On Fri, Sep 5, 2008 at 12:03 PM, Milos Rancic <millosh(a)gmail.com> wrote: Hm. The answer is at the talk page: "Removal of the ISO 639 requirement. Technically, all that is needed is an RFC 4646 code (like zh-min-nan). The language code identifying the localisation is traditionally RFC 4646, as is the HTML "lang" attribute. And there is no reason why multiple wikis can't share a language code -- there are plenty of language group codes which provide for miscellaneous minor languages and dialects. They just have to have a unique domain name. The domain name does not have to be the same as the language code."

On Fri, Sep 5, 2008 at 1:02 PM, Tim Starling <tstarling(a)wikimedia.org> wrote: And one more question, just to be sure: Internet standards (like HTML/XML language tags or so) are using RFC 4646 codes, not ISO 639?

On Fri, Sep 5, 2008 at 2:36 PM, Tim Starling <tstarling(a)wikimedia.org> wrote: Then, I see no reason why to stay with ISO 639 codes. RFC 4646 codes are more flexible (=better) and Internet standards are based on those codes. The only issue which I see is: Do we have some problems with language codes which mean one language in ISO 639 and a different one in RFC 4646?

Gerard Meijssen wrote: You are making a requirement that a language proposal must have a *unique* ISO 639-3 code. That is much more restrictive than what I am suggesting. I'm saying that we shouldn't need a unique language code of any kind. All we need is a domain name and something to put in the lang attribute. As far as I'm concerned, the lang attribute can be "und", and the domain name can be a 7-letter transliteration. Both can be changed at any time after the creation of the project. We've been ahead of the standards bodies in the past, and we've moved wikis when they've caught up. I see no reason why we can't do it again. -- Tim Starling

Gerard Meijssen wrote: No, on meta I removed the ISO 639-3 requirement entirely. I replaced the ISO 639 code in the request section with "RFC 4646 code (if available)". In another post: No, I'm not suggesting that we use RFC 4646 in the same way that you have been using ISO 639-3. I'm suggesting that we make our own decisions on which languages should be included, independently of any standards body. An RFC 4646 code is useful, if there is one, but it shouldn't be necessary. -- Tim Starling

Gerard Meijssen wrote: Or another Egyptian Arabic? Everyone makes mistakes. It's not a problem if the decision is easily reversed. The community let us know what the problem with Siberian was, and we deleted it. No big deal. -- Tim Starling

On Fri, Sep 5, 2008 at 7:44 PM, Gerard Meijssen <gerard.meijssen(a)gmail.com> wrote: Gerard, I love you as a friend, but you are really pain in the ass :) When you feel comfortable in linguistic area, you use those arguments; when you feel comfortable in Internet standardization area, you use those arguments (as you convinced me in private conversation a month or two ago that standards like HTML are using ISO, not RFC/BCP codes), and, finally, when you feel comfortable in politics, you use political arguments. So, please, if you want to, let's say, use linguistic arguments, then just try to think a little bit out of the scope of your knowledge -- yes, there are some people who know about languages better than you. And I am sure that they would be glad to help to Wikipedia. If you have a problem with asking, I may ask them. And, if you want to use other arguments, please, use them consistently. In this particular case, there are two options: (1) To use political methods or (2) to use expert methods. If you don't want to ask for expertise some university, then you have to ask the community. In both cases LangCom is not an untouchable quasi-political-quasi-expert body, but a body which gathers expertise to support the community. Community is not here to support LangCom (as well as community is not here to support, let's say, stewards -- stewards are here to support community).

On Fri, Sep 5, 2008 at 8:21 PM, Gerard Meijssen <gerard.meijssen(a)gmail.com> wrote: ISO is a political organization, not an expert (in this case, a linguistic) one. Relying on one political organization, when we are talking about science, is stupid. I don't say that they are useful, but it *is* stupid to follow blindly one political organization. Fortunately, they are not the body which declares standards on Internet. And, fortunately again, IEEE is an expert body. As well as I am sure that they are reasonable enough to accept an advice which Wikimedia would give to them. There are some basic linguistic *facts* which ISO doesn't assume. At the area of Serbia, Montenegro, Bosnia and Herzegovina and Croatia there are four different linguistic systems. But, only one covers all three-four ISO codes: Shtokavian. Other three: Chakavian, Kajkavian and Torlakian don't have their ISO codes, but they are much more distinct from all five standards (Serbian, Croatian, Bosnian, Montenegrin and Serbo-Croatian). (So, yes, linguistically speaking, there should be four Wikipedias at the linguistic area of those countries, but not Serbian, Croatian, Bosnian and Serbo-Croatian, but Shtokavian, Torlakian, Chakavian and Kajkavian.) Also, I would like to ask you which Hainan (a Chinese island, south of the Mainland) languages you would consider as possible Wikimedian laguages? Let's say, two persons from their Tai-Kadai group came and ask for two different projects? No ISO codes -- no Wikipedia? It is really highly bureaucratic measure to deny to, let's say, half of million of humans to have a project in their native language just because official China is, also, a bureaucratic machinery which recognizes just "Li people". (Fortunately, Li peoples were supportive to communists and they are well protected now; except from bureaucracy.) There are a lot of such problems all over the world. While LangCom made regulation in relation to proposals like Siberian or Zlatiborian are, it is not there to forbid new projects because of some bureaucratic rules, but to help to old and new Wikimedians with expertise. When someone from IEEE wants to see what is a language, I am sure that in 99% that person goes to Wikipedia :) So, I am sure that we have enough of influence on IEEE and its auxiliary bodies :) Actually, I said that your (singular) arguments are inconsistent; I didn't say that your (plural) positions are. It is reasonable to expect that your positions may be changed, while I am not so convinced in change of your arguments ;) (Unlike an ISO politician, which would ask, firstly, their private interests, then their national interests, and after them, some old academic who would say something about, let's say, dialects gathered in the 19th century which describes the situation from the 18th one.) Where I am sure that I am right (while you are sure that I am not :) ) is that LangCom have to take a clear position: if you are willing to be a mediator between the community and experts, you should behave like that; and if you are willing to be a political body, you should behave like that. In the first case (and I am sure that LangCom's intention is this one), ISO is far from any kind of relevant institution for language diversity. Sorry to say, this is true for SIL, too. English Wikipedia articles are much more relevant than both of those organizations. (BTW, en.wp articles are more relevant than a couple of linguistic encyclopedias which I read, too.) In relation to ISO and SIL you may be sure only that there is some name in some area which some persons use to name their language.

I think we're all tired of the rules lawyering. 2008/9/5 Milos Rancic <millosh(a)gmail.com>om>:

On 9/6/08, Mark Williamson <node.ue(a)gmail.com> wrote: It was retired before ISO 639-3 was published, and even if it weren't it would fail the requirement that it "be sufficiently unique that it could not coexist on a more general wiki". Identifier: yib Status: Retired Retirement remedy: Merged into [eng]; entirely intelligible with English. http://www.sil.org/iso639-3/documentation.asp?id=yib -- Yours cordially, Jesse Plamondon-Willard (Pathoschild)

Hoi, If you were to request a Wikipedia for Yinglish, it would be eligible for a community to start this project and be considered. When a substantial corpus has been produced, we will ask an expert if it is indeed Yinglish. If we can not get this from an expert WE select, we will not accept what you call Yinglish. Consequently we already do as you suggest; we check not on something theoretical, "does this language exist" but we test based on something practical "is this indeed the language you use the code of". Thanks, GerardM On Sun, Sep 7, 2008 at 4:52 AM, Mark Williamson <node.ue(a)gmail.com> wrote:

2008/9/5 Tim Starling <tstarling(a)wikimedia.org>rg>: Have we merged Serbo-Croatian yet? The problem is that such wikis tend to be serious nationalist POV forks which does nothing for our reputation in the relevant areas or get the communities working together long term. -- geni

On Fri, Sep 5, 2008 at 7:29 PM, Gerard Meijssen <gerard.meijssen(a)gmail.com> wrote: It is not so hard to ask for expertise. One of the tasks of LangCom is to ask for expertise if members of the committee are not sure. And if three universities, for example, say different things, then LangCom may decide without problems. I am sure that there is no one university which would say that Siberain or Zlatiborian are languages.

Gerard Meijssen wrote: The BCP 47 series has always used the most recent version of ISO 639. And RFC 4646 makes several references to reserving space for future developments in ISO 639. And it reserves all three-letter codes. RFC 4646 predates ISO 639-3. I think we can use a defacto merger of ISO 639-3 and BCP 47 while we are waiting for the next RFC to come out. In fact, that is exactly what we have been doing with MediaWiki language codes since the release of ISO 639-3. For Internet protocols, it's not a choice of one or the other, BCP 47 is the only game in town. -- Tim Starling