Wikipedia talk:WikiProject Spoken Wikipedia/Archive 1

From Wikipedia, the free encyclopedia

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.
Archive 1 Archive 2 →

Contents

Names

We should have a consistent naming convention. Is there any reason not to use the full exact name of the article for the name of the .ogg file? Fuzheado | Talk 23:41, 13 Apr 2005 (UTC)

I agree that the article name should be used to name the OGG file. This would be easier if there was a "move" function that renames files. :-) — Timwi 14:59, 14 Apr 2005 (UTC)
  • Should we recommend that the filename be the same as the canonicalized article title? Demi T/C 07:25, 2005 Apr 14 (UTC)
    • Yes, I think we should. :-) Does anyone object if I rename the existing recordings? — Timwi 10:45, 14 Apr 2005 (UTC)

I didn't rename any files, but I put this recommendation in the "Production Notes" section Demi T/C 17:16, 2005 Apr 14 (UTC)

Could you guys please come up with better names for these things than "All your base are belong to us" => "aybabu.ogg", perhaps the full title + .ogg or en- + full title + .ogg —Ævar Arnfjörð Bjarmason 21:01, 2005 Apr 14 (UTC)

  • I moved your comment here, where it was already being discussed--hope you don't mind! As you see, we have put that recommendation in the guidelines already. If there a way to rename files, I would have done it, but it seems to me for tagging reasons it'd be much better for the recorder to re-upload: the license goes with the file better and they can ask for the old file to be speedy-deleted. Demi T/C 21:32, 2005 Apr 14 (UTC)

Updates

This is a great project, and I very much like it, but my question, I guess, is what happens when an article is changed or expanded? Is a new sound recording going to be made?—Ëzhiki (erinaceus europeaus) 15:50, Apr 14, 2005 (UTC)

  • I would think so, more or less. Good judgment should probably be used as to the stability of the version being recorded--picking a version in the middle of an edit war would be a bad idea. I think it's also why it's a good idea to start on the featured articles first, since they are less likely to be rewritten or greatly expanded. Demi T/C 17:04, 2005 Apr 14 (UTC)
  • How about simpy re-recording the (substantially) updated/added paragraphs, and simply inserting them in where neccessary? Flipping between different voices might sound odd for small bits, but for whole paragraphs it might sound okay... ?? --Frankie Roberto 15:06, 29 Apr 2005 (UTC)

File Format

  • Is there any value in saving a lossless format like FLAC as well as the ogg? So it could be reprocessed in the future more easily?
    • No, by that time the article will need to be re-recorded anyway. silsor 20:35, Apr 13, 2005 (UTC)
  • Do we like the new bitrate? 48 kbps sounded a little "watery" to me, but maybe it's my ears or headphones. Demi T/C 17:20, 2005 Apr 14 (UTC)
    • Actually its likely your encoder is not set to "high quality". 48kbps is the standard speech bitrate for testing purposes, anything more than that is pointless space wasting. I'm a professional audio engineer and I can assure you if your planning on a long term speech recording program, 48kbps, 44.1khz mono is the way to go.  ALKIVAR 18:11, 14 Apr 2005 (UTC)
      • As I understand it, for vorbis encoding, you can meaningfully either set "quality" or a nominal (average) bitrate, but not both, when encoding. Do the files sound right to you? Demi T/C 21:35, 2005 Apr 14 (UTC)
        • They sound a little tinny, but I think most of that is due to the poor microphones. The watery sound is very very minor and mostly because of the variable bitrate IMO. I always prefer a fixed bitrate.  ALKIVAR 15:17, 15 Apr 2005 (UTC)
      • I'm experimenting with some of the encoding parameters (and other codecs, too). I suspect that the 44100 Hz sample rate may be too high--ogg might do a better job at the lower bitrate with a lower sample rate. I might try 16000 or 32000 Hz and see what the results are. However, I suspect that the "non-CDDA" sample rates might throw off some players, since almost all players will have been coded with CD ripping in mind (I noticed, for example, that Buzkashi.ogg, with a 96000 Hz sample rate, caused some problems for me and for Raul654; I haven't had a chance to try it on my commercial ogg-enabled mp3 player. Demi T/C 22:14, 2005 Apr 16 (UTC)

I did some recording and listening tests. One of the things I tried was a lower sample rate: 16000 Hz instead of 44100 Hz, since someone had told me this was "adequate" for voice recordings. It is, but it lacked "brightness"--I could definitely tell the difference. I also experimented with various encoding qualities and found that I can hear the difference between "quality 8" (about 128 kb/s) and "quality 3" (about 48 kb/s--our current recommendation). The "low quality" encoding resulted in audible compression artifacts. "Quality 5" (about 64 kb/s) was better, however, so the upshot is that I will continue recording at 44100 Hz mono, but I'm going to bump up the average bitrate to 64 kb/s ("quality 5"). I'm giving the quality numbers here because I think you can enter those into Audacity under the Ogg Export preferences. Demi T/C 07:05, 2005 Apr 17 (UTC)

  • Interesting, I've never used Audacity so I dont know how they set it up. In the software I use you can set a specific bitrate, and it will save to that. I think i'll have to load up Audacity and see what is comparable so that we can nail down this ambiguity.  ALKIVAR 10:36, 19 Apr 2005 (UTC)
    • Awesome (there's some additional discussion on encoding below, too). While you're at it, I'm trying to get a handle on the audio levels advice on the project page, particularly about compression. I can apply a "compand" filter to the audio to compress and expand the volume and level it. What it seems to want as input is: an "attack" and "delay" (for increasing and decreasing volume, in seconds--the "time over which the absolute value of the input signal is integrated to determine its volume." As a trial I've specified this as -0.3,1.0 (taking these numbers from an example). It also wants a curve of decibel points telling it what levels to boost: for example, I've used -30,-15,-20,-12,-4,-8,-2,-7 (which boosts -30 to -15, -20 to -12, and decreases -4 to -8 and -2 to -7); here I'm just taking a guess at what curve will converge at -12 dB. Anyway, any information our clues you could provide on the better use of this would be greatly appreciated. Demi T/C 23:15, 2005 Apr 19 (UTC)
    • After some trial-and-error and listening, I settled on the following compander settings for Evolution.ogg: compand 0.1,0.3 -60,-60,-30,-15,-20,-12,-4,-8,-2,-7 3

Codecs

I just thought this was interesting in terms of the free codecs from Xiph:

   27483537 Apr 15 00:10 Caesar_cipher.flac
    4172650 Apr 15 00:17 Caesar_cipher.ogg
    2402619 Apr 15 00:25 Caesar_cipher.spx
  122213836 Apr 15 00:10 Caesar_cipher.wav

Other notes

  • as a side note, freematrix radio (freematrix.us) would be willing to air these after some review and talk about a way to put them into some sort of way to make it good for a radio show, and maybe something else, contact Apple in #freematrix, #wikinews, or #wikimania on irc.freenode.net (I might join #wikipedia soon enough). and we can work something out.
  • GFDL compliance - do we need to worry about GFDL guidelines, like having to cite the most significant authors? Do we need to provide a spoken list at the end, or perhaps speak the Wikipedia URL, or append the GFDL license?
    • I think to be safe the reader should recite the license notice at the end, or merely state it's distributed under GFDL. Ghost Freeman 10:17, 13 Apr 2005 (UTC)
    • The German recordings seem to begin by stating that the material is from Wikipedia, mentioning the URL http://www.wikipedia.de/ and inviting the listener to look up the article there, and that it is licensed under the GFDL, mentioning the URL http://www.gnu.org/copyleft/fdl.html, before proceeding with the actual article text. Perhaps we should do the same. — Timwi 13:22, 13 Apr 2005 (UTC)
      • I begin my recordings with "<Article name>, from Wikipedia, the free encyclopedia, www.wikipedia.org." and end each one with "This sound file and all text in the article is licensed under the GNU Free Documentation License, available at www.gnu.org/copyleft", NPR-style. User:Luigi30 (Υσηρ ταλκ ΛυηγηΛ) 02:24, 14 Apr 2005 (UTC)
        • I'm following Luigi30's convention of the opener and tail. Also, since the Ogg container gives us the option of several additional tags, I added the following to the file itself (besides an artist, date and title tag; this is an example from my Lithia Park recording above):
          source
          From Wikipedia, the free encyclopedia www.wikipedia.org
          copying
          Licensed under the GNU Free Documentation License, http://www.fsf.org/licensing/licenses/fdl.txt
          article
          http://en.wikipedia.org/wiki/Lithia_Park
          version
          08:43 2005 Apr 11 UTC

Diagrams (inline); Quotations

I've run into a couple of situations where the proper reading isn't clear. Let me say how I've handled it and then we can discuss it so that, if someone wishes, they can follow a consistent practice.

Most diagrams and images I just make no reference to, since they are "decorations" that add to the article but aren't necessary to read. However, occasionally there are inline tables or diagrams where the text doesn't make sense unless the diagram or table is included, perhaps because the text makes reference to something the diagram says or introduces it. In that case, I've introduced the diagram by saying "View diagram: " and then a brief description of it, before continuing with the text.

Secondly, I considered introducing (long or block) quotations by saying "Quote" or something similar; in the end I just changed the cadence of my reading to try to make it clear I was reading something else: more conversational, and ended with the quote source.

Demi T/C 22:34, 2005 Apr 15 (UTC)

I used the words "open quote" and "close quote". — Timwi 23:11, 15 Apr 2005 (UTC)

Templates

  • This discussion was ongoing at 2005-06-28 archive, and may appear later in this archive, in another archive, or the current edition of the talk page.

Date

I've copied this discussion from the template's Template talk:Spoken Wikipedia:

Wouldn't it be nice to record the date the recording was made, so that users can actually tell how up-to-date it is?

  • Not just the date, record the exact revision it was taken. Also a link to the comparison between that revision and the current revision (i.e. the equivalent of clicking the (cur) link adjacent to that revision on the history page) might be useful. Thryduulf 09:11, 17 Apr 2005 (UTC)

Unfortunately, you can only link to old revisions or the (ever-moving) current revision - there's no permanent address for the revision that is currently visible. A null edit serves as a workaround. --Andrew 09:17, Apr 17, 2005 (UTC)

hmm, true at present, although I understand the developers are working on this. Another work around is to view the newest non-current revision and then click the link to view the newer edit. I think this generates a permanent url that basically says link to the version that is one newer than the one with the id given. Thryduulf 12:34, 17 Apr 2005 (UTC)
I don't think so: Doing that on this page gives you the URL http://...Template_talk:Spoken_Wikipedia&redirect=no, whereas a permanent revision URL is of the form http://...Template_talk:Spoken_Wikipedia&oldid=12429402. Maybe as a temporary measure, until there is a fix in the code, the template should contain a timestamp of when it was created? Or else just have the poster first add the template and then find the revision URL of the previous version. — Asbestos | Talk 10:02, 18 Apr 2005 (UTC)

In how many places should we put the version? Below, there is the suggestion to include it in the audio file, which I'm not sure is that useful, as I think about what it will mean to re-record my versions as the articles change. I do put the version on the file's description page in addition to the project page. Should it go in the template as these folks suggest? Demi T/C 21:11, 2005 Apr 18 (UTC)

Requests for Spoken?

Shouldn't we have a section for spoken articles? If not, I guess this will do. Ghost Freeman | Talk 21:32, 16 Apr 2005 (UTC)

  • I've added a section for "Articles to read" on the project page where requests can be suggested Demi T/C 22:27, 2005 Apr 16 (UTC)

Text-to-speech

I'm a little surprised to find this project starting up instead of something that trained and operated a text-to-speech program...it seems like that would provide a lot more coverage a lot faster, would make it a lot easier to keep things up to date, and wouldn't require special audio equipment. Though having actual people read the articles is a nice touch. On the other hand, it's not like people can't use their own text-to-speech programs in combination with their web browser, which sort of brings in to question the entire effort. Though perhaps it is worthwhile to make audio available to people who don't own or who can't set up such software. It would also be nice if there were some collaborative way to give pronounciation hints on words that a computer would mess up. -- Beland 11:27, 17 Apr 2005 (UTC)

  • Well, if you can listen to eleven minutes of synthesized speech reading an encyclopedia, you're a hardier person than I! In any case, I note the popularity of talking books for the blind as well as audio books (recorded by people), and I think it's a useful endeavor. Besides the utility of spoken articles for the blind, it could be useful for English learners (for whom syntehsized speech would be much less useful) as well as just those who like it. I don't see this project as discounting any of the useful text-to-speech tools someone might use, but it also seems to me that adding spoken media where we can is worthwhile, too. Demi T/C 18:56, 2005 Apr 17 (UTC)
    • I'm with Demi on this. It's remarkable how bad text->speech technology still is. I tried listening to some from the German Gesprochene Wikipedia, and it just doesn't work. After 15 seconds, it causes a brain cramp. That said, I'm open to the idea if better technology exists. Fuzheado | Talk 07:39, 19 Apr 2005 (UTC)
      • However, good text to speech gets it is still boring listening to the same voice. And as people say it is currently highly variable in quality. This is a great project. :ChrisG 23:44, 20 Apr 2005 (UTC)
        • I worked with a blind student at a summer job, and I really got to hate his screen reader! I can't imagine how anyone could listen to it for hours. Even though the project is time-consuming, I support doing it with real people. Plus, the vocal variation is nice. It's neat to hear all kinds of accents and voices when listening. It sort of sums up the "anyone can contribute" feel of greater wikipedia, huh? Ckamaeleon 20:07, 19 January 2006 (UTC)

Versioning

It seems like a good idea to announce the version date of an article in the spoken introduction, no? -- Beland 11:31, 17 Apr 2005 (UTC)

  • Not a bad idea. Any other comments? Demi T/C 18:03, 2005 Apr 17 (UTC)
  • Although I think this would be better at the end than the beginning. A lot of front-matter makes it less interesting, in my opinion. Demi T/C 18:34, 2005 Apr 17 (UTC)

Table of completed recordings

  • This discussion was ongoing at 2005-06-28 archive, and may appear later in this archive, in another archive, or the current edition of the talk page.

A better idea?

  • This discussion was ongoing at 2005-06-28 archive, and may appear later in this archive, in another archive, or the current edition of the talk page.

Placing and naming

  • This discussion was ongoing at 2005-06-28 archive, and may appear later in this archive, in another archive, or the current edition of the talk page.

Speech synthesis

Festival seems to be close to state of the art for speech synthesis; however, it doesn't build on OpenBSD. Can someone please install it and use one of the 16-bit diphone voices to run a reasonable long article we've recorded through? We're getting a lot of questions about "why aren't we just synthesizing the voice?" and I think it would be fruitful to the discussion to have a spoken article and text-to-speech-ed article to compare. Demi T/C 00:24, 2005 Apr 18 (UTC)

  • It would probably make sense to have a speech-synthesised file for comparison. The German Wikipedia has one, too. They never state this, but to me it kind of looked like they were saying "See, this is what speech synthesis is like, isn't it useless? This is why we're recording things ourselves." — Timwi 11:56, 18 Apr 2005 (UTC)
    • It was indeed horrible, the one at Gesprochene_Wikipedia. Fuzheado | Talk 07:41, 19 Apr 2005 (UTC)

Current featured article on Aramaic

I've just been trying this out on the current featured article which is quite long and involves some real pronunciation humdingers. So far I've reached 15 minutes and am still less than halfway through the article. I haven't been reading unduly slowly but I haven't been pausing either. I stopped there and saved the resulting ogg file -- 8MB so far. Looks like one file containing a straight through read isn't going to be feasible. Maybe I should try recording individual sections of the article. -- Derek Ross | Talk 04:14, Apr 18, 2005 (UTC)

  • Maybe. I just recorded Evolution, which is pretty long: I ended up at over 34 minutes and a 15 MB file, but I plowed through anyway. It's feasible, but maybe it would be more manageable in sections. If we wanted to do this for long articles, we'd have to consider:
    • The intro would be very short--should it be in its own section file or together with the next section?
    • Is there inconvenience associated with downloading multiple files?
    • How do we let people know what they all are? Make a template that accepts a list?
  • On the other hand, it might make change control a little easier and each individual file is a little more managable. Demi T/C 10:05, 2005 Apr 18 (UTC)
    • I could reasonably imagine having two files for each article: One with just the lead section, and one with the complete article. Once you have the complete article, you can trivially derive the shorter file from it. — Timwi 11:56, 18 Apr 2005 (UTC)
I think this project is a good idea, and, as one of the main contributors to Aramaic language, I thought I'd give it a go. The pronunciation 'humdingers' aren't so bad, probably because I wrote most them (!), but the file is big. It would be best to split the file up, but that would disrupt the flow of the article for listeners. Each file would need the top and tail suggested on the project page, and speakers may want to introduce the next file at the end of the previous one. Any thoughts? --Gareth Hughes 12:25, 18 Apr 2005 (UTC)
How big is your file? Have you uploaded it yet? Maybe someone can convert it for you to a lower bitrate, but you'd have to upload it first :-). — Timwi 12:30, 18 Apr 2005 (UTC)
I'm working on squeezing that file down. I've uploaded the lead section quite easily: you can find it at Image:Aramaic language (lead section).ogg. --Gareth Hughes 12:52, 18 Apr 2005 (UTC)

A Welsh priest is probably perfectly placed to do justice to the pronunciation both of Ahiqar and of Achaemenid. The rest of us probably need to take a deep breath for one or both of them. ;) -- Derek Ross | Talk 14:22, Apr 18, 2005 (UTC)

Thank you for the vote of confidence, I think! I'm using Audacity to record the article. Can anyone suggest the best way of recording the sections of the article and preparing the audio for upload? --Gareth Hughes 14:32, 18 Apr 2005 (UTC)

End material

At the end of any article there can be a vareity of common sections, such as See also, References, External links, footnotes, and so on. I think that most of those should be read out with the exception of "external links." In most cases the title of the link on the page, not the url. Is there a consesnus on how to handle this material? Cheers, -Willmcw 00:25, Apr 19, 2005 (UTC)

  • I don't think there's a standard yet. What I've been doing is reading all the See also, References and External links information including the link URL. It's tiresome to listen to, but I do it on the logic that since the article is essentially over, a listener can always hit "stop." Demi T/C 01:11, 2005 Apr 19 (UTC)
  • If the URL is short, you can say it if you want, but if it's weird (ie has ?, =, etc) you can just give the title of the page and the domain it's hosted on, or some other info that'll let someone find the page again. I'd definitely do this for sources, references, See also, but External links seems optional, unless, for example, the article is on a company, and the link is to the company's web page. --brian0918™ 01:27, 19 Apr 2005 (UTC)
  • I'm inclined to speak See also and important References but not External links because there is not much use for the latter in a spoken form. Fuzheado | Talk 16:51, 21 Apr 2005 (UTC)

Too big/splt

I recorded Robert Oppenheimer. The article is 42 minutes long and is over 26 megs. Wikipedia has an 8 meg limit on uploaded files. I have separate tracks for each section so splitting it is not a problem at this end, but how would multiple files be handled by the template? Thanks-Willmcw 02:11, Apr 19, 2005 (UTC)

  • Are you sure you couldn't cut down on the filesize? Talk to some of the people on here who have longer files with small filesizes. --brian0918™ 02:21, 19 Apr 2005 (UTC)
  • Also, you can upload it to Commons, they have a higher limit.
I tried to upload it to the Commons, but got the same error message about an 8 meg limit. I can't find a file-size limit documented anywhere there. -Willmcw 05:30, Apr 19, 2005 (UTC)
  • Well, Evolution.ogg is 15MB + and I uploaded it to Wikipedia. It takes a while, and today it took me three tries to do so. Anyway, Evolution is 34 minutes long, and that's the longest file currently in the completed articles list. Your Oppenheimer file is encoded at an average bitrate of 86 kb/s or so, which is higher than what we're recommending (48 kb/s) or what I'm using (64 kb/s). If you don't have direct control over the target bitrate, try setting your Ogg encoding quality to something lower and see what you get. In any case, I do think we have to address the issue of split files at some point. Demi T/C 05:38, 2005 Apr 19 (UTC)
Thanks for the feedback. I used Audacity set at 44.1 khz mono, so I don't know how it grew so large. Does anyone know of a program that can lower the bitrate of a file? Also, does anyone know why I'd be unable to upload a large file if the 8-meg limit is not definite? Finally, virtually all featured articles are long, so with a focus on them there will be lots of big files. Would the (old) list of most-accessed files be another source of articles to work from? Cheers, -Willmcw 05:46, Apr 19, 2005 (UTC)
Ah, I found the OGG quality settings and have been able to reduce the size of Robert Oppenheimer slightly. That resulted in a mere 16meg file (at Audacity/OGG quality level 0). I uploaded it successfuly, but for some reason Wikipedia has decided that it is an image. I think I'll try a shorter article next time. ;). Thanks for the help, everybody. Great project. -Willmcw 08:51, Apr 19, 2005 (UTC) (veteran of "Readings for the Blind")
Wikipedia doesn't "decide" that something is an image, but all files are uploaded into the namespace which, due to historical reasons, is named "Image". You can link to the file directly using [[Media:Robert Oppenheimer.ogg]], or link to the description page using [[:Image:Robert Oppenheimer.ogg]] (notice the leading colon, which tells the software not to try to inline it as an image). — Timwi 09:19, 19 Apr 2005 (UTC)

Splitting or abridging

Listen to this article (2 parts)
Icon of loudspeaker
Note: the audio file may be out of date as changes are made to the article. (audio help)
Listen to this article (3 parts)
Icon of loudspeaker
Note: the audio file may be out of date as changes are made to the article. (audio help)

On the subject of splitting files--I agree that many of the FAs are going to result in quite large audio files: 30 minutes to an hour. I propose the creation of new {{Spoken Wikipedia-n}} templates (n 1 - 5: I don't think we have much need for splitting into more than 5 parts), along with a guideline that audio files should be under 10 MB in length and split on a section boundary. What does everyone think? Is there someone with greater template skills than I that can put this in one template somehow? Demi T/C 08:30, 2005 Apr 19 (UTC)

This brings up something we need to decide - do we always want to speak the whole article? Is it reasonable or desirable to make an abridged version of an article that makes it more "ear friendly?" Or perhaps only speaking certain sections? Writing for reading's sake is quite different than writing for the ear. I myself have taken to editing one of the articles (Pulitzer Prize) before recording it, in order to split a really long sentence into two shorter ones. Fuzheado | Talk 08:47, 19 Apr 2005 (UTC)
Boy, I wish I'd edited the Robert Oppenheimer article before reading it. Some of the sentences went on for clause after clause, popping back to life just when I thought they'd be winding down. Featured articles are good, but long. To get the project going we also need a source of good short articles in order to build up a useful library. The Wikipedia 1.0 project is looking at using only the introductory paragraphs of articles, but their list may not be ready for a while yet. Any other suggestions for logical targets? User:Demi, the template looks great. For many articles discrete files for sections would make more sense. I assume it'd be possible to replace or pipe the "Part 1" with "Section headers". If not, that's ok too. Cheers, -Willmcw 11:07, Apr 19, 2005 (UTC)
Ha, yes I've taken to the habit of finely editing the article before reading it aloud. So it winds up taking about an hour to do a 7 minute recording, because I correct mistakes, split long sentences, make things work for listening, etc. I'm trying some articles that benefit from spoken pronunciation, so a test case is Chinese-oriented dim sum. Also, I've found it a bit easier to make something resembling a radio script (see Dim sum/spoken) which has the intro written out and the GFDL at the end. Fuzheado | Talk 12:02, 19 Apr 2005 (UTC)
I've done a little of that--a very little. I see what we're doing as "speaking the article" rather than "making an audio presentation of the article's subject." As such, faithfulness is important (to me at least). In the case of bad writing, I've corrected some of it--after all, we're editors too, and if something is bad writing when spoken aloud it's probably just bad. If I made any but quite minor changes, though, I'd probably make them and wait a day or so before recording it, just to give other editors a chance to have their say, if any. Demi T/C 18:22, 2005 Apr 19 (UTC)
On the template, Willmcw--I'm not sure, because I'm a newbie with the template code. However, I also see these recordings as eventually being available through some system outside the main article (Wikipedia by phone? On CD?), so maybe we shouldn't rely too much on the structure of the article itself? Demi T/C 18:22, 2005 Apr 19 (UTC)
Good point. Re: list of potential articles, I came across this list of so-called "core topics". Many of them are reasonably short and stable, and they of presumed value. It's not official, or even vetted, so I don't know if we'd want to adopt it. However it's a good starting point. Cheers, -Willmcw 20:56, Apr 19, 2005 (UTC)
Ah good. I see that list happens to point to "Simple English" articles. Those are shorter (and simpler!) than the non-simple English articles. If we read a "simple English" article, is it still appropriate to link it back to the non-simple English article? And vice versa? hmmmm. -Willmcw 23:35, Apr 19, 2005 (UTC)
Simple is its own Wikipedia, so I would classify something like that in the same vein as an interlanguage link. We can have it, but it should be clear it's not the same thing. I wouldn't use our template on the en: article page for an article read from simple:, myself. Demi T/C 18:40, 2005 Apr 21 (UTC)
If we do split articles into parts or sections, how do we handle the linkage? Do we speek the license at the end of each part? Say "From Wikipedia..." at the beginning of each? Say "Continued on part 2..." and/or "Continued from part 1...?" Demi T/C 23:54, 2005 Apr 19 (UTC)
My inclination would be to read the article title as "Worms, part 1. license info." Maybe even make it "Part 1 of 3 parts". Then at the end, "continued in Worms, part 2, end license info". In any case, the full license info should be in each recording. -Willmcw 23:59, Apr 19, 2005 (UTC)

OGG spec

In the document now, it says use OGG at 48 kbps, but Audacity and other say that with Ogg you cannot exactly rate limit anything since its VBR, and it only has a 1 to 10 slider. Should we not specify instead what quality to use? Fuzheado | Talk 07:43, 19 Apr 2005 (UTC)

  • Maybe. See my notes about sampling and encoding quality and bitrate above. The command-line encoder allows you to specify a target bitrate, or a quality from 0-10. I find, with that encoder on my audio files, 3 is about 48 kb/s, 5 is about 64 kb/s and 8 about 120 kb/s. However, Timwi tells me that he encodes with quality set to 5 and his resulting files are 110 kb/s or so. I think it's a good idea to provide recommended settings for our "Recommended tools" section, but I don't know what it should be. Demi T/C 08:14, 2005 Apr 19 (UTC)
    • I just encoded Dim sum at level 3, which is about as low as I can go without it sounding quite "wormy." Fuzheado | Talk 11:58, 19 Apr 2005 (UTC)

Audacity HOWTO

Since a lot of people are using Audacity, it might be good to have a "howto" or cheat-sheet on how to use it; which menu options to use, what buttons to press, etc. I've done that for the tools I'm using, but an Audacity version might be helpful for people coming here to make their first recording. Demi T/C 18:26, 2005 Apr 19 (UTC)

See also

  • This discussion was ongoing at 2005-06-28 archive, and may appear later in this archive, in another archive, or the current edition of the talk page.

Template on Talk Pages

I don't think we should have the template on Talk: pages -- it's not the place for it. I wrote on User talk:Brian0918:

I also think the link is better solely on the article page. For me, a useful rule of thumb is that things that primarily target the readers/users/audience of Wikipedia belong on the actual article, and things that are directed at editors belong on the Talk page. In this case, a link to a recording of the article targets the readership (er, "listenership"!), rather than editors, and would be better (IMO) just on the article page, and not on the Talk: page. — Matt Crypto 02:13, 20 Apr 2005 (UTC)

  • But the Spoken Wikipedia template links to the WikiProject, just as other talk pages have wikiproject templates at the tops of them. It's similar to the featured article template, which isn't specifically about editing the page, but has links to ways to contribute. I talked to Raul654 and other WikiProject members, and they're fine with it. We could probably change it in the future, when we get more people. We could just create another Spoken Wikipedia template specifically for talk pages (but that should still have a link to the audio file of the article). --brian0918™ 02:18, 20 Apr 2005 (UTC)
Well, the "Featured" status of the page is useful metadata about an article for editors, as it is an end-result reflecting peer-review and stability of an article. (I would, incidentally, advocate that a Featured template would do well to be on the main article page itself, since the information is useful to readers, too). However, a Spoken template doesn't really provide a great deal of useful editorial information, at least as I see it. It may link to the Spoken WikiProject, but that's not really like other WikiProjects which exist to coordinate subject areas. It's clearly not a hugely important issue, but I think we need to be careful how much clutter we add — for example, on Talk:Caesar cipher, it becomes difficult to see the Table of Contents. Adding a Talk page template involves an (admittedly quite small) cost, yet provides no real benefit for editors, or at least as far as I can see. It might raise the profile of the WikiProject a little, but I'm sure there's better ways of getting the (most excellent) Spoken WP project known than adding clunky templates to Talk pages. — Matt Crypto 02:32, 20 Apr 2005 (UTC)

Alright, I created a template specifically designed for the talk page. See the WikiProject's page. I'll go through and add them to the appropriate articles. --brian0918™ 02:42, 20 Apr 2005 (UTC)

Thanks, although I'm still unconvinced this Talk page template is of any utility other than an advert for Spoken. That's not necessarily a bad thing, of course, but is probably off-topic on Talk: pages for articles on arbitrary topics — at least, that's my take. But, again, it's no big deal, and I'm happy to go with whatever people prefer. — Matt Crypto 02:50, 20 Apr 2005 (UTC)
So far, I've talked to Raul654, Demi, Luigi30, and silsor, who are all fine with it. Check out Talk:Caesar cipher. All those WikiProject templates are just like this one (advertisements?). --brian0918™ 02:56, 20 Apr 2005 (UTC)
OK, I'm happy to accept it if the consensus is that the templates on the Talk: page are beneficial. I would, however, argue that "topic" WikiProject templates are different in two ways: 1) It'd be inappropriate to put a Project template on an article page, unlike a Spoken template; and 2) An editor interested in editing a article in a certain field is especially likely to be interested in any associated WikiProjects which concern that topic. There is no special reason why an editor on arbitrary topics would be interested in the Spoken WikiProject. Would we be happy with links to Wiktionary, WikiQuote etc on the Talk pages? — Matt Crypto 03:15, 20 Apr 2005 (UTC)
Sure, anything that promotes any aspects of Wiki is fine by me. I didn't find out about several of the other projects until I randomly came across them. --brian0918™ 03:19, 20 Apr 2005 (UTC)
Well, perhaps. To argue an straw-man extreme, you could include links to every single Wikipedia project on every single Talk page, but that wouldn't help people find things — it would just increase the noise of the project. We need to weigh things up, and see if the benefit of the link is worth it. In my opinion, a good way to keep the clutter down is to try to stay on-topic. There are plenty of other places on Wikipedia to promote worthy projects, but an article's talk page is primarily for discussion about ways to improve that specific article. Of course, it is appropriate to include links to WikiProjects in the same field, or to mention the Featured Article status. But, for me, links to projects like Spoken, or Wiktionary, or WikiQuote (or whatever) just seem too generic. (They're fine on the article page itself, of course). — Matt Crypto 03:28, 20 Apr 2005 (UTC)
Wiktionary and WikiQuote are sister projects though, which get main page advertisement. Spoken is just a WikiProject, and can't get the same type of advertisement that other WikiProjects do (because it's not about a specific topic). So we have to advertise whenever possible. Once the project gets much larger, we might consider getting rid of the talk page template. --brian0918™ 03:33, 20 Apr 2005 (UTC)
OK, I guess this is the crux of the problem I have: the benefit is all for the Spoken project, and not for the articles themselves. I would say that if you're going to include an advert on an article Talk page, it should be for the benefit of that particular article. If all that these Talk-page templates provide is advertisement for Spoken, then I think they're misguided. I do, of course, understand the importance of getting people informed, but we have plenty of places like the Village Pump, the Community Portal, etc, which are appropriate to point people towards things like this. — Matt Crypto 03:41, 20 Apr 2005 (UTC)
Current WikiProject talk page templates suggest either editing the article or visiting the project page to join the project. We don't really have the option to edit the audio file, so we have to go with what we can do, visit the project page. --brian0918™ 03:48, 20 Apr 2005 (UTC)

Possition of temptate with no CSS

Just thinking about this a bit. basicaly, if i was blind, then wouldn't i want to know about the recorded version befor having m text reader speek the full normak version befor hand. To demostrate what I meen, disable css in firefox for a moment, and you will see that the spoken version comds below the artcal text, this seems non-sensical to me, as i would have probably just lissened or read the full (text) artical befor seeing... oh, i could have just lissened to that recording of it.

Does anyone have any problems with placing the template at the TOP of the page? tooto 19:15, 20 Apr 2005 (UTC)

Apparently it goes against policy. I think what will eventually happen is Spoken Wikipedia will become a sister project, eliminating that problem. --brian0918™ 19:21, 20 Apr 2005 (UTC)

Name of this project

I regret having started this project under a name that was a literal translation of the German equivalent. I'm beginning to think that "Spoken Wikipedia" sounds really rubbish. Would anyone object to a significant name change? (The sooner we make such a change, the less work it will be.) — My suggestion for a new name would be "Audio Articles", which personally I like much better, but of course I am open to other suggestions too. — Timwi 21:51, 20 Apr 2005 (UTC)

  • I personally am fond of the name Spoken Wikipedia; but I care enough to onject, per se. BLANKFAZE | (что??) 22:04, 20 Apr 2005 (UTC)
  • I actually like Spoken Wikipedia. My guess is that eventually this will (or at least should) become a sister project, so it should have a comprehensive name like that (as opposed to Audio Articles)... maybe Audiopedia? --brian0918™ 22:05, 20 Apr 2005 (UTC)
  • No problems at all with Spoken Wikipedia. Fuzheado | Talk 00:00, 21 Apr 2005 (UTC)
  • I think "Spoken Wikipedia" is clearer than the above alternatives. 119 03:33, 21 Apr 2005 (UTC)
  • "Spoken Wikipedia" is good, I think. The thing that makes it awkward is prefacing the name with "WikiProject." It might look less awkward if we just referred to it as "Spoken Wikipedia" in our "collateral"--that is, the templates and so forth. Demi T/C 05:25, 2005 Apr 21 (UTC)
  • Spoken Wikipedia. silsor 05:37, Apr 21, 2005 (UTC)
  • I think "Spoken Wikipedia" is OK. — Matt Crypto 14:30, 21 Apr 2005 (UTC)

Short featured articles

I've created a list of short featured articles that haven't been on the Main Page. By requesting that they be on the Main Page, and then turning them into spoken audio, we can more quickly and easily increase the publicity of the project. --brian0918™ 12:59, 21 Apr 2005 (UTC)

An idea?

What do folks think of the idea of getting someone who's a fan of Wikipedia to record the article on themselves. No, I don't mean Jimbo, I was thinking more along the lines of someone like Wil Wheaton. Thoughts? It might be fun, publicity wise, for the Spoken project. -- Seth Ilys 20:23, 21 Apr 2005 (UTC)

  • Good idea. That would be neat, if you could actually get someone to do it ;-) BLANKFAZE | (что??) 20:33, 21 Apr 2005 (UTC)
  • If anyone's been thinking about joining the FSF, tell them I sent you. If I get three referrals the prize is RMS (or Eben Moglen) will record an "answering machine" greeting. I was thinking instead of asking him to record an article. Demi T/C 02:19, 2005 Apr 22 (UTC)

Ideas for the future

Once we get more articles, we should probably change the layout. Instead of the article template linking to the project, it'll link to Wikipedia:Spoken articles, which will have the layout of a sister project main page, or of Wikipedia:Featured articles, and list the articles we have in the appropriate subject headings. The list will be made of links that don't go directly to the files or file pages, but to subpages, Wikipedia:Spoken articles/XXXXX, which will be fancier pages that have file information and a direct link to the file (bypassing the ugly and confusing Image: page). --brian0918™ 21:19, 21 Apr 2005 (UTC)

Perhaps we could also have a link to the audio file above the language links, rather than a template in the article? Joe D (t) 21:26, 21 Apr 2005 (UTC)

How would you go about doing that? I think that would go unnoticed by most people. --brian0918™ 21:28, 21 Apr 2005 (UTC)
Less noticable than the interlanguage links? Joe D (t) 21:42, 21 Apr 2005 (UTC)
In another vein, it might be nice to have a special tag to label the media files. I've been labelling mine {GFDL}, but maybe a {Spoken-GFDL} or similar tag could be made to mark them as spoken article files. That way the files could be easily found, in case they get droppped from an article, or some other break in the chain. There are numerous special tags like that for images, so it wouldn't be unprecedented. -Willmcw 22:04, Apr 21, 2005 (UTC)

Position of links

Copied from Wikipedia talk:Spoken articles: hadn't realized I wasn't on the project page...

This spoken articles project seems like an excellent idea, and I'm amazed at the amount of work that's been put into it. I just have one question: all the links on articles to the spoken version are down at the bottom of the article. This generally means that you only realize it exists after you have read the article, which seems to rather defeats the purpose.

If this project ever starts covering a significant proportion of Wikipedia articles, and people with difficulties seeing/reading start to keep an eye out for spoken versions, I think it would need a notice that stands out somewhere near the top. If there were a small speaker symbol or something up near the top, people would know instantly that there was a spoken version.

Minor point, I'm sure. It just always strikes me as odd whenever I see a link to the spoken word version after having read the entire article. — Asbestos | Talk 23:08, 21 Apr 2005 (UTC)

  • I know, there is a way of doing this (I think) but I'm loathe to add yet more instructions to the steps for recording and listing an article. Demi T/C 02:00, 2005 Apr 22 (UTC)
  • Perhaps a disambig-like:
    Spoken version available.
    119 02:14, 22 Apr 2005 (UTC)
    • I like that idea. --brian0918™ 02:17, 22 Apr 2005 (UTC)
    • I was thinking more along the lines of a <div style="display: none"/> section at the very top of the page with a link to the audio file. That way no one can complain of clutter but a screen reader will see it (I think). Demi T/C 02:23, 2005 Apr 22 (UTC)
    • FYI, see {{User:Demi/spoken-header|Filename.ogg}} for what I mean. Demi T/C 04:22, 2005 Apr 22 (UTC)

Help fix these links

The article link should be of the format [[:Image:FileName.ogg|ArticleName]] at Wikipedia:Spoken articles. Click through the link, and on the file page, replace the contents (except for the copyright notice, which is usually {{gfdl}}) with the following text:

{{Spoken article entry|
file_name= 
|title= 
|time= 
|file_size= 
|user_name= 
|date= 
|accent=
|}}
  • File_name should be of the form: name_of_file.ogg
  • Title refers to the title of the actual article.
  • Time is the length of time in minutes and seconds of the article, for example: 12:34
  • File_size is the size of the file, for example: 12.6 MB
  • User_name is the name of the user who read/uploaded the file, for example: Willmcw
  • Date should be a link to the revision that was read, for example: 2005-04-18
  • Accent is the regional accent of the reader. See Regional accents of English speakers for the possibilities.


FWIW, I've been using [[Media:foo.ogg]] instead of [[:Image:foo.ogg]], is there a reason to use one rather than the other? I think the "Media" tag is more accurate. Fuzheado | Talk 02:40, 22 Apr 2005 (UTC)
  • The [[Media:]] link makes a link directly to the sound file, which is convenient, but doesn't provide some of the convenience of the [[:Image:]] link, which links to the description page (now adorned with the template created by brian0918). I think brian0918's intention with this is twofold: 1) Preparing for the day (if it hasn't already arrived) when the list of completed recordings is impractical for a single page and 2) Creating a more accessible "public face" for the recordings (since the project page is more about instructions for participants rather than information for listeners). Not putting words in his mouth, just saying. Demi T/C 04:16, 2005 Apr 22 (UTC)
  • I don't think we really need the table in the project page anymore. That should cut down on the work load. --brian0918™ 04:20, 22 Apr 2005 (UTC)

Template Discussion

There's some disagreement over the links which should appear in our {{Spoken Wikipedia}} template. Can interested parties please visit the talk page and weigh in? Thanks! Demi T/C 18:25, 2005 Apr 22 (UTC)

Suggestion

I think it might be worth summarizing the goals and reasons for the project at the top of the project page. For me, I assumed screen readers were all anyone needed since I know a few blind people that use them and had never heard them mention they weren't sufficient. In any case, having the reasons for the project right at the top should help. - Taxman 14:51, Apr 23, 2005 (UTC)

They are about as "sufficient" as Internet Explorer is to browse the web. People get "used" to it and don't complain simply because they don't know what they are missing. — Of course, I agree to your suggestion. :) — Timwi 21:08, 23 Apr 2005 (UTC)

Problem with my computer

I can't play .ogg files on anything (Real Player, Windows Media Player, etc.) on my computer. -- Tony Jin | (talk) 00:21, Apr 25, 2005 (UTC)

Try Winamp. --brian0918™ 00:22, 25 Apr 2005 (UTC)
Actually try installing anything that supports ogg by default. http://www.vorbis.com/download.psp  ALKIVAR 07:03, 25 Apr 2005 (UTC)
Winamp doesn't support OGG (or Unicode or a lot of other things). I recommend foobar2000 or XMPlay. — Timwi 14:21, 25 Apr 2005 (UTC)
I recently reformatted, and Winamp was one of the first things I installed... and it plays ogg's just fine. --brian0918™ 14:31, 25 Apr 2005 (UTC)

Adding OGG tags

Does anyone know of an easy way to add these file ID tags? I use Winamp for the task but it has a very clumsy interface for doing so. Thanks, -Willmcw 23:34, Apr 28, 2005 (UTC)

  • I'm using oggenc and doing them at encode time. The vorbis-tools package from Xiph includes the vorbiscomment command, which is a command-line tool designed for this task. I don't suppose that helps the Windows users much, however. Demi T/C 18:44, 2005 May 4 (UTC)
Thanks. Another good reason to switch to Linux. In the meantime, I've found many Windows programs which claim to allow tagging of OGG files, but so far Winamp is the only one which allows custom fields. And when I've looked at Winamp-tagged files with those other programs they don't show the tags that I've added in Winamp. Demi, would you mind downloading media:1755 Lisbon earthquake.ogg to see if the tags are showing up properly? If I'm stuck using Winamp I'd at least like to make sure that it is working correctly. Cheers, -Willmcw 19:22, May 4, 2005 (UTC)

Self-referencing

Remember that huge debacle that resulted in Wikipedia:Spoken articles being moved to Category:Spoken articles because the template on the article page linked to Wikipedia: instead of an external link? Well, all we had to do was use {{SITENAME}} in place of Wikipedia in the link name. So, it would've been: [[{{SITENAME}}:Spoken articles]] which creates Wikipedia:Spoken articles, and this would be preserved on third party sites as whatever their sitename is. Should we consider moving the Category: page back to Wikipedia: (since the problem is now solved, and moving it back would reduce the wording confusion and clutter of the article page). --brian0918™ 08:22, 30 Apr 2005 (UTC)

I agree, but I'm not in the avoid self-references cabal so it doesn't really matter. --SPUI (talk) 08:23, 30 Apr 2005 (UTC)

Summary proposal

Having read about ten articles, I have learned to complete an article with less time spent preparing and finishing it. Even so, it is very time consuming. The shortest featured articles are around 2,000 words, about 15 minutes of air time, at least an hour of production time, more with all things considered. Long articles are prohibitively time consuming for most of us. On the other hand, it'd be swell to read at least part of as many articles on topics as possible, in order to make it into a useful reference resource rather than just random articles.

What we need is a sub-project: reading the introductions of articles. Most articles use the initial paragraphs as a summary, so simply reading everything above the TOC would provide a simple yet comprehensive way of choosing the material, and it would be apparent to others as well. It'd be a subdivision of the others templates-"Full/Abridged/Summary"-or something like that. We might toss in the "see also/refer also to" internal links, but avoid all the long links and references, formulas, diagrams, etc. Plus, the introductions necessarily change less often than their full articles, so a mere introduction is likelier to be current.

As an example, look at The Temptations. It's a great article of about 5500 words, none of them hard. But reading it would be a big project, including reading many lists and references, and a big file, maybe 20mb. OTOH, its intro is four well-written paragraphs totalling just 340 words. We could do ten introductions in the time it would take to do one full article of that size. I could see us recording the summaries of the entire 1000-entry [1], but recording the full articles of the entire list looks daunting and tedious. And all of the featured articles. Or all of the articles on some related topic of inerester to the speaker. And the files would be relatively small therefore quick to download and listen to.

To implement this we would have to make some changes to the protocols. Probably the filenames would be something like "Article_title_(intro).ogg" (and "Article_title_(abr).ogg"). The templates would need to allow for additional filenames, labelled as "summary/abridged/full (pt1, pt2, etc.)." We'd probably want to find a way to differentiate between the {{spoken}} tags used in article lists, maybe a grey {{sumspoken}} for articles with a summary only. Naturally, as the proposer I'm volunteering to do all that set-up work (help appreciated). What do other editors think? -Willmcw 08:00, May 1, 2005 (UTC)

Hearing no objections, I'll move forward with this. Cheers, -Willmcw 21:16, May 3, 2005 (UTC)
I don't see any problems with this. I don't want to discourage those who want to do full articles, so let's somehow make it clear that just because an "(intro)" recording exists doesn't mean you shouldn't tackle the full article, if you're so inclined. You might examine the multipart templates ({{Spoken Wikipedia-2}} and friends) for some ideas on how to handle the various parts.
Oh, the other thing I'd prefer not to see is us paraphrasing or summarizing using words that don't appear in the article. The article texts we're reading are wiki-vetted; so while I think it's fine to excerpt them, by reading just the introduction, for example, I'd prefer us stick to the text when reading (since our spoken versions are less plastic than the wiki-text). Demi T/C 18:41, 2005 May 4 (UTC)
Since writing that I've investigated more fully. I have discovered that my random example, The Temptations, has an unusually good introduction and that many articles which would otherwise be perfect candidates have very short intros. Also the added complexity of having a two-tiered system is would almost require a parallel project. So I'm in less of a hurry. I'll keep looking into it though. I do think that the difficulty of recording an article seems to grow almost exponentially with its size, and I wish we could find some scheme of shortening articles. The Simple English wikipedia has short, easy to read articles, but I doubt that anyone would think it right to link spoken versions of those to the full English articles. I agree that free-form abridging should probably be avoided. Abridgement by deleting lists and weblinks may be more acceptable. Half of the featured Roy Orbison article is taken up with several lists of his songs. Even a short list of external links can take a long time to read: sometimes it's necessary to spell them out letter by letter. Whew.
On the other matter, I have been surprized at the poor quality of writing even in featured articles. The facts may have been vetted, but the grammar is sometimes tortured and even the article structure may be convoluted or repetitive. Speaking prose aloud is a more stringent test of writing than merely skimming it for accuracy. There's nothing we can do about that, but a certain measure of re-writing is sometimes necessary. Usually it's just a matter of breaking up run-on sentences and incorporating informative photo captions. Anyway, it's a great project, thanks for getting it started. Cheers, -Willmcw 07:58, May 6, 2005 (UTC)

Direct links

It would be nice to have direct links to the audio files. Clicking through to a separate Image:Article.ogg [sic] page is awkward, especially when starting from Wikipedia:Featured articles or Category:Spoken articles. I think some people may also be surprised to find that clicking on the spoken article icon leads to the image page for that icon, not the recording. Unfortunately there seems to be no way to address either of these issues currently — should that be considered a bug? Even [[Image:All your base are belong to us.ogg]] only produces a link to the "Image" page (and worse, the file is treated as an image).

There should probably be a Sound: or Audio: namespace for these, too... can't imagine that would be easy to arrange at this stage, though. ᓛᖁ♀ 06:25, 4 May 2005 (UTC)

Oh, I've noticed the [[Media:...]] syntax. Heh, I'd forgotten the Media: links just redirect to the Image: namespace. Could some combination of these be used in the template? ᓛᖁ♀ 06:35, 4 May 2005 (UTC)
The Media: namespace links direct link, not redirect. BLANKFAZE | (что??) 11:52, 4 May 2005 (UTC)
I agree it would be most convenient to have a direct link to click on (in the template). The tradeoff is that we would like to present the user with some additional information before they stream/download the file (like its size and length) and some other metadata that appears on the "Image" page. Would it work better to have a link like this: Listen to this article (info)? However (not that these are really in use yet), I'm not sure what this implies for the multipart templates. As for the links on Category:Spoken articles, I can't remember why we linked to the description page and not the file. Maybe brian0918 could comment? Demi T/C 18:32, 2005 May 4 (UTC)
I agree Demi, the image/info links you suggested seems much better. Craigy 18:44, May 4, 2005 (UTC)
The (info) idea sounds good. For multiparts, the info link could go to the first Image: file, which would have links to all the other parts. --brian0918™ 18:50, 4 May 2005 (UTC)
I've implemented Demi's suggestion and made the 'Listen to this article' a direct link with '(info)' next to it. If there any objections or suggestions, please amend my edits to the table. Craigy 01:35, May 5, 2005 (UTC)

Pronunciation of GNU?

Perhaps the project page should have a standard on pronunciation of GNU. I've heard that it's supposed to be said "g'nu" (as in the animal), but I suspect most people will instinctively read out each letter "gee en you"; in fact I probably would if I didn't know "better".

I haven't added anything to the start page, because there needs to be some consensus about which pronunciation is most suitable.--Ejrh 05:54, 6 May 2005 (UTC)

  • According to the GNU project, it's "guh-noo." However, as it's an acronym, I don't think spelling it out is incorrect, either. I'm okay with people choosing either according to their preference; but they should probably not pronounce it as the word gnu. Demi T/C 07:12, 2005 May 6 (UTC)

I think Ejrh is right, we ought to add that to the project page. It is not at all obvious. I wouldn't have known if a helpful editor hadn't corrected me. -Willmcw 07:34, May 6, 2005 (UTC)

  • Yes having either option is good (and "guh-noo" is more emphasised than "g'nu" I guess ;-) Thanks for adding it, Willmcw.--Ejrh 13:15, 6 May 2005 (UTC)
Put more scientifically, the animal gnu is pronounced IPA['njuː], but the project GNU is pronounced IPA[gə'nuː].