Wikidata:Property proposal/Computer Science Ontology topic
Jump to navigation
Jump to search
CSO topic id
[edit]Originally proposed at Wikidata:Property proposal/Authority control
Motivation
[edit]An important resource for people working with computer science research, by Open University and SpringerNature.
- Also includes formatter for RDF resource: https://cso.kmi.open.ac.uk/topics/$1.ttl
- Can serve per-entity RDF, JSON, and other formats
- One item will often link to multiple CSO topics (which themselves are linked by relatedEquivalent), see first two examples
- Maybe I'm blind but I couldn't find a previous proposal for this
Vladimir Alexiev (talk) 16:30, 12 December 2022 (UTC)
- I updated the examples to show that CSO topics really are labels. But they are connected through relatedEquivalent, and the main label is designated as preferentialEquivalent. You can see that eg in https://cso.kmi.open.ac.uk/topics/semantic_technologies.ttl
<https://cso.kmi.open.ac.uk/topics/semantic_technologies> ns0:relatedEquivalent <https://cso.kmi.open.ac.uk/topics/semantic_web_technology> , <https://cso.kmi.open.ac.uk/topics/semantic_web_technologies> , <https://cso.kmi.open.ac.uk/topics/semantic_web_applications> , <https://cso.kmi.open.ac.uk/topics/semantic_web> , <https://cso.kmi.open.ac.uk/topics/semantic_technology> ; ns0:preferentialEquivalent <https://cso.kmi.open.ac.uk/topics/semantic_web>
- So: we should link only to preferential topics, and use the others as alt labels. I've tweaked the proposal to show this.
- I also changed the name to "CSO topic id" as per User:SM5POR below Vladimir Alexiev (talk) 14:10, 14 December 2022 (UTC)
- I included "%" in the regex because some topic URLs use percent encoding (eg computer_aided_design_%28cad%29).
- While https://cso.kmi.open.ac.uk/topics/computer_aided_design_(cad).ttl is more readable and resolves, the semantic data inside uses the percent-encoded form of URL. Vladimir Alexiev (talk) 15:29, 14 December 2022 (UTC)
Discussion
[edit]- WikiProject Informatics has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. Notified participants of WikiProject Authority control WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. --Vladimir Alexiev (talk) 16:30, 12 December 2022 (UTC)
- Support I think the CSO user interface is really nice ... I would love such an interface for Wikidata. While I am not that interested in the CSO I think mix-n-matching it with Wikidata should help us find some notable topics for which we currently lack data items. --Push-f (talk) 04:51, 13 December 2022 (UTC)
- Comment Is there a reason for using the word "topic" instead of "ID" at the end of the label, such as the CSO itself using this word for its identifiers? Using "ID" seems to put it more in line with other external identifiers ("topic ID" appears to be used for various news organizations, which the CSO isn't a case of). --SM5POR (talk) 07:55, 13 December 2022 (UTC)
- I called it "topic" because it's text-based (doesn't look like an id). We can add "CSO id" as an alias? Vladimir Alexiev (talk) 14:29, 13 December 2022 (UTC)
- External identifiers are sometimes intuitive text labels rather than cryptic alphanumeric strings, but their properties are still conventionally labelled "ID" to show that they are reference strings, not items. A few exceptions are GitHub topic (P9100) and PhilPapers topic (P3235) though.
- The word "topic" does not relate to the format of the identifier, but rather to its role in the external database, to contrast it with other types of items you can look up in the same database. Even properties for news organizations, where "topic" look-ups are common, are labelled "topic ID", and may still use a cryptic string (like BBC News topic ID (P6200) does, differentiating it from BBC programme ID (P827) or BBC Things ID (P1617)). The description for PLOS Thesaurus ID (P10609) mentions "topics" in the context of scientific publishing, but it uses an alphanumeric identifier too.
- I see that pretty much all other ontology references are alphanumeric identifiers (KBpedia ID (P8408) and Ontology of units of Measure 2.0 unit ID (P8769) are exceptions). Searching for a property label containing "ontology topic" yields no relevant hits, however.
- If you still want to avoid "ID" in the label, I'd suggest "entry" rather than "topic", similar to how Gujarati Vishwakosh entry (P9863), Merriam-Webster online dictionary entry (P11130) and The Britannica Dictionary entry (P11263) use it. Other dictionaries like ODLIS ID (P9374) use "ID" in spite of having (at least partial) plain text identifiers. SM5POR (talk) 09:24, 14 December 2022 (UTC)
- These are definitely topics. So either "CSO topic" or "CSO topic ID". -- Vladimir Alexiev (talk) 14:03, 14 December 2022 (UTC)
- I called it "topic" because it's text-based (doesn't look like an id). We can add "CSO id" as an alias? Vladimir Alexiev (talk) 14:29, 13 December 2022 (UTC)
- Support Boyan-bechev (talk) 15:39, 13 December 2022 (UTC)
- Support Chervenski (talk) 16:17, 13 December 2022 (UTC)
- Support --93.123.21.124 16:20, 13 December 2022 (UTC)
- @Vladimir Alexiev, Push-f, Boyan-bechev, Chervenski, 93.123.21.124: Done--Tinker Bell ★ ♥ 17:52, 21 December 2022 (UTC)