Archives/2007-03-23

From LangCom

This page is no longer maintained and may be outdated: moved to m:Special projects subcommittees/Languages/Archives/2007-03-23.
Archived discussion
This is an ongoing email discussion about the Kabyle Wikipedia and localization.


1.

Jesse Martin (Pathoschild)
23 March 2007 09:59


Hello,

I propose the subcommittee approval of the Kabyle Wikipedia. This language is well-established, has a very wide audience, a large interested community, a unanimous discussion on Meta, and a standard code. In addition, the test project is very successful with over 200 pages and daily activity by multiple users.

Although the project is not localized due to the technical difficulty of localizing without a wiki, the community is more than strong enough to localize with Special:Allmessages once the wiki is set up; there's no reason they would use an English interface. This translation can be added to the localization files by the time a second Kabyle project is approved. This opinion was seconded by Aphaia, member of the translation subcommittee.

Yours cordially,
Jesse Martin (Pathoschild)

2.

Berto 'd Sera
23 March 2007 23:10


Hoi!

> the technical difficulty
> of localizing without a wiki,

Pls explain. What's more difficult for them in respect to any other linguistic entity? Why can’t they use Betawiki?

I'm quite against making *any* exception whatsoever. If we allow a wiki without localization (no matter which one and why) we either allow them all or promote the idea that LangCom makes arbitrary decisions based on personal sympathy and internal mafia.

We may mean good but that's the way anyone would interpret such a move. We just canceled all pending applications and keep on hold tens of requests because they miss a localization file, make 1 + 1 and you'll have the meaning that the community will attribute to any single exception...

Do we need that?

Bèrto ‘d Sèra

3.

Jesse Martin (Pathoschild)
23 March 2007 23:52


Hello,

There is no exception to make; we never agreed to localization as a prerequisite and it is not part of the policy.

BetaWiki is complex to edit and is not affiliated with the Wikimedia Foundation. This method requires that editors register to a third wiki (Meta, Incubator, and Beta-Wiki), explicitly define required preferences, acquire special permission to translate the interface (the "translator group"), and learn to use the translation interface. Every translator must do this; since the only instructions for this complex procedure are in English, bilingual users will need to translate the instructions to the relevant language for those who aren't. Furthermore, users cannot translate important wiki-based pages such as the sidebar.

On the other hand, translating live consists of going to "Special:Allmessages", clicking a redlink, and filling in the translation.

If BetaWiki were merged into the Incubator as an extension, it would be much more workable; users could translate the entire wiki simultaneously. However, it is not very workable as a separate, unaffiliated, complex procedure on a personal test wiki.

Yours cordially,
Jesse Martin (Pathoschild)

4.

Berto 'd Sera
24 March 2007 12:51


Hoi!

Okay, let's try and see some light in the tunnel :)

1) I maintain PMS in betawiki myself and made its first interface by editing AllMessages myself, single-handed. It takes but a couple of days in all. BTW, I use a 16kbits dialup on a very aged soviet telephone line. You don't need a broadband connection to use goodwill and there's nothing complex in reading an expression and writing an equivalent in your native language. If even the translator cannot understand an expression, then HOW do we expect a not bilingual community to be able to use it??????????
2) Mandatory localization SHOULD be in the policy (it was in the draft we first published and approved, if this was changed I never voted to allow its deletion). I'll never give green light to a project that is not localized (just as won't accept bots on the incubator to fake a community where there's none).
3) The fact that betawiki is outside wmf is irrelevant. Volunteers aren't part of wmf either, and we don't feel disgusted by what they do. If we were to accept only internal stuff then no public translation interface should be available, we would have the whole job done by paid translators in Micro$oft $tyle. But that's not our case.
4) It's 100% true that betawiki misses the extensions, yet it's just a couple of things (like special chars) we are talking about, and the sidebar, which is better decided by community discussion anyway. There's nothing weird in using betawiki (try and make a PO translation for a linux interface) and ALL translation processes (and decisional life at meta level) are documented in english anyway.
5) A community that has not at least one user who is fluent in english and willing to represent it at meta-level is going to become autistic. We may not like it, but it's a PLAIN FACT. We have more than enough political wikies maintained by hallucinated sects to deal with.
6) I can't see what's the problem in having one more user in another wiki, people do not pay to register a user in Betawiki, afaik. If we have a procedural problem with Betawiki (like giving people a translator role) we should address it as general procedural problem, not as an act of delicacy towards langcode ZXX. Rules are rules; we cannot act as emperors whose thumbs may save a life or terminate it.
8) I take it that for smaller linguistic communities this may well be the first impact with the technology. Yet this didn't keep those users to show up in the incubator. I'm myself from a linguistic community of less than 2 millions native speakers (mostly elderly people) with a 2% alphabetization rate. We could make it, they can make it.
9) I will strongly support any community action directed to help people learn. It's not just empty words. I have almost no spare time, but if needed just give them my email and tell them to make any request for explanation they may need at any time. I'll be happy to help with minimal requirements, but I won’t change my vote to make anyone's life easier, including mine.

In short, I officially oppose the creation of ANY unlocalized wiki. You can record this opposition vote as unchangeable for any further such request.

Bèrto ‘d Sèra

5.

Jesse Martin (Pathoschild)
24 March 2007 01:11


Hello,

As far as I can tell, none of the policy drafts at <http://langcom.wikimedia.org/wiki/Language_proposal_policy> have ever required localization. What do the other subcommittee members think of requiring full localization before approval?

Yours cordially,
Jesse Martin (Pathoschild)

6.

Berto 'd Sera
24 March 2007 01:15


Hoi again,

Jesse , pls don't take my position as a personal will to blunt lock-gates against any new project. As a matter of fact I'm spending a huge number of hours in translating wikipedia articles in rare slavic languages that are full of political assertions that (if translated and published) would *seriously* damage our public image.

You don't want donors to know that their donations are used to publish revisionist history, a blatant usage of terms that equivalent to *nigger* in their racial derogatory value, etc, etc. So what many of us do is keeping a monitoring eye on as many languages as possible and eventually get to the point in which some Committee will decide of the most dangerous events.

It's quite true that having a real community is not a warranty about the neutrality of a wiki, yet as a matter of fact most of the weirdest wikies happen to be managed by a couple of overheated extremists. Maybe if we had a required a wider number of users for a starting edition this problem would have been reduced.

If we want users to take wikipedia seriously we must not make a toy of it since the very start. I'm positive you just want to help people, yet there is no good thing in life that one can get for free. Valuable things always come from hard work and a good wiki is no exception.

Bèrto ‘d Sèra

7.

Jesse Martin (Pathoschild)
24 March 2007 01:22


Hello,

Did you mean to send that last message to the whole subcommittee or only to me?

I agree with what you're saying; those are the problems I wanted to prevent when I wrote the policy draft a few months ago, and it is one of the reasons we require a successful test project before approval. However, I don't think localization of the interface itself before approval rather than after will affect the viability of the communities or the veracity of the content.

Yours cordially,
Jesse Martin (Pathoschild)

8.

Berto 'd Sera
24 March 2007 13:35


Hoi!

Yes, it was for all. I'm getting sleepy (8.35 am on the eastern front) and I simply landed on the wrong button :) I wrote a number of things about localization for small communities here: http://eng.i-iter.org/making-new-semantic-domain

As a matter of fact making a wiki interface is "inventing a semantic domain". That cannot be done by a lot of concurrent hands because it will result into a unusable jam. A community should really choose one or more interface editors and it should discuss about the language they "create".

Being immediately understandable to a native speaker who never really used technology is 90% of the marketing behind a small language wiki. We had our first 5 nation-wide press articles just because our literal translation of the term WEB sounded funny (actually I just picked it up from a friend as I found it funny, too). Our term "Ragnà" (web) prompted the first printed appearance of text in the piemontese language on the national press in 150 years, because then they moved to discuss our GNU license, etc etc.

We keep having press coverage mainly because the italian local languages do translate those english words that the italian interface imported as unchanged. It's not just a small procedural step we talk about, interface language is a big asset for a wiki, maybe the biggest asset a wiki can have because it defines how usable (or cryptic) that wiki will be.

Bèrto ‘d Sèra

9.

GerardM
24 March 2007 01:57


<this user has not agreed to public archival.>

10.

Sabine Cretella
25 March 2007 13:32


Hi, I just came back home: well localization is the basics ... and for Neapolitan I still have a hard time to get people to localize stuff in the right place: so new projects should avoid these problem in a first place. If they learn to maintain the localization directly on betawiki from the beginning they will always do it - if you allow for anything else ... you get projects that even after half a year are not localized. See the so-called Tarantino wikipedia: http://roa-tara.wikipedia.org/wiki/Pagene_Prengep%C3%A1le

If they have time to write the articles they should also have time to do the localization ... or am I wrong?

Having the localization done before a wiki is opened was always the goal.

Sorry for being short - have to put things away.

Ciao, Sabine

11.

Jesse Martin (Pathoschild)
27 March 2007 21:38


Hello,

There seems to be a consensus in favour of requiring localization, so I'll accept it. I've implemented it as a requirement in the language proposal policy. I thus propose the following procedure to streamline and simplify the Kabyle localization. This will be a good test case for future localization efforts.

  1. Set up a specific section on the request discussion page to answer questions, provide updates, and coordinate localization.
  2. Contact every user who has expressed an interest in contributing of this effort.
  3. Establish a group of trusted users who will coordinate the localization in the Kabyle community and form the initial group of administrators. Doing this will allow users to begin the wiki prepared, minimize conflict, and let users work together with the future administrators to ensure that they are a good choice *before* complex and conflictive desysop'ing is the only appeal.
  4. Set up a small number of subcommittee members to take official charge of this effort. This ensures that [a] guidance is always available, where normally subcommittee members might be busy elsewhere or forget about the effort, [b] it is easy for editors to find help, and [c] users are sure of our support and interest. Other subcommittee members would help as much as they desired, of course, but these few members would make it a point to be available as much as possible.

What do you think of this process?

Yours cordially,
Jesse Martin (Pathoschild)

12.

Jon Harald Søby)
28 March 2007 12:02


Well, after reading the slaughtering thread, I think there are some good arguments against demanding a localisation beforehand. Localisation is very hard work, and even I, who have been involved with wikis for more than two years now, in a lot of ways (as common user, anon, sysop, steward, bot operator), don't even know what some of the messages in "core" on BetaWiki actually mean, and have never encountered them.

What I suggest is that we make a list of the real core messages, perhaps around a hundred of them or so. What I suggest is that the main page, and all pages linked from it in the menus, etc, are to appear fully translated to begin with. Then, when they start the Wikipedia, they can translate the rest, if they feel like it. But it shouldn't be a requirement to have everything translated. (Not even Norwegian, which has more than 100 000 articles, have a fully translated MediaWiki message system; most of what's untranslated are those stupid and pointless Exif- messages, though, while the rest is translated.)

-- Jon Harald Søby
http://meta.wikimedia.org/wiki/User:Jon_Harald_S%C3%B8by

13.

GerardM
28 March 2007 04:24


<this user has not agreed to public archival.>

14.

Jesse Martin (Pathoschild)
28 March 2007 09:04


Hello GerardM,

I proposed the selection of a small number of editors based on the advice of Bèrto 'd Sèra earlier in this discussion:

"As a matter of fact making a wiki interface is "inventing a semantic domain". That cannot be done by a lot of concurrent hands because it will result into a unusable jam. A community should really choose one or more interface editors and it should discuss about the language they "create"."

Do you disagree with this advice?

Yours cordially, Jesse Martin (Pathoschild)

15.

Berto 'd Sera
28 March 2007 10:07


<this block of text is marked as private.>

EXIF msgs... they are pure bullshit, I'm lucky me and my dad have always played with cameras, but I really hated them when I translated them :)))) I'm also positive 100% translations can't be made (new messages appear/get changed almost every day). It's one more reason to have someone keeping a constant eye on the UI.

At least all commands should be translated (and the copyright statement, too, because it's a contract and end-users have the right to understand what they implicitly sign), but whether you translate EXIF msgs or the like of it IMHO is largely a matter of choice.

This should be a process by which a community "configures" its own wiki and learns how a wiki works and what admins are supposed to do with it. My best option would be an incubator mailing-list, in which translators and would-be admins could get to know each other and help each other.

Anyway... what if we stop talking theory and we start an experiment? We can use it to document the process and later make adjustments while we learn from a case of study. Let's begin this Kabyle translation, so we can address real problems as they come.

Bèrto ‘d Sèra

16.

Jesse Martin (Pathoschild)
28 March 2007 19:56


Hello,

I agree; localizing every last message is unnecessary. Perhaps we should determine the files to be localized in discussion with the Kabyle community; this collaboration will simplify the process for us, improve the accuracy and quality of our selection, and the involvement of the waiting community will alleviate impatience.

GerardM has stated that all users should translate the interface simultaneously; Berto has stated that they should select a few users to work on the interface. I prefer the latter, as I proposed above. Any other opinions? Are there any objections otherwise to my above proposal (contact the users, centralize discussion and collaboration)?

Yours cordially,
Jesse Martin (Pathoschild)

17.

GerardM
24 March 2007 01:35


<this user has not agreed to public archival.>

18.

Berto 'd Sera
29 March 2007 09:21 --- Hoi!


LOLOL Gerard, I didn’t expect a quote from Simon & Garfunkel J


Okay, let’s look for a convergence point:

  1. I’d like to have an interface in the community for us to be able to remain in contact with them while they localize (and not to give them the impression that we simply “forgot about them”).
  2. Nominating an interface is one more request and can sound as a limit for the community


I’d say we can start with no fixed organization, but just an informal interface. My point is that most of what we will fix as a “procedure” should come from practical experience, and it should be some sort of “endless blog” from the people who start-up the communities (not us, that is, but the native speakers involved in making new wikies). In the end it’s a sort of “entrance fee” we request. People get a wiki and pay for it by delivering a new localization for mediawiki software.


Bèrto ‘d Sèra

19.

GerardM
29 March 2007 09:42


<this user has not agreed to public archival.>

20.

Berto 'd Sera
29 March 2007 09:49


Hoi,


Sorry, what code are we talking about?


Bèrto ‘d Sèra

21.

GerardM
29 March 2007 11:32


<this user has not agreed to public archival.>

22.

Jesse Martin (Pathoschild)
29 March 2007 12:33


Hello,

There doesn't seem to be any opposition to centralizing discussion at <http://meta.wikimedia.org/wiki/Talk:Requests_for_new_languages/Wikipedia_Kabyle>. This is a pretty logical choice, so there's no need to nominate or vote on it, nor establish any codified procedure.

I don't oppose a free-for-all localization; this will be a good chance for the users to get to work together, and mistakes can easily be corrected by other editors. Berto, do you oppose this? There doesn't seem to be any support for assigning subcommittee members to help, so I'll just make sure to watch the discussion to make sure all their questions are answered.

If there's no opposition to a loose collaborative method (centralized discussion, free-for-all localization), I will announce that and get it going as soon as possible.

Yours cordially,
Jesse Martin (Pathoschild)

23.

GerardM
29 March 2007 12:43


<this user has not agreed to public archival.>

24.

Berto 'd Sera
29 March 2007 14:24


Hoi !


That’s good news J


Bèrto ‘d Sèra

25.

Jesse Martin (Pathoschild)
29 Mar 2007 14:34


Hello,

Alright then. If there's no opposition to the adjusted proposal, we'll begin tomorrow. This isn't a codified procedure, just a way we can proceed with the Kabyle Wikipedia.

1. Set up the request discussion page to answer questions, provide updates, and coordinate localization. 2. Contact every user who has expressed an interest in contributing of this effort. 3. Help any interested user begin localizing.

Yours cordially,
Jesse Martin (Pathoschild)

26.

Berto 'd Sera
29 March 2007 16:41


Hoi!

Okay for me, I would add a sort "contact structure". I mean, we speak different languages, I just started correspondence with the Ingush because we have a common russian, maybe there are other people who can volunteer their languages with other fallbacks.

The further we get from the West the more important it will be to be able to deliver assistance in non-english form.

Bèrto ‘d Sèra

27.

Jesse Martin (Pathoschild)
01 April 2007 05:40


Hello,

With no opposition, I've started the localization process and mentioned it in the relevant mailing list discussion.

Yours cordially,
Jesse Martin (Pathoschild)

Navigation