Wikipedia:Reference desk/Archives/Computing/2010 November 29

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

November 29

Add-ons and other issues

If I uninstall an add-on on the beta version of Firefox 4.b7 due to incompatibility issues, will it also be automatically uninstalled in the 3.6.12 version? 24.189.87.160 (talk) 03:59, 29 November 2010 (UTC)

Oh yeah, and while switching in between both versions, my reload button has mysteriously disappeared in 3.6.12. 24.189.87.160 (talk) 05:56, 29 November 2010 (UTC) Nevermind that, it just got relocated, and I couldn't tell it apart from my blue background. 24.189.87.160 (talk) 07:52, 29 November 2010 (UTC)

Are you using the same profile for both versions? If so, then probably yes. However I would have thought as a beta version Firefox 4 should keep its profile separate from your current install version, but it might not. In the future you might like to check out FirefoxPortable which can install alongside any existing Firefox you have on your computer and won't change or interfere with any of your current settings. 82.44.55.25 (talk) 10:34, 29 November 2010 (UTC)

Thanks for the link to portable Firefox. Is there any way I can set up separate profiles for each version of Firefox I have running? 24.189.87.160 (talk) 10:58, 29 November 2010 (UTC)

Questions about wireless carrier technologies (GSM, CDMA...)

Until recently, it used to be that, there were 2 different technologies used for Canada's wireless "Big 3" networks: GSM used by Rogers/Fido and CDMA used by Bell and Telus.

Ever since 3G (and 3.5G 3G+ and whatever other marketing gimmicks they use) I cant keep track of all the new ones.

Is the HSPA (including HSDPA/HSUPA/HSPA+) 3G network used by the Big 3 companies compatible with each other in the sense that unlocked HSPA phones can work on all 3 networks? What about "travelling on other networks" (basically domestic roaming) does that work seemlessly if you have one network's phone and you are in an area only covered by the other network's 3G?

Another question... if you buy a new 3G HSPA phone from Rogers/Fido, would it work on their 2G (GSM) network if you are in an area with 2G coverage only?

Same question for Bell and Telus, are the new 3G HSPA phones capable of using the old CDMA network?

One last question, the two new independent companies, WIND Mobile and Mobilicity, are they both on UMTS networks? Do their phones work with 3G HSPA networks or 2G (GSM or CDMA) networks?

Answer as many of my these as you can please... i'm totally baffled by all of this and whatever u can clarify would be of great help Roberto75780 (talk) 08:22, 29 November 2010 (UTC)

First, this may be inaccurate (I live in Australia) but phones that are not too old (like 2 years) should still have GMS or CDMA capabilities (depending on the carrier). General Rommel (talk) 09:11, 29 November 2010 (UTC)

I work for a wireless carrier here in the U.S. So, although I'm not an expert on the situation there in Canada, I can say that interoperability depends on the phone. The phone has to support the access method (CDMA or GSM) and the frequency used by that carrier. I understand that Bell and Telus use CDMA for 2G and HSPA (which uses GSM) for 3G. I also understand that they use the same frequencies. So, Bell and Telus phones should work on both networks for both data and voice. Their customer-service reps may not help you move your phone between their networks, but you can always do this manually if they refuse by reprogramming your phone on your own. There's generally a code you enter on the phone to enter the programming screen, from where you can enter the SID (system ID) and the MIN manually.

As for Fido, it uses GSM for everything (both 2G and 3G), right? In general, GSM phones are not compatible with CDMA networks and vice-versa. There are some phones that operate on both (like the BlackBerry Storm 2), but they are rare. Here in the U.S., there are two GSM carriers (T-Mobile and AT&T). Although they both use GSM and HSPA, they use different frequencies. Some GSM phones support both frequencies, so they work with both T-Mobile and AT&T. But, from what I understand Telus and Bell in Canada use the same frequencies for both CDMA and GSM, so they are compatible. But Fido does not appear to be compatible with those carriers.

As for WIND Mobile and Mobilicity, the "List of Canadian mobile phone companies" entry says that they use HSPA. HSPA uses UMTS.--Best Dog Ever (talk) 09:39, 29 November 2010 (UTC)

The above reply is correct but there are a few things worth adding. There are plenty of handsets that work between CDMA and GSM for voice (known as "world phones") and among them are most blackberries, and some new Android phones. However, the frequency differences of the 3G technology are where you will find a lot of inconsistencies, and while you may get voice functionality your data speed will be limited to 2g (or 1x or edge or whatever you want to call it.) If you are just worried about voice capabilities, verifying that the phone is CDMA and GSM capable is sufficient (although to actually move between carriers usually requires "unlocking" the handset.) So, while phone shopping be sure to check the specs for the base technology (gsm/cdma/both) and 3g technology (hspa, umts, and frequencies.) If you are set on having the widest amount of compatibility, a Blackberry is probably the best choice. --144.191.148.3 (talk) 16:20, 30 November 2010 (UTC)

Python as a calculator

... generates these results:

>>> 4 * 0.18
0.71999999999999997
>>> 4 * 0.08
0.32000000000000001

Why not 0.72 or 0.32? Is that the same in other programming languages? Quest09 (talk) 09:30, 29 November 2010 (UTC)

See IEEE floating point, which is how most computers represent numbers with fractions.

Basically 0.72 is represented as

1*1/2 + 0*1/4 + 1*1/8 + 1*1/16 + 0*1/32 + 1*1/64 ... or 0.101101...

Likewise, in decimal notation 1/3 can not be exactly represented; for those used to the trinary number system where decimal 1/3 would be represented as trinary 0.1 they might be suprised that our decimal number system can't represent it exactly.

The only reason we use the decimal system is that we have 10 digits on our hands. —Preceding unsigned comment added by Csmiller (talk • contribs) 10:06, 29 November 2010 (UTC)

How do proper calculators avoid giving these binary-biased results? 213.122.68.179 (talk) 13:18, 29 November 2010 (UTC)

Arbitrary-precision arithmetic. You can do this in Python (and other languages) too:

              from decimal import Decimal
              Decimal('4') * Decimal('0.18')
              # gives: Decimal('0.72')

              # but if we try to express that Decimal as a float:
              float(Decimal('0.72'))
              # gives 0.71999999999999997, for the reasons noted above.

-- Finlay McWalter ☻ Talk 13:31, 29 November 2010 (UTC)

You seem to be talking about decimal floating point, which is not the same as arbitrary-precision arithmetic. That's what Python's decimal module implements, and I think it's what most pocket calculators use. -- BenRG (talk) 17:22, 29 November 2010 (UTC)

A further point is that since a calculator stores only a few numerical values and has limited display room, it's easy to design it to store more digits than it displays. So when it multiplies 4×0.18, it might really be calculating 0.71999999999999997 internally and just rounding it to 0.7200000000 for display. In fact I don't think this technique is widely used, but it's possible. You could find out by subtracting numbers whose valuess are close together, like 1/3−0.333333333. --Anonymous, 03:17 UTC, December 1, 2010.

I should note that the "problem" with floating point isn't because it's binary. As Csmiller notes, there's nothing special (especially good or especially bad) about decimal, and you get rounding errors and repeating representations regardless of how many fingers you have. The space used to store, and the time taken to calculate, floating point ops are essentially fixed (they're not related to the numbers in question) which means it's a handy scheme for applications where performance is important, and exact representation and small errors aren't (like controlling robot arms or 3D graphics). Arbitrary-precision arithmetic, which represents numbers as a string of digits (it's how you were taught in school) uses storage and takes time in proportion to the string-sizes of the numbers (see computational complexity of mathematical operations) - it's good when you need accurate representation (e.g. for financial applications) and don't care so much about speed or memory use. But arbitrary precision has issues with irrational numbers (it has to truncate them at some arbitrary position, otherwise it'd never finish). Note incidentally that when the Arbitrary-precision arithmetic article says arbitraries are slow because they have to be "implemented in software", that's abject nonsense: they don't have to be implemented in software (it's a lot more work to do in hardware than fp, but it's quite tractable) and their speed issue is a fundamental result of how they work. -- Finlay McWalter ☻ Talk 14:02, 29 November 2010 (UTC)

What normally happens is the programmer knows of these precision problems, and rounds any floating point number to a few decimal places before reporting them to the user, thus avoiding the large number of decimal places. CS Miller (talk) 14:48, 29 November 2010 (UTC)

The problem that the original poster is talking about is a side effect of using binary instead of decimal. 4 * 0.18 is 0.72 exactly in decimal floating-point. That's why using the decimal module fixes the problem. Decimal floating-point arithmetic is routinely used in the financial industry for this reason. Also, "arbitraries are slow because they have to be implemented in software" isn't abject nonsense; it's true that on 99%+ of computers out there, binary floating point has special hardware support while alternatives must be implemented in software. Sure, arbitrary precision requires extra work in any case, but there's a big slowdown up-front just from the software simulation.

"Arbitrary-precision arithmetic" makes me think of computable real arithmetic, which can retain full precision even with irrational or transcendental numbers. I don't know if there's a computable real package available for Python, though. -- BenRG (talk) 17:22, 29 November 2010 (UTC)

An even easier way of circumventing this issue is converting floating point numbers into a string. Without importing anything, simply try:

 str(4*0.18)

. —Preceding unsigned comment added by Trustinchaos (talk • contribs) 17:00, 29 November 2010 (UTC)

That's just rounding, as has already been mentioned: str() on floats formats them with, I believe, 12 significant digits, which hides such trivial "errors" as these. For example, I get str(0.3/0.1)=='3.0' but repr(0.3/0.1)=='2.9999999999999996'. --Tardis (talk) 20:38, 29 November 2010 (UTC)

Yes, str(n) only formats your floating point numbers like in the case above, but it is expressed more concisely and doesn't let you choose the number of significants digits, which is set to 12 in str(n) and is optional in round(n[, ndigits]).Trustinchaos (talk) 11:44, 30 November 2010 (UTC)

Microphones and Sound cards

Hi, I've toying with the idea of getting a new microphone and sound card, which microphones and sound cards produce a crisp and clear recording of speech, such as File:The Game.ogg? I currently have a Windows Vista laptop with a Realtek High Definition Audio. Thanks --George2001^hi 11:45, 29 November 2010 (UTC)

Is there any reason why you feel you need a new sound card? Nil Einne (talk) 15:51, 29 November 2010 (UTC)

Wikipedia:WikiProject Spoken Wikipedia/Recording guidelines has some guidelines about recording speech; there's a lot more to it than the microphone (and really any sound card should work fine). Wikipedia:WikiProject Spoken Wikipedia/Recording guidelines/Sample configurations lists some of the microphones people have used; very few seem to have anything other than basic hardware. -- Finlay McWalter ☻ Talk 16:26, 29 November 2010 (UTC)

Thanks, for the help. I was just wondering if the sound card was poor or outdated. --George2001^hi 17:27, 29 November 2010 (UTC)

(e/c) Be wary of buying a microphone that requires a lot of auxiliary equipment. To demonstrate what I mean, my experience: I used to use a very cheap, old headset for recordings. Its sound was a little shallow and tinny as a result (made worse by the fact that I have the naturally higher voice of a young woman), but it was otherwise clear and fuzz-free. About two or three years ago, in order to play an older female character in a more professional recording, I was asked to get a serious amateur microphone that would have a richer sound, to help balance the limitation of my voice wrt this character. Without really knowing what I was doing, I bought a decent-ish electret condenser microphone. I then found that it required pre-amplification, so I had to buy a very small pre-amplifier. Also, didn't have the right port for the jack, so I purchased an expensive (for what it was) converter. Once I finally got it all working, the sound was rich and warm, yes, and I sounded better. But, because it wasn't that directional, was sensitive, and not top of the range, I got a lot of fuzzy background noise, picked up from my computer fan. I then had to make a small sound booth, rather like this, to house the microphone. Also had to stack books on a surface before the microphone would reach my standing height. I'd spent way more than I had intended to, had big books and a cumbersome foam-filled box cluttering my desk, and still found it fiddly and bothersome to set up and use. In the end, I scrapped the approach altogether and bought a new, quality headset: I get fine sound (not as warm, but good enough), freedom to record standing or sitting, no unwanted noise (from fans or speakers), and it's USB and plug-and-play. I also don't get variations in the volume just because I happen to shift my body position. Bonus: also perfect for Skype, gaming, etc and easy to stow away. If you're not attempting to set up an actual recording booth at home, then I suggest keeping it simple and multi-functional. Go for a good headset perhaps—there are dozens with great reviews. As for sound cards, the one you have should be fine; probably only an audiophile or a professional would be able to tell the difference between the average sound card and a top-end one. Maedin\^talk 18:04, 29 November 2010 (UTC)

Thank you, now that's going the extra mile - your advice is greatly appreciated and will taking into account. Sorry if I sound bothersome, can I ask what your quality microphone is? Thank you. --George2001^hi 18:36, 29 November 2010 (UTC)

Are you referring to the serious microphone or the headset? The microphone that I gave up on was a Sony, like this, only not quite that expensive, if I recall correctly. As you can tell from the reviews, it's meant more for recording music than speech. The headset I use now is a Logitech, but I couldn't find the model (it's older) on Amazon, but likely comparative to this one: [1]. I imagine that one of the cheaper ones would also do great (depending, of course, on your intended usage), as would a number of other good brands in headsets. A general search for headsets will bring up a lot of decent options. As a very general guide, I wouldn't spend less than £15 or more than £30, but ymmv!  :) Maedin\^talk 19:01, 29 November 2010 (UTC)

Computer modelling of Artificial consciousness

"Computer modelling shows that even consciousness can be generated with very small neural circuits....only a few thousand could be enough to generate consciousness." from http://www.sciencedaily.com/releases/2009/11/091117124009.htm ... 92.24.176.72 (talk) 14:32, 29 November 2010 (UTC)

The above question has been asked on the science desk; just giving any of the denizens here who don't answer the science desk a heads up (how may are there?) CS Miller (talk) 16:19, 29 November 2010 (UTC)

Processor upgrade

Firstly may I state that I know upgrading the processor on my laptop would void the manufacturers warranty and that it can cause compatibility issues. I fully understand that, but I need to check two things:

-A) I have an Acer Aspire 5542 with a AMD M880G motherboard with a AMD Athlon II Dual-core M320 processor. Initially when I bought my laptop this was sufficient but I'm starting to feel the effects of not having a powerful enough processor. I have sufficient RAM for my needs and a decent graphics card, so I'm assuming my poor performance is due to the processor. Is this chip set soldered in to the motherboard and thus unable to be upgraded?

-B) Assuming it's upgradable, where can I find which processors I would be able to upgrade to?

I can't find about either question in the computer manual or through a quick Google search, so I'm seeking advice from you guys. May I also state that I have no intention of doing this myself and it would be done through a technician, but I want to avoid the cost of getting advice from one beforehand. Thanks for your help. Regards, --—Cyclonenim | Chat  17:58, 29 November 2010 (UTC)

A processor is never soldered into a motherboard. The processor models that are compatible with your system are listed at List of AMD Turion microprocessors#Turion II / Turion II Ultra. You should be able to use any of the "Caspian" types -- the ones that use the S1G3 socket. Looie496 (talk) 19:04, 29 November 2010 (UTC)

Any idea where I could purchase a Turion II Ultra M660? And ah right, I just read an article that said sometimes they're fixed to the board? Never mind! This is why I love Wikipedia and the reference desk! Thanks Regards, --—Cyclonenim | Chat  19:06, 29 November 2010 (UTC)

If this is what your motherboard looks like, you can see the empty CPU socket in the image. That is where you'd remove/replace the CPU. As for finding a CPU, there are many places that sell processors. For example, you can try the first hit I got from Google. -- kainaw™ 19:21, 29 November 2010 (UTC)

Does it involve any complicated working with the BIOS or anything though? I consider myself competent with computers but not much more. If I literally just have to swap things over then I'd be willing to that, or maybe if I can find a detailed guide. Regards, --—Cyclonenim | Chat  19:46, 29 November 2010 (UTC)

Firefox and Keyboard Shortcut customization

Mozilla Firefox's official Keyboard Shortcuts guide explicitly states: "Firefox does not provide any method of customizing keyboard shortcuts." This seems like intentional, heavy-handed, and forceful, especially for a free and open-source software package that allows all kinds of other customizations. I suspect this decision has something to do with maintaining "brand consistency" or precluding an entire class of phishing attacks, but I'm looking for an official explanation behind this design decision. Can anyone help me find an official explanation from Mozilla (a developer-blog, program documentation, mailing list, etc.) explaining why they do not allow custom keyboard-shortcuts? (Addendum, I found this background theory on Mozilla key bindings, but I'm really looking more for a high-level, "we chose to do it this way because ..." explanation). Nimur (talk) 18:55, 29 November 2010 (UTC)

I think they chose to do it this way because they generally tend to leave non-essential stuff to extensions and add-on packages. Installing keyconfig takes about a minute.—Emil J. 19:21, 29 November 2010 (UTC)

Yes, and there are plenty of other key configuration plugins for Firefox. Vimperator radically changes the key bindings to use those of the Vim text editor. There are many other such plugins. –Tom Morris (talk) 02:37, 30 November 2010 (UTC)

Office 03, Excel, tilde issue in VLOOKUP

Is the tilde some kind of reserved character in Excel, or does it confuse Excel in some other counterintuitive way?

I have a large spreadsheet of client investment data by account, a direct but seriously ugly dump from one of our reporting systems. In order to more easily make sense of this data, I need to map basic account numbers to a "Holding Type". I've added a proper column, filled it with a VLOOKUP call to an external data table, thusly:

   =VLOOKUP(E2, [XrefTable.xls]Sheet1!$A$2:$E$506, 5, FALSE)

This works wonderfully -- and then feeds into a Pivot Table -- except for one thing:

A dozen or so account numbers contain a "~" as the leading character. No problem, I sez to myself, I should create the XrefTable data with the tildes anyway since they don't ever change. But, ALL of the VLOOKUP calls FAIL and return "#N/A" for any account that begins with that damn tilde.

This is clearly documented in the VLOOKUP help: If VLOOKUP can't find lookup_value, and range_lookup is FALSE, VLOOKUP returns the #N/A value. The XrefTable is sorted on account number, and most of the numbers are strictly numeric, but some aren't, and I have tried both of the "sort anything that looks like a number" options.

So, WHY can't VLOOKUP find a lookup value that starts with a tilde in the first place?? It can't be a known bug in an 8-year-old product, can it? I thought it might treat ~ as some kind of metacharacter and throw off the sort or lookup, but can't find anything in the online help.

Anybody ever run into this issue before, and solve it?

Thanks much, DaHorsesMouth (talk) 20:07, 29 November 2010 (UTC)

• I should probably also mention :-) -- I understand the part about using ~ when typed literally into a search box or the like. In this particular case, since I'm not going to be the only user of the resulting table, I have a strong aversion to editing the raw data. The tilde in question appears as the leading character in two data fields, not as a literal in my function or something I'm entering directly.

Besides, even if ~ says "do not modify the next character", shouldn't it STILL match the two strings properly?

DaHorsesMouth (talk) 21:00, 29 November 2010 (UTC)

MS suggests that you'd normally use a double tilde ("~~") in these cases [2]. As you can't modify the raw data you'll probably have to use an expression in the spreadsheet to modify the search string before handing it to VLOOKUP. Blakk and ekka 13:38, 30 November 2010 (UTC)

Oops, my mistake! I meant .dat file (continued from here)

I just looked at it again, and it's a .dat file, not a .bat. should I do? (if it means anything, it's 35 kB)--72.178.134.134 (talk) 21:20, 29 November 2010 (UTC)

If its a .dat file, then its ussually a serialized data file of some sort. If you don't know what to do with it, you probably should either delete it/or ignore it.Smallman12q (talk) 21:29, 29 November 2010 (UTC)

A .dat file attached to email may be winmail.dat, a container of the actual attachment sometimes sent by Outlook clients. Unilynx (talk) 07:26, 30 November 2010 (UTC)

DVD -/+R

Working on an iMac G5, OSX, version 10.5.8. I want to download some movies and burn them to disk. I've never done it before, though I've used Toast Titanium to burn some pdfs to CDR disks. So I have a couple of questions. I see the DVD+R and DVD-R formats. I looked at the article on them but didn't get much out of it other than that it probably doesn't matter but DVD+R is slightly better. I just wanted to first confirm that these type formats should work with my computer... yes? and wanted to know if the brand matters much and if so, what brand is recommended. Second, I want to burn a movie that is more than two hours long. I went to a store and looked at the disks and they all list 4.7 GBs and "120 minutes." Does that mean it can only store 120 minutes of material so I would have to break the movie up into parts? (I have no idea how to do that).---141.155.159.27 (talk) 23:17, 29 November 2010 (UTC)

What format will you be downloading and saving your movies in? Most commercial dvds are 8.something GB, while writable DVDs are the mentioned 4.7. A program like DVD Shrink (which unfortunately appears to be Windows only) will shrink a DVD file to the desired size (trading quality for a smaller size). Buddy431 (talk) 23:40, 29 November 2010 (UTC)

Check the label on the edge of your DVD drive tray, if that doesn't say, look up its model online, though it probably won't matter as you suspect. The brand / type only matters if you need to support extra-fast writing (which requires a more durable disc), or want particularly long-lived backups; also some have better labels/surfaces than others. As you can see at DVD there are capacities other than 4.7GB (and this will be decimal, not binary GB [4700 MB, not 4810]), but 4.7 is by far the most popular and simplest to write to. The "120 minutes" is a gross estimate for the ignorant masses. You can store much more than 120 minutes of media, but the longer it is, the lower the quality (the sum of audio quality and video quality [including resolution]) must be to fit in the given space. This is particularly problematic if you want to create a DVD-Video disc, as they use an incredibly old and comparatively inefficient format (MPEG-2). If you want to create a disc that you can give to someone for them to watch most anywhere, DVD-Video is probably your only sane choice. If it's just for yourself, it will save you a great deal of time, effort, and cost to setup a media PC, which you can transfer media to over the network or via USB stick. As for software, you might try ConvertXtoDVD or tovid. ¦ Reisio (talk) 00:20, 30 November 2010 (UTC)

I appreciate your responses but you're almost speaking a foreign language. Re:Buddy431's response: What format? I have no idea. I thought I could use Titanium Toast to burn the DVD as that is the only program I have. Will that work? Is there a free better program I could download? Re Reisio's post, you're talking over my head. I visited the links for setting up a media PC and DVD-Video but it's all Greek.--141.155.159.27 (talk) 01:34, 30 November 2010 (UTC)

Where will you download the movies from? That will allow us to determine what the file format is. Buddy431 (talk) 02:44, 30 November 2010 (UTC)

Titanium Toast appears to be a powerful program, and should do whatever you need to with it. Using a mac, your alternatives are limited. However, searching "free DVD burner for mac" nets this list. The Xilisoft product, for example, has a free trial version that watermarks your output. If you like it, you can buy it for 50 USD, cheaper than the Roxio product that you're looking at. Some of the other ones can be had for even cheaper, and offer free trial versions (but that limit how much you can convert with the free trial version). It's likely (though I've not tried any of the products) that the more expensive ones are either more powerful or easier to use (or both).

You can try any of these programs for free to see if they do what you want. DVDs themselves are cheap, so if you screw up, it's no great cost. Try the Xilisoft product and see if you can get your full program on one DVD. If you can, you can buy the full program (or just live with their watermark). Buddy431 (talk) 03:11, 30 November 2010 (UTC)

Before buying anything, check this first: do you have a DVD burner in your computer or only a CD burner? If your burner isn't built for DVDs, not amount of software will make it so. I'm not familiar with iMac specs, so my apologies if this is obviously not an issue, but it's worth checking before proceeding. Matt Deres (talk) 16:14, 1 December 2010 (UTC)