Talk:Interproject Babel template standardization (proposal)

This is the talk page for the Babel template standardization proposal. For the actual proposal(s), see the content page. Please keep all discussions related to the proposal on this talk page.

For your information

http://meta.wikimedia.org/wiki/Multilingual_MediaWiki Thanks GerardM 19:06, 17 January 2006 (UTC)

Of particular relevancy it will be for the WiktionaryZ project, the idea of a professional level is really tempting for it will destinguish among others professional translators. Now in a way having people state that they are professional without stating the profession does not make much sense. The relevancy of a profession is also to do with their specific vocabulary. GM

• I agree, there should be at least two different ones, for writers and for copyeditors, both preferably outside the numbering system of Babel: meaning you would put en-5 on your userpage to indicate you are near-native and an additional template to indicate your occupation as "professional writer". The specific vocabulary-issue (technical terms) would open the door to lots of more templates ('user can translate psychological terms', biological terminology etc.) - maybe we should stick to the good old fashioned list-principle that we have now.--Fenice 19:35, 17 January 2006 (UTC)

I'm for the six levels. en:User:Nightstallion, 21:31, 17 January 2006 (UTC)

Stick to the ususal pattern of listing alternative proposals.

I've refactored the proposals in order to keep them separated into more or less independent parts. For example, all proposals made so far have included an xx-0 level, so I've split that one into a separate (and hopefully uncontroversial) subproposal. Similarly, I've split off the subproposals to rename the project and to translate the templates, since I see no reason to tie these into the debate on the number of levels to include. My goal with this is to have each subproposal address a single issue that we may actually discuss and eventually vote on without people having to pick-and-mix pieces from different alternatives.

If you feel I have "watered down" your proposal in the process, please go and edit it to make it say what you want. But do try keep the subproposals as independent as possible, otherwise this can easily turn into an administrative nightmare that will never gain consensus on anything. --Ilmari Karonen 00:42, 18 January 2006 (UTC)

No, Ilmari, no way. If you want to discuss and promote your proposals do so on relevant talk pages. This main page will stick to the usual rules of listing alternative proposals. I am going to add the usual numbering now in order to make that clear, even though I personally believe in giving each proposal a name instead of a number. --Fenice 06:52, 18 January 2006 (UTC)

I've been trying to follow the advice at en:Wikipedia:How_to_create_policy#Guidelines for creating policies and guidelines, specifically point 5 as (unofficially) elaborated at en:Wikipedia:How to hold a consensus vote. But if you insist otherwise, and given how few other people have boithered to comment so far, I'm willing to let the matter be for now. I'll reformat my proposal to match yours in style instead. --Ilmari Karonen 15:14, 18 January 2006 (UTC)

[Partially written before Ilmari Karonen's post above, in response to Fenice.] What? Ilmari Karonen's goal was, unless I'm missing something, to "have each subproposal address a single issue that we may actually discuss and eventually vote on without people having to pick-and-mix pieces from different alternatives". What's wrong or unfair about this? (And, more specifically, how is it not perfectly in line with point 5 at en:Wikipedia:How_to_create_policy#Guidelines for creating policies and guidelines?) I really don't understand your complaint. EldKatt 15:17, 18 January 2006 (UTC)

Babel box style

I like the babel boxes at w:User_talk:TarmoK#bilingual_babel_boxes. Gerard Foley 07:11, 18 January 2006 (UTC)

User's language skills (see Babel):

et Emakeel on eesti keel. Native speaker of Estonian. en-3 Advanced level of English. fi-2 Kohtalaisesti osaa suomea. Intermediate level of Finnish. es-1 Nivel básico de español. Basic level of Spanish. ru-1 начального уровня на русском языке. Basic level of Russian.

Not a bad idea at all. I wouldn't oppose this, but I still do think this should be made an independent (sub)proposal, since it doesn't really depend either way on the number of levels chosen. --Ilmari Karonen 15:40, 18 January 2006 (UTC)

Bilingual boxes are more useful to foreigners than most people think: even someone who doesn't speak a language himself might still be able (or need!) to tell language A from language B in print. --Ingeborg S. Nordén 14:55, 2 August 2006 (UTC)

Curious - no 4-level proposal so far...

It's somewhat curious that no-one's proposed a system with levels 0, 1-4 and N yet, given that this seems to be the most widepread variant in actual use at the moment. (I actually did add a four-level alternative while refactoring the proposals, but it got reverted.) Is everyone here unanimously agreed that this particular system is no good, or has no-one just got around to proposing it yet? --Ilmari Karonen 15:49, 18 January 2006 (UTC)

Probably the latter. I think it would be a bad idea to leave it out. Of course, it would be a lot easier to deal with if we split the proposal up into its various aspects, as you tried to do earlier. I still don't see why this is not a good idea. EldKatt 15:55, 18 January 2006 (UTC)

I have restored some content, including the 4-level proposal, from an older version of the page.

The 4-level system does not seem so widespread to me. Commons and Meta both use 3 levels. Because en.wikipedia and en.wikiquote nominally have a level 5, I have not yet found a project using 4 levels other than en.wiktionary. --Kernigh 05:26, 24 January 2006 (UTC)

Off the top of my head, at least de.wikipedia and fi.wikipedia used 4 levels. Maybe fr.wikipedia too? Also, en.wikipedia had 4 levels until less than a month ago. --Ilmari Karonen 12:26, 25 January 2006 (UTC)

Just checked, and the French, German, Russian, Chinese and Japanese Wikipedias all list 4 levels (although at ja.wikipedia only a few xx-4 templates exist). Those are all pretty large projects. On many of the pages, however, it's clear that the fourth level is a relatively recent addition, with some parts of the page still referring to a three-level system. --Ilmari Karonen 12:39, 25 January 2006 (UTC)

Measuring proficiency

I agree with suggestions to keep specialised kinds of language expertise out of the Babel system. I am a professional interpreter but prefer to say "advanced" for my second language rather than "near native", for a number of reasons (see this discussion), and it would seem strange to be level both 3 and 5!

One of the problems with both the second and third proposal is while "native" is on top, lower levels contain definitions such as "Advanced — though you can write in this language with no problem, some small errors might occur." This one-dimensional numbered system fails, as most native speakers I know make "small errors" when writing; some are atrocious. Native speakers who are 12 years old, or whose socio-economic/cultural background hasn't emphasised writing and education, or who are just bad spellers, are going to be (nonsensically) both above and below this so-called "advanced" level. These definitions really describe one's ability to write to a particular standard, not their fluency (or nativeness) in speech. Proficiency in language can take so many different forms that a one-dimensional system that tries to grade people into ever-narrower categories becomes meaningless — one person may be a perfect speller and never make a grammatical error, but use convoluted sentence structure and produce dry prose. Another may be a sloppy speller and make mistakes, but their writing sparkles.

So I support the first proposal, although I suspect too many people are already using the xx-4 templates for us to get any support deleting them. I strongly believe that the four levels cover everything we need to know for practical Babel purposes, and the addition of more (and more!) levels is doomed to fall over just like the eponymous tower. A highly-gradated system breeds competitiveness rather than community spirit, and unfairly forces those who would prefer to keep things broad to judge and rank themselves against each other. Editors who desire more detail than four simple levels, however, can always add additional user boxes ("Spelling Bee Champion"), or user-page/talk-page descriptions. Ntennis 03:57, 20 January 2006 (UTC)

I agree with everything in this comment—well, maybe almost everything. I strongly think standardization across projects and languages is needed. I support a system that keeps things simple while maximizing utility. Designating between advanced, near-native, and professional levels of fluency is silly. It doesn't help build the projects, it just allows users to categorize themselves into ever-smaller self-defined constituencies. That may be appropriate for unofficial user templates (or may not—I think each project should decide for itself) but it is clearly not useful to the goals of the Wikimedia Foundation, and somewhat elitist beside (whatever that counts for—I think we can fairly say it doesn't foster a sense of community).

I would clarify, though, that language levels other than N are non-native, e.g. "Non-native fluency in Russian at intermediate level" to clarify that the purpose of Babel templates. I don't feel that self-classification of one's proficiency in one's native language(s) is helpful; for second languages, however, this is appropriate.

I feel it would also be wise to clarify that the Babel templates indicate a user's ability to contribute to the project at a certain level of proficiency. Again, the templates' utility to the project is only for measuring skills in contributing, not rating someone's literacy or fluency. For instance, in general I speak Spanish at an intermediate level of proficiency, but I certainly could not contribute to Wikispecies in Spanish at an intermediate level. On the other hand, someone with a lesser proficiency than I in Spanish may be able to contribute to Spanish Wikispecies at a higher level of proficiency, if they're familiar with the specialized vocabulary that project deals with.

Also: in addition to the "four simple levels" to which Ntennis refers (which I assume to be 1-3 and N), I think level 0 is useful. Imagine for example a user who enjoys contributing images. S/he could visit a Wikipedia article in a language they speak and contribute an image to that page; s/he could also follow the interwiki links to that article in other languages and contribute the same image, without any understanding of those languages. It would be very useful in this case to have a template that says, "Hey, I don't speak your language, so if you need to talk to me, unfortunately I can't understand what you're saying unless you say it in one of the languages I speak." Communicating this without a template is tricky because the user would have to know how to say "I don't speak Croatian, I only speak Catalan" in Croatian, a problem which of course is multiplied for every language to which the user contributes. The other useful situation would be e.g. a user contributing to en.wiki who states on their user page that they live in or are from Japan, or who edits a lot of articles about Japanese subjects; another user on en.wiki tries to communicate with them in Japanese, not realizing that the user does not speak Japanese. This could also happen within a country, e.g. a user on fr.wiki states that s/he is from Switzerland; another user tries to communicate with him/her in German, not realizing s/he does not speak German. So there are many cases where level 0 could be appropriate and useful. Of course, overuse of 0 templates would be silly and pointless, and should be discourages, but I don't feel this should be a major concern in the overall consideration of level 0 templates.--Tetraminoe 05:43, 25 January 2006 (UTC)

You are right about the level-0 templates; also, defining levels 1 to 3 as explicity non-native is a great idea. So now that we agree, let's go implement the NEW WIKIPEDIA POLICY! Only joking. Seriously, though, where is this proposal going? How does it get beyond the proposal stage? Ntennis 01:16, 7 February 2006 (UTC)

0-5 plus N is fine

I'm heartily in favor of a system that distinguishes native speakers from foreigners who can still write fluent, coherent articles about a specialized topic: the latter would count as xx-5 fluency IMO. However, I also propose that all levels be available for all modern languages, not just three or four prominent ones. I have defended the need for other xx-5 Babelboxes on the main WP talk page for this project (copying the same text here might be considered spamming). --Ingeborg S. Nordén 14:44, 2 August 2006 (UTC)

ILR-based de facto?

Does anyone disagree that we should use the ILR system? I think that there must be categorization-by-linguistic-ability systems out there, which are more precise than ILR. I strongly believe that ILR was chosen in English Wikipedia, because of it's national oigin (US). And its use then spread to other language projects, by the quick-copying-templates habit and the fact that the English Wikipedia was the First Of All :-) Dead3y3 02:21, 11 October 2006 (UTC)

What is the "ILR" system ? GerardM 11:55, 11 October 2006 (UTC)

See w:en:ILR scale (you are right, I should tag it as "scale", not "system. My mistake). It is also referred in the article page here. --Dead3y3 22:24, 12 October 2006 (UTC)

If ILR-based, then 0-5 (ILR + 0 for project lang)

I think that the best approach is to follow the ILR, on which our sytem has been based. Since the case is that we haven't made an independant redefinition of the levels, the fast and the precise to the existing level definitions choice is the ILR itself.

Or else we'll have to redefine the levels (down-to-top approach). But I think that this is impractical; it requires linguistic professionality by the majority of us; we will have to understand the slight differences between different language schemas, or this discussion will have 3-4 specialists as the only participants, and propably will be abandoned.

The N Level is by definition the ILR Level 5. To minimize possible reactions from users, Template:User xx should be a Redirect to Template:User xx-5, in all projects. And also the color should be changed from the current red to the xx-N green (I don't know if it is the case for all projects except en.wp).

The 0 Level should be strictly restricted to the project language. The use of it to the so-called de facto international laguages will result in a mess. De facto is not in the vocabulary of many many people, who will start to react by "who says that this lang or that lang is de facto international?". And propably the reaction will result in hundreds of 0-level temps, or in the depart of hundreds of frustrated users. I don't want to scare anyone. I simply think that this is what will happen in that case. --Dead3y3 02:21, 11 October 2006 (UTC)

There is a huge gap between professional users of a language and native people speakers of a language. I have had the pleasure of living in the UK for several years and my English proved to be technically better than of some of my colleagues. Professional means that you get paid for working with specific aspects of that language. It does not say necessarily anything about the quality of the language. People who indicate there professional use of a language claim some specific expertise.. GerardM 12:00, 11 October 2006 (UTC)

In WiktionaryZ we support ISO-639-3 exclusively with levels 1 to 5 and N. GerardM 12:00, 11 October 2006 (UTC)

I know this and I agree. There is a huge gap.

See w:en:ILR scale (I also posted the link above). I think that it classifies the professionality in a language as Level 4, and the native (or near-native) as Level 5. So the distinction between the two linguistic ability categories is already there. All we have to do is redefine our correspondency between the Level number and the ability at a language, to fit the ILR. --Dead3y3 22:31, 12 October 2006 (UTC)

merge "native" with "highest level"

I added a sub-section here (diff history link)

Esteban.barahona 19:06, 27 August 2008 (UTC)