take it from your remark here that you, at least, believe that's not the case.
One of the RFCs that I'd like to see closed was my own question, which of course I'd like to see resolved in favor of my proposal. By my count, there are seven supports and one oppose, but not having unanimity, I would think I was out of line to say the issue is decided in my own favor. - Jmabel (talk) 16:06, 29 July 2018 (UTC)
I think we've had admins and non-admins close RfCs before. The big thing of course is neutral evaluation of consensus. I've closed the particular one you reference above. -- Ajraddatz (talk) 01:18, 30 July 2018 (UTC)
I suppose you can close it and if someone strongly feels otherwise it can be reopened. If no one closes, we drown in open RFCs. --RAN (talk) 05:59, 30 July 2018 (UTC)
Many rfc are just closed for being stale. I don't think Q497757 has an outline yet. Maybe my comment at Wikidata:Project_chat#Proposed_property_for_Instagram_tags helps (the actual outcomes there are slightly different though). Everyone is free to have their own view on how to contribute in the best way. This somewhat invalidated if people don't actually contribute or if they do the opposite when they do it themselves. If you feel too hesitant about it, you might not want to do it. --- Jura06:25, 30 July 2018 (UTC)
@Jmabel: Perhaps you could... open an RfC about it? (I'd favour the creation of a policy or guideline instructing uninvolved administrators or other uninvolved experienced users to close stale formal-discussions-which-need-to-be-closed after 30 days, sort of like w:en:Wikipedia:Closing discussions. There seems to be a lot of places where Wikidata just doesn't have essential policies and guidelines; e.g. the English Wikipedia has 62 policy pages but Wikidata has just 14.) Jc86035 (talk) 09:50, 30 July 2018 (UTC)
Some of this seems clear, in terms of current practice: there is no firm rule on who can close an RFC, but it's generally the job of admin or other experienced user; it's OK to close one without conclusion just for being stale (30+ days) even if it hasn't reached a conclusion; and if it was closed without conclusion, it's OK if someone who disagrees re-opens it. But none of this has answered my other question, which I will now ask for the fourth time: How does one close an RFC? Obviously, people are closing them, and there are presumably some reasonably uniform actions they are taking when they do so. Please see Commons:Commons:Deletion_requests#Procedure for the sort of think I'm asking for. - Jmabel (talk) 15:30, 30 July 2018 (UTC)
Self references?
How do you perform a metareference (Q4307154) on wikidata? For instance, if I have a book (Q571) that contains the name of the author, how would I reference that? Likewise, if I have a film (Q11424) that contains a list of the cast members in the credits at the end, how would I create a reference that references the current subject? --U+1F360 (talk) 14:14, 29 July 2018 (UTC)
We add urls when there is no property. "Described at url" is correct, unless we have blacklisted Wikia, and since we have it as a property, I can't image that it is blacklisted. Should we revert Special:Contributions/Nikkimaria's deletions and give her another warning about seeking consensus before mass deletions, she has been warned previously about deleting links to Familypedia and again for deleting links to Findagrave. She has a habit of deleting links to sources that she says are unreliable without giving statistical evidence. She uses "unreliable" and "crowd sourced" as synonyms. --RAN (talk) 01:45, 30 July 2018 (UTC)
Where is the consensus for mass addition of such links in "described at URL"? I suggest it would be appropriate to get a clearer understanding of the community view of what exactly this property should be used for to avoid such concerns, as was raised when the property was first proposed. Is it intended to be used for every single URL that describes the item? Only those that add value to the page? Some other case? Nikkimaria (talk) 02:33, 30 July 2018 (UTC)
When we are overwhelmed with urls we can decide then what needs to be trimmed, we are not at that point. Since we have Wikia Fandom wikis as a property, I see no reason to not have the subpages of those wikis as links for episodes and characters. --RAN (talk) 05:55, 30 July 2018 (UTC)
To improve the security of our readers and editors, permission handling for CSS/JS pages has changed. (These are pages like MediaWiki:Common.css and MediaWiki:Vector.js which contain code that is executed in the browsers of users of the site.)
A new user group, interface-admin, has been created.
Starting four weeks from now, only members of this group will be able edit CSS/JS pages that they do not own (that is, any page ending with .css or .js that is either in the MediaWiki: namespace or is another user's user subpage).
You can learn more about the motivation behind the change here.
Please add users who need to edit CSS/JS to the new group (this can be done the same way new administrators are added, by stewards or local bureaucrats).
This is a dangerous permission; a malicious user or a hacker taking over the account of a careless interface-admin can abuse it in far worse ways than admin permissions could be abused. Please only assign it to users who need it, who are trusted by the community, and who follow common basic password and computer security practices (use strong passwords, do not reuse passwords, use two-factor authentication if possible, do not install software of questionable origin on your machine, use antivirus software if that's a standard thing in your environment).
EditGroups can now revert batches which created items. These items are deleted in the revert batch. This feature is only available to administrators but other users can easily request revert from the tool.
populationWDvsOSM: a new script that compares the population in OSM place nodes and boundary relations with Wikidata
user:Lakokat is back to adminship, congratulations!
Terminator, a tool by Magnus Manske that finds Wikidata items that lack a label, description, or article in a specific language, has now a version 2 (blog post)
Further work on supporting Senses on Lexemes - specifically persistent IDs for Senses and preventing the deletion of the last Gloss of a Sense (phabricator:T198032, phabricator:T198199)
Adjusted colors of the Query Service UI to be more in line with Wikimedia style guide (phabricator:T192235, Thanks abian!)
We'd like to invite you to the weekly Technical Advice IRC meeting. The Technical Advice IRC Meeting (TAIM) is a weekly support event for volunteer developers. Every Wednesday, two full-time developers are available to help you with all your questions about Mediawiki, gadgets, tools and more! This can be anything from "how to get started" and "who would be the best contact for X" to specific questions on your project.
The Technical Advice IRC meeting is every Wednesday 3-4 pm UTC as well as on every first Wednesday of the month 11-12 pm UTC.
If you already know what you would like to discuss or ask, please add your topic to the next meeting: Technical Advice IRC meeting
They are not the same. Some demonstratives are demonstrative pronouns, but not all demonstratives are pronouns. Some demonstratives are adjectives, determiners, or other parts of speech depending upon the language. --EncycloPetey (talk) 21:07, 30 July 2018 (UTC)
Normally, the statements should determine it. It's just that the labels of Q282301 in fr/fi/it/la etc. are different from en. What they link to is yet another problem. Maybe it's better to create a new one and split Q282301 between the new one and Q34793275. Once done, Q282301 can be deleted. --- Jura21:18, 30 July 2018 (UTC)
If Floscher is watching: Soldiers of the Great War (Q55831179) is a three volume work, so if they would please work on that entry, I will have an example of how the data fields should work. volume (P478) for the Encyclopedia Britannica has the value 22, the total number of volumes, but the field looks like it is meant that each volume gets its own Wikidata entry ... I am confused. Thanks all! --RAN (talk) 00:42, 31 July 2018 (UTC)
@Richard Arthur Norton (1958- ): Google Books ID (P675)dswMAAAAYAAJ and Internet Archive ID (P724)SoldiersOfTheGreatWarV1 are both about the first volume, so they should not be assigned to the Wikidata item about all three volumes as a whole. You should create one Wikidata item for each volume and put the respective Google-Books-IDs and Internet-Archive-IDs on those items.
@Floscher:Ok, I see now how "volume" works, an entry for the combined work and an entry for each volume of that work. How is "edition or translation of" handled?
OK, I see you now changed to instance_of=version from instance_of=volume to get rid of the warning. It just seems odd when there was only a single publication of the book and never a reprinting or a translation. It isn't very intuitive. --RAN (talk) 17:53, 31 July 2018 (UTC)
Link to a Wikimedia commons image containing "&"?
I just tried to add an "image" statement consisting of [[File:Bruemmer with M&G posters.jpg|thumb|Bruemmer with M&G posters]]. It would not accept this, saying, 'Could not save due to an error. File names are not allowed to contain characters like colons or slashes. Only paste the file name after "File:", please.'
No. This is not a game of guesswork, KlaasZ4usV. Bruemmer with M&G posters.jpg works fine. Making the change you suggest, to the string the OP was entering, does not work and is, therefore, not good advice. --Tagishsimon (talk) 05:15, 31 July 2018 (UTC)
Thank you. Yes, that did what I wanted, sort of. The [[File:Bruemmer with M&G posters.jpg|thumb|Bruemmer with M&G posters]] syntax used in Wikipedia, for exaample, includes a figure caption.
I do not mean use in reference, but links to articles, is it possible to suggest properties for 10 copies of the same site?! --David (talk) 10:40, 31 July 2018 (UTC)
I just created Daniele Ganser (2017). Les Guerres illégales de l'OTAN (in French and German). Demi-Lune editions. ISBN978-2-917112-39-7. WikidataQ55839635. based on a review of that book by Gabriel Galice in François Honti; Serge Halimi; Ignacio Ramonet; Claude Julien; Benoît Bréville (eds.), Le Monde diplomatique (in French and English), Paris: Le Monde Diplomatique SA, ISSN0026-9395, OCLC473074470, WikidataQ858121 of 2018-07, p. 24. I have three questions regarding this:
What's the preferred way to create a Wikidata entry for this book review? I found book review, WikidataQ637866. When I went to "What links here", I concluded that some of those links were NOT of the preferred form. So I decided to ask.
What's the preferred way to enter the publisher? It's Demi-Lune, Paris. When I tried to enter that, it took "Demi-", and dropped the Lune.
The Mediawiki software parsed {{cite Q|Q858121}} above as, "https://mondediplo.com/ Le Monde diplomatique Check |url= value (help), Paris: Le Monde Diplomatique SA, Wikidata Q858212". The url generating the "Check |url=" message is to the English-language edition. (It's currently produced in 71 editions in 27 languages. 33 of the editions are electronic. Should we be happy that the Wikidata entry on Q858212 only links to the French and English, and not the other 31? ;-)
1. Might I suggest looking at the format of references on statements for, say, Douglas Adams (Q42) to give you an idea of the approach taken here.
Thanks. I've got 5 "Statements" (instance of, title, author, language, and publication date). It seems I should create a separate Wikidata entry for the review, then enter the Q number as a reference for each of those? That seems like more work than should be necessary. ???
@DavidMCEddy: What is the format of your book review ? It is a book or an article in a newspaper ? Answer that question and then check again what is the recommandations on help:Sources. As the book review is another document, this requires a new item, different from the cited book. Snipre (talk) 19:09, 31 July 2018 (UTC)
2. You would need to add a new item for the publisher, before the publisher could be used as a publisher (P123) value in a statement for a book.
Done. (After providing English names for 9 apparently different Demi-Lune lakes, all in Quebec -- then finding the web site for the publisher and finding it was NOT in Paris, even though Le Monde Diplo said it was.) Thanks again. DavidMCEddy (talk)
If you are "Unhappy but unsurprised", I'm fine. I've done my duty to report this unhappy but unsurprising feature of Wikidata. Thanks once more. DavidMCEddy (talk) 07:44, 31 July 2018 (UTC)
Regarding '3' - the version of {{Cite Q}} on this Wiki is, AFAICT, a direct copy of that from English Wikipedia. If you feel it needs additional, or different, functionality, feel free to make a change, or to request one on its talk page. But if you wish to cite a different edition of a book, you should first create a Wikidata item about that edition, then use the QID from that in the template. That said, note that you can specify a local value for |url=, which will override the one transcluded from the item cited. This is explained in the template's documentation. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits13:54, 31 July 2018 (UTC)
Torrent data dumps
Hi, since google couldn't I've came to ask here:
Are there any recent data dumps that can be downloaded via torrent? – The preceding unsigned comment was added byShouldaistay (talk • contribs) at 19. 7. 2018, 20:00 (UTC).
Merged Matěj Suchánek (talk) 07:13, 4 July 2018 (UTC)
WikiProject Informatics has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.
⟩ named after (P138) 
Thank you. that is what I was looking for. --Jarekt (talk) 17:55, 3 July 2018 (UTC)
Support For the shutdown, at least one day. It is pretty serious. I called several MEPs' offices today through this website, and I encourage everyone to do the same, so at least they know that there are people who care about this. It is a free call since you are called to your number and then you are put through to the MEP office. If that website doesn't work for you, you can also try this one by the Mozilla foundation, where you even have a script of what to say.--Micru (talk) 14:48, 3 July 2018 (UTC)
Done
Notified participants of WikiProject Informatics/Software
Strong oppose continue creating items for cebwiki, GZWDer, you're just polluting Wikidata, just stop that work, you're feel free to use your bot on other areas. --Liuxinyu970226 (talk) 10:05, 16 July 2018 (UTC)
Neutral While this matches my second point above (excluding any creation of items that only have Geonames property), I wouldn't support this kind of solution easier, as there are still no potential good sanctions to cebwiki, which is required to prevent more wikis from such behaviors. --Liuxinyu970226 (talk) 14:38, 16 July 2018 (UTC)
based on a review of that book by Gabriel Galice in François Honti; Serge Halimi; Ignacio Ramonet; Claude Julien; Benoît Bréville (eds.), Le Monde diplomatique (in French and English), Paris: Le Monde Diplomatique SA, ISSN 0026-9395, OCLC 473074470, Wikidata Q858121