Talk:Language committee

Please add any questions or feedback to the language committee here on this page.

Dear members of WMF Language Committee (Amire80, N-true, Antony D. Green, Bèrto 'd Sèra, Maor X, GerardM, Jon Harald Søby, Klbroome, Arria Belli, MF-Warburg, Evertype, Millosh, Baba Tabita, SPQRobin, Santhosh Thottingal, Satdeep Gill, ZaDiak).

For your huge efforts on language diversity and especially on small language projects I want to present you the Flower of the month .

Best regards, --Holder (talk) 20:01, 2 June 2017 (UTC) Reply

Thanks! --MF-W 03:34, 13 June 2017 (UTC) Reply

I would like to point out that the Ingush Wikipedia (incubator:Wp/inh) might be ready ([1]). It's been active for a long time and considering its size it is outgrowing the incubator :) --OosWesThoesBes (talk) 16:26, 17 June 2017 (UTC) Reply

Agreed. We wanted to approve it already half a year ago but found nobody to verify the content. We'll start a new attempt. --MF-W 23:01, 18 July 2017 (UTC) Reply

So, I've been working on incubator:Wt/sco, and I was wondering if the status page (Language committee/Status/wt/sco) could be re-activated with new stats. --AmaryllisGardener ^talk 23:18, 27 June 2017 (UTC) Reply

@MF-Warburg: Would you know how to get this done? (I see you deleted the page) Thanks, --AmaryllisGardener ^talk 01:01, 28 June 2017 (UTC) Reply

@AmaryllisGardener: Hi. LangCom isn't using those old status pages any more. If a test project seems on its face to meet approval criteria LangCom does a deeper-dive investigation and then makes a decision. As someone who is effectively a clerk for LangCom, I would share the following:

• Your test project is probably large enough to meet the size criterion for approval, though adding additional content will always help.
• Both incubator:Wt/sco and Requests for new languages/Wiktionary Scots have links to both an activity tool and a localization statistics tool. Check both of those.
• In order for a project to be approvable, it must have an active test community. Generally, this means that there must be at least three active editors—"active" defined as having at least ten edits in a month—for at least three consecutive months, and then continuing on until the project is approved and exported. So far, this project has only you, and only for one month. (My edits last month were in the nature of housekeeping and don't count, and there were only six.)
• For second and subsequent projects in a language, further work must be done on translating the interface at translatewiki.net. Certainly all messages in the "core MediaWiki messages" group must be translated, and notable additional progress beyond that must happen.
• Once the rest of this has resolved, LangCom may still wish to ask what the relationship of this project would be to the Scots language category within English Wiktionary. Do the projects include duplicate information? Do they both need to exist? I don't know that LangCom will give you any trouble about this, but you may wish to consider in advance what your plans might be.

Since you're a very experienced Wikipedian with mop experience, I'd encourage you to apply for test administrator status at incubator:I:RFTA. That gives you a partial administrator toolkit that you'd probably find helpful.

Any questions, please let me know. StevenJ81 (talk) 14:12, 28 June 2017 (UTC) Reply

@StevenJ81: Thanks, I'll follow some of your advice, and will try to recruit more editors to work on the wiki. --AmaryllisGardener ^talk 18:29, 28 June 2017 (UTC) Reply

Proposals for closing projects/Closure of Ilocano Wikipedia has been closed by me as obviously invalid. I'd appreciate if a LangCom member could review my closure and confirm it or revert it. Thanks, —Marco Aurelio 11:07, 29 June 2017 (UTC) Reply

(not LangCom member) I concur with MarcoAurelio's assessment. StevenJ81 (talk) 12:47, 29 June 2017 (UTC) Reply

Very good. --MF-W 11:32, 30 June 2017 (UTC) Reply

@MarcoAurelio: Technically you can just close it without invoking IAR per the phrase "Any other responsible WMF user, not affiliated with the wiki proposed for closure, may also close a discussion considered disruptive, for stated cause, pending review, and it should not be re-opened except by a LangCom member." on CPP. --Rs chen 7754 06:25, 1 July 2017 (UTC) Reply

Indeed, that's why I wrote "very good". --MF-W 23:07, 2 July 2017 (UTC) Reply

The discussion about the Closing projects policy is discussed at Talk:Closing projects policy#This is not working. --George Ho (talk) 02:43, 30 June 2017 (UTC) Reply

Hello dear community!

Already some time ago, the Board decided that it does not need to be informed anymore by the language committee for a possible veto when a new language version should be approved. It was then decided that "the community" will be informed before an approval and be given one week to raise objections, if there is something serious Langcom didn't see. So far, we always forgot to do this. However now we remembered and inform you that we want to approve

• Bashkir Wikibooks (see phabricator:T173471)
• Hindi Wikivoyage (see phabricator:T173013)

If you have objections, please use this page and base your argumentation on the language proposal policy. --MF-W 15:31, 28 July 2017 (UTC) Reply

This is nice news for us. We have quite active users and we are ready to work as an independent project.--ZUFAr (talk) 17:46, 28 July 2017 (UTC) Reply

The creation of these wikis will now be requested. --MF-W 16:01, 10 August 2017 (UTC) Reply

Done. See phabricator tasks as above. StevenJ81 (talk) 18:02, 18 August 2017 (UTC) Reply

I'd like to bring up some common problems (of varying levels of severity) I have seen in newly created wikis, from my perspective as a regular (non-admin) user. This is all based on impressions I've developed since I started paying closer attention to new wikis, back when the Maithili Wikipedia was created in Nov 2014. (Although, in that time I haven't actually looked through every newly created wiki.) I am hoping that problems such as these can somehow be identified and addressed earlier in the process of wiki creation (say, while requests for new languages are being processed — at the very least, interested users looking at such requests could be advised to look out for these issues).

Lack of local admins

Of the 15 most recently created content wikis, 2 currently have no local admins at all: kbpwiki (created 23 June 2017) and dinwiki (created 12 July 2017). This may just reflect normal "lag", since local admins are promoted only after gaining local consensus, but it would be nice to have at least 1 active admin (who knows the language) right off the bat, to deal with any issues that arise once the wiki opens for editing.

Non-localized namespaces

Again, of the 15 most recent wikis, 2 seem to still have English (canonical, non-localized) prefixes on most of their namespaces: adywiki (11 February 2016) and jamwiki (2 May 2016). This makes it highly likely that in the future, Phabricator tasks will be required to localize these namespaces (and then pages may need to be moved, maintenance scripts run, etc.). This can be avoided by just requiring that all namespace names be localized before a wiki is created. (I thought that was the policy, but apparently not?)

Incorrect initial article counts

Many, if not most, new wikis show initial article counts that are too high (sometimes impossibly so) right after importing is finished. This gets fixed when the wiki is recounted for the first time by the updateArticleCount.php maintenance script (which gets run on all Wikimedia content wikis except Wikibooks on the 21st of each month), but it seems obvious that there is still at least one bug remaining in the MediaWiki software related to article counting and wiki creation / importing. Fortunately, it doesn't appear to be as bad as the old bug that increased the count by 1 for every individual revision of each page imported (rather than for each page), but it's still annoying. This problem can easily be "corrected" (that is, worked around) by simply requesting the running of updateArticleCount.php (or initSiteStats.php?) on new wikis as a matter of course after importing is finished.

Misunderstandings about article counting

A separate article-count issue: Many new wiki communities report "wrong" or "broken" article counts even after the wiki has been recounted for the first time, because no one understands that, by default, only pages (non-redirects) in content namespaces (usually just the main namespace) that contain at least one [[wikilink]] to a page title on the same wiki (whether the target page exists or not) get counted as articles (AKA "content pages"). (Clearly part of the problem is the subtlety of the "link" method of article counting!) Part of the process of getting approval for a new wiki should be an explicit choice of which article counting method the wiki wants and, when appropriate, which namespaces are "content namespaces" — hopefully with a good understanding of the implications of these choices.

Suboptimal wikicoding

When the contents of new wikis get imported from the Incubator (or Wikiversity Beta, or Multilingual Wikisource), internal-style wikilinks are (mercifully) converted from forms that work at those multilingual wikis (say, [[Wp/din/Foo|foo]]) to forms that work on the new standalone wiki (in this case [[Foo|foo]]), but this conversion is essentially ad-hoc and not automatic — and, as I've discussed elsewhere, it doesn't always give the best results (in this case [[foo]] would be best, assuming initial-case-insensitivity). I don't know if any other conversions are attempted, but some other examples of widespread poor wikicoding practices I've seen in newly-imported wikis include:

• external-style links that should be normal internal wikilinks or interlanguage/interwiki links;
• level-one =Section headings= in article content;
• other kinds of improper section nesting (e.g., using level-3 headings without level-2, etc.);
• article text beginning with a section heading (or un-marked-up isolated line of text) that duplicates the article title (I can see why people might want to do that while the content is still in the Incubator, but it's not appropriate once it's at the standalone wiki);
• extra whitespace that makes no difference when page is rendered (e.g., whitespace at the end of lines [except inside templates, of course!], multiple spaces or tabs between words or sentences [sometimes dozens of them!], etc.);
• [[File:]] links that specify both "thumb" and "right" ("thumb" implies "right", so the latter parameter is redundant);
• [[File:]] links that hardcode the sizes of thumbnails in pixels, sometimes at ridiculously large sizes (since hardcoding thumbnail sizes circumvents user preferences, it's usually better to use the default size [by not specifying one], or to use the "upright" parameter to rescale the thumbnail relative to the default size);
• (Etc.)

My point is, it would be great if some basic wikicode cleanup (or at least the creation of some list[s] of detected problems) were incorporated into the importing process somehow.

I know these items are all over the map in terms of who could actually address them (in particular, the last point about wikicoding should probably [start to] be addressed at the Incubator itself, before the content gets imported), but I just wanted to lay this all out in one place. Perhaps some additional instructions are needed at pages like Language committee/Handbook (requesters) (and maybe Template:New wiki request) and/or incubator:Help:Manual (and equivalents at Wikiversity Beta and "Old Wikisource")?

Any thoughts from users who regularly are involved with new wiki requests and/or creation? - dcljr (talk) 09:30, 4 August 2017 (UTC) Reply

When we will have Wikipedia in Montenegro language. That differs from Serbian. 'Gorski Vijenac' is written by Njegos. Language has 2 more letters. Language is spoken by 232.000 people. For thr formation 3 people voted in favor, while 2 were against. I would ask some of Wikimedia team to create Montenegro Wikipedia. Sonioa (talk) 00:17, 5 August 2017 (UTC) Reply

@Sonioa: If you really want to "ask" someone, ask langcom@lists.wikimedia.org, but you're unlikely to be able to complete your answer. --Liuxinyu970226 (talk) 22:20, 4 August 2017 (UTC) Reply

Anyway @Amire80, GerardM, Jon Harald Søby, SPQRobin, and Satdeep Gill:@Millosh and MF-Warburg: I don't know how to judge if this user is asking the actual "Wikimedia team" or not, or just, is this request likely to be rejected or verified? --Liuxinyu970226 (talk) 12:05, 5 August 2017 (UTC) Reply

I don't understand your question. --MF-W 17:19, 5 August 2017 (UTC) Reply

@Sonioa: See Talk:Language proposal policy#Why ISO discrimination?

Let's be really clear. At the present time, Montenegrin does not have a language code; it's considered a variation or dialect of Serbian. Now, that may or may not be accurate, or what you want to hear. But for right now it's reality.

The first thing that the Montenegrin community needs to do if it really wants its own Wikipedia is to apply to SIL for an ISO 639–3 language code. So far I have not seen any evidence that there has ever been such an application—not by the community here, not by the Montenegrin government, not anyone. If you apply for such a code you might or might not get it. But if you don't apply for a code you'll never get it. So go apply for a code.

• If you get a code, then you can start working on a Montenegrin Wikipedia test at Incubator.
• If you apply for a code at SIL, you can temporarily start working on a Montenegrin Wikipedia test at Incubator; we'll give you a temporary code then. But we won't do so if there isn't even an application pending at SIL.
• If you apply for a code and the application is denied, there may be a small possibility that LangCom will allow this anyway. But as I am not a member of LangCom, I can't guarantee they will do that. But I can pretty surely guarantee that if you don't even try to get a language code, there is no chance LangCom will allow a Montenegrin wiki. Is that clear enough? StevenJ81 (talk) 13:53, 7 August 2017 (UTC) Reply

I see that SIL feels this would fall to an ISO 639–2 code. Not sure why that should be. You'll have to wait until there is a formal answer from the Library of Congress. StevenJ81 (talk) 03:31, 8 August 2017 (UTC) Reply

Why wait? The last modification of 639-2 from LoC is adding Standard Morocco Tamazight (zgh) which is happened 5 years ago, and maybe somewhat dead. I don't think we really should open this request for more than a decade. --Liuxinyu970226 (talk) 12:47, 9 August 2017 (UTC) Reply

I also suggest to decline this request as nothing happened in LoC, not surprise, the LoC is a lazybone of America. --111.30.229.38 02:22, 14 August 2017 (UTC) Reply

Good point, ah I'm not possible to send email to langcom mailing list as my E-mail provider don't allow such sending. --Liuxinyu970226 (talk) 23:57, 18 August 2017 (UTC) Reply

Cf. #Notification about proposed approvals

Hello. The Langcom intends to approve Fiji Hindi Wiktionary. If you have objections to that based on the language proposal policy, please say so on this page. --MF-W 19:15, 11 August 2017 (UTC) Reply

Having the Hungarian and Japanese Wikivoyages marked as "approved" in the table at Requests for new languages#Wikivoyage and yet having them just "sumbitted" (i.e., status = open) on the individual request subpages is a bit confusing. Should the table entries and subpages be changed to waiting, instead (with a comment= by a langcom member explaining the situation)? - dcljr (talk) 00:48, 18 August 2017 (UTC) Reply

@Dcljr: Note: Requests for new languages/Wikipedia Innu-aimun, Requests for new languages/Wikipedia Khowar and Requests for new languages/Wikipedia Tabasaran are likely (marked as Verified as eligible on RFL list but still submitted on their subpages). --Liuxinyu970226 (talk) 14:18, 18 August 2017 (UTC) Reply

@Dcljr, Liuxinyu970226, and MF-Warburg: I can't speak to those Wikivoyages, which fall under a special transitional rule that I don't completely understand. As for the three that Liuxinyu970226 mentioned:

• I boldly went ahead and edited Requests for new languages/Wikipedia Innu-aimun to read "eligible", based on the fact that MF-Warburg actually wrote a comment to that effect on the page.
• With respect to Requests for new languages/Wikipedia Tabasaran: Liuxinyu, you added that entry yourself, with a status of "eligible", at this diff: Special:Diff/16256268. Unless you actually saw this discussed in the LangCom discussion group around that time, you should have inserted it with a status of "open". I'll let you decide what the right thing to do there is.
• As far as Wikipedia Khowar goes, I'm not going to rehash the whole situation there. I suspect that MF-Warburg didn't really intend to change the status all the way back to "discussion". At the same time, the general frustration level at the time was such that when a different user edited that page to status "eligible", MF-W reverted him, because he didn't want to risk an edit war. So that's where it stands now. I suspect that the status should really be "eligible"—it's a real language, with a language code, and a reasonable population of speakers. But to avoid stirring up old problems, I'm going to let MF-W decide for himself how he wants to handle it. StevenJ81 (talk) 17:10, 18 August 2017 (UTC) Reply

I've reset the Tabasaran one, no idea why I made such mistake when digging unlisted requests in that time. --Liuxinyu970226 (talk) 23:54, 18 August 2017 (UTC) Reply

Requests for new languages/Wikisource Hindi has been in limbo since 2008, while many other Indic languages now have their own Wikisource... we need a Hindi Wikisource for the hundreds of texts in s:mul:Category:Hindi and s:mul:Category:हिन्दी. Aryamanarora (talk) 15:56, 20 August 2017 (UTC) Reply

@Aryamanarora: I'd urge you to combine those categories into s:mul:Category:Hindi so that LangCom members can get a better sense of the overall picture of this test.

Additionally, for projects to be exported from their incubators (whether from Incubator, or from Multilingual Wikisource or from Beta Wikiversity), there needs to be a regularly active community. That means that every month (for a minimum of three months) there are at least three editors who make at least ten edits each in the test project. Since Hindi Wikisource is divided into two categories, it's not clear to me whether you actually reach that level now, but it doesn't really look like it to me. So please try to pull together a community to contribute regularly, and then LangCom can look into approving the project. StevenJ81 (talk) 15:55, 21 August 2017 (UTC) Reply

@StevenJ81: Let me say it again: since 2008. 9 years. Nevertheless, I'll see what I can do once I'm on my computer. Aryamanarora (talk) 16:00, 21 August 2017 (UTC) Reply

@Aryamanarora: Sorry, but I'm afraid that mostly doesn't matter. The concern of WMF and LangCom is that when it approves a project, the project will continue to have active participation. And it judges that by whether the test project is currently active, not that it's been out there for 9 years. (Note, also, that WMF and LangCom do not want to approve a project which is primarily the work of a single individual. That, too, helps explain this "community" requirement.)

My guess is that because Hindi Wikisource test has been out there 9 years, LangCom will give you every benefit of the doubt. But there still needs to be an active community involved. StevenJ81 (talk) 16:05, 21 August 2017 (UTC) Reply

@StevenJ81: Thanks for the help, I'm just a bit frustrated with how long it's been. It'll be easy to get 3 contributors I'm sure. Aryamanarora (talk) 16:17, 21 August 2017 (UTC) Reply