Talk:Translation strategy

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search


Good doc. If I may, I'm not sure how CX can be an example of "one place we can keep up to date" (I understand it's a cross-wiki service, hence something different IMHO). --Elitre (WMF) (talk) 11:53, 7 April 2016 (UTC)[]

I don't think that's what Trizek meant when he added that sentence. I've removed it for now. Trizek, could you explain what you meant? /Johan (WMF) (talk) 22:39, 9 April 2016 (UTC)[]
My point was to think about having cross-wiki things. If the tool is available on each wiki, that may help us to reach people. Trizek (WMF) (talk) 07:48, 11 April 2016 (UTC)[]
User:Johan (WMF), some more thoughts:
  • I believe we need some insight about the motivation of current and wannabe translators (I'm looking for existing resources BTW). While looking for them outside our movement may be necessary, it may not grant good results. We know that paid professionals struggle a lot with our "jargon" and that what they deliver is not optimal (and we are talking about people who translated contents related to community consultations: I'd guess technical translations would be even more challenging). A lot of context is often necessary to figure out what something is even supposed to mean. --Elitre (WMF) (talk) 16:20, 8 April 2016 (UTC)[]
True, but this is equally true of translating technical products in all contexts where there's an existing jargon in the target language. I'm not saying this is the way to go, but I think that highlighting it is one way to go. The first step to help newcomers would be to make sure they have somewhere to turn when it comes to terminology – both technical and Wikimedia-specific jargon. And, speaking as a volunteer translator, our translations are often of varying quality as it is, which is worth remembering.
I'm planning to talk to some of translators this coming week. /Johan (WMF) (talk) 22:39, 9 April 2016 (UTC)[]
  • When we're involving Wikimedian translators, why do they answer our calls for support? Do they just have enough time, do they think they're doing a service to their community, do they just enjoy translating, other? (AFAIR, our biggest projects do not even have -deliberately, I'd guess- local documentation explaining and encouraging how non-content translation can be done.) What I found out with the VE translathons is that acknowledgement and small prizes go a long way (see the second last slide here). We get much development work done through paid programs like GSOC/Outreachy, and that's something to keep in mind, although it may look not related here. --Elitre (WMF) (talk) 16:20, 8 April 2016 (UTC)[]
Interesting. What kind of prizes? I've read some research – I can't find the article now, but I'll do a new attempt to find it again later – that suggests that small rewards might actually be less motivating, because you focus on the small reward instead of the intrinsic value of what you were doing (the thing that motivated you from the beginning), at least if you consider what you're doing creative to some extent. I might be overreacting to one (1) study I once read and your empirical experience should probably be considered more important, but I feel this resonates a lot with how I behave as a volunteer translator and editor. /Johan (WMF) (talk) 22:39, 9 April 2016 (UTC)[]
Wikimedia merchandise of their choice, given a certain budget. --Elitre (WMF) (talk) 14:16, 20 April 2016 (UTC)[]
We have a budget for that. We can nominate our best translators, based on involvement or quality (despite that last point is difficult to measure). Trizek (WMF) (talk) 16:05, 21 April 2016 (UTC)[]
(That meant they were given a limit. --Elitre (WMF) (talk) 08:28, 22 April 2016 (UTC))[]
  • deployed the Thanks feature only recently AFAICT, and I'm not sure that's enough to make someone feel that what they do is immensely useful and appreciated. So if I loved translating, in 2016, what would make the Wikimedia environment the most appealing destination to work on? ;) The fact that the movement's mission is the best in the world, alone, won't be enough. Why shouldn't I join communities of "subbers" instead, who'll place my name prominently in the credits (random example for TED videos, as well)? When someone is translating the interface for us, there isn't even a clear correlation between the place where you work and where the content ends up.
    About a year ago: "2015-07-01: Thanks and Echo extensions were installed on this wiki." I think that is not too bad. I cannot have too many extensions such as LiquidThreads that cause so many problems, as I am doing the most maintaining there. Fortunately Echo and Thanks have worked well, though Thanks could be even more integrated: it is not visible in the translation interface. --Nikerabbit (talk) 08:50, 10 May 2016 (UTC)[]
    I just got a question from a translator who inquired about the practice of adding one's name to translation, to get some credit for it. Of course, this is not how attribution is usually done on our projects when it comes to adding/editing content. OTOH newsletters like the VE one do not really have an immediate history link that one can click to find out who worked on that issue for a specific language (you can get there: it's just not as immediate). If we can't provide direct attribution, I wonder what would change if we provided at least a more direct/better worded link to the history page for each language. Dunno how hard it would be to do that with the LUA module we use... but I'd be definitely interested in A/B testing this. --Elitre (WMF) (talk) 16:47, 7 July 2016 (UTC)[]
  • In an ideal world, we'd be translating the interface "in context", just right-clicking on the string to translate. --Elitre (WMF) (talk) 16:20, 8 April 2016 (UTC)[]
I think it makes sense for this to mainly – but not exclusively – focus on what we can do with existing tools, because the development of new ones are mainly out of our hands, so to speak, but we should definitely also help out by listing things. /Johan (WMF) (talk) 22:39, 9 April 2016 (UTC)[]
Many people think that in-place translations would help. I disagree, it is opposite of trivial to implement and can never cover everything. I think it is not a good use of our limited development resources. --Nikerabbit (talk) 08:50, 10 May 2016 (UTC)[]
Elitre (WMF) originally wrote the post above as one. Since I realized it was much easier to follow the discussion if I answered inline when previewing this, I took the liberty to do so, and copied her signature to the individual paragraphs as well. I hope that's OK. Johan (WMF) (talk) 22:39, 9 April 2016 (UTC)[]
Elitre has good ideas above. --Nikerabbit (talk) 08:50, 10 May 2016 (UTC)[]

Some thoughts[edit]

Tracked in Phabricator:
task T129088

A searchable glossary would be absolutely amazing, both for technical terms and wiki jargon. There have been many times where I'm stuck looking for a good translation of a single word, and end up either just not bothering to translate, or I end up making up a word and putting the English word in parentheses. A glossary would help immensely for this. While I'm at it: I'd like it to be easier to reuse translations. There are far to many instances where you need to translate say an introduction paragraph or something along those lines, and somebody has probably translated it before, but because it's hard to know that, you end up translating it anyway. This wastes effort which could be put to better use. We sort of have this with the suggestions to the right of the text translation box, but I find that there are no suggestions when the text is more than say five or so words long, despite the fact there have been made identical translations before.--Telaneo (User talk page) 20:52, 2 May 2016 (UTC)[]

Agreed with all you write. Even if there is no suggestion - which I understand is mainly for performance reasons when long texts are to be translated - there should be a button asking for a more lenient search, no matter how long it lasts. Human translation time will always be a lot longer and more expensive.
I am working on glossaries on the basis of already existing translations of the translate extension. Expect first results in a few months or so. ::--Purodha Blissenbach (talk) 12:49, 9 May 2016 (UTC)[]
I'm also working on defining such a glossary, more on the usability side. Trizek (WMF) (talk) 13:21, 12 May 2016 (UTC)[]
I think a glossary will be very helpful. Even if it is just a regular page with a table. --Lsanabria (talk) 15:48, 30 May 2016 (UTC)[]

What is the point of translation administrator status ?[edit]

I have been thinking to create a wiki page as glossary for some weeks. So I am happy to see others need the same.

Also: I don't understand why everybody can edit the pages but only "translation administrators" can facilitate the translations. Sometimes I want to translate a page but the translation markups have not been created. --Pols12 (talk) 20:58, 2 May 2016 (UTC)[]

Pols12, I can't answer as to why the system has been built the way it has (I'm sure Niklas could explain, though), but as it it looks now, one reason that immediately comes to mind is that translation administrators can also remove pages from translation, which means they're removed from the translation system and all translations become ordinary pages. This can't be undone, but has to be manually fixed – remarking the page for translation, manually readding the translations in the right places. It would be very effective vandalism, because it takes so much longer time to fix than to destroy. /Johan (WMF) (talk) 00:02, 5 May 2016 (UTC)[]
It doesn't fix the structural problem, of course, but you could always apply for translation admin rights. /Johan (WMF) (talk) 00:04, 5 May 2016 (UTC)[]
Thank you Johan (WMF) for your answer. I could request those admin rights, but an admin has community pressure. A member who has not request any right is still entitled to make a few mistakes... --Pols12 (talk) 23:20, 5 May 2016 (UTC)[]
What you describe doesn't exist. The expectations for usage of translation administration functionality is the same whatever the group that assigns the permission. Nemo 17:28, 9 May 2016 (UTC)[]
Actually it should be pretty easy to revert even the removal of pages from translation. Deletion on the other hand... --Nikerabbit (talk) 08:52, 10 May 2016 (UTC)[]

New websites[edit]

  • Is there any reason to believe new websites help anything? --Nemo 17:25, 9 May 2016 (UTC)[]
    I don't think a website will be of any help whatsoever on its own. If we put it up, link to it from Twitter a couple of times and leave it there, it will do absolutely nothing. I see it as a tool to use in communication with potential translators – a step to introduce the concept in the conversation.
    So you answered how a website can not help. My question stays unanswered. Nemo 12:07, 10 May 2016 (UTC)[]
    "[A] tool to use in communication with potential translators" – my experience, trying to do outreach my volunteer capacity, is that I have much better success if I can link to simpler explanations off-wiki first, as a stepping stone to getting people involved, perhaps because the wiki layout is cluttered with the tools we have not yet convinced people they should use (to the degree that many will use Wikipedia on a daily basis for years and yet be unaware of the existence of the edit button). /Johan (WMF) (talk) 13:57, 12 May 2016 (UTC)[]
  • Do you think it's possible to make sure that all the people with relevant knowledge have access to such new website? Like, all translation admins, all translators etc. Nemo 17:25, 9 May 2016 (UTC)[]
    With access you mean e.g. editing? /Johan (WMF) (talk) 18:05, 9 May 2016 (UTC)[]
    For instance. Nemo 12:07, 10 May 2016 (UTC)[]
  • Not sure if creating a totaly new site will help. Maybe we could try to reactivate Meta:Babylon (and maybe change the name?). Besides adding a comma and removing an outdated link, it has not been updated at all this year. I think removing the translations will help to keep it more up to date. I could make a few edits but I am not a translation administrator and since the source language of most (all?) documents here is English, I don't think having it translatable really helps anyone. --Lsanabria (talk) 15:41, 30 May 2016 (UTC)[]
Could be this isn't the best idea, given that people generally seem to disagree with me. I'll rethink it and see if I can come up with better arguments to convince you, or if I should let myself be convinced. :) /Johan (WMF) (talk) 15:48, 31 May 2016 (UTC)[]
Translating is work, and motivating people who are not already or genually inclined is hard, most specifically in fields where the results are not very visible or enduring.
Talking arbitrary people understanding two languages into translating may be generally less clever. There are so many questionable or even wrong translations that are hardly revised… --Purodha Blissenbach (talk) 18:43, 31 May 2016 (UTC)[]

As most people didn't find the arguments for this convincing, this has been removed for the time being, until better arguments can be made or a clearer need has been shown. /Johan (WMF) (talk) 11:39, 17 June 2016 (UTC) (previously unsigned, signature added by Johan (WMF) 00:06, 18 June 2016 (UTC))[]

Thanks! Nemo 21:41, 17 June 2016 (UTC)[]

Who, when, what?[edit]

Who is this strategy for and who is going to execute it with what resources? Just wondering why I have not seen pings to Language team and other people working on translation tooling to participate and comment. --Nikerabbit (talk) 08:55, 10 May 2016 (UTC)[]

I'm primarily focused on the things that can be done without engineering resources, but also collecting information the team can use, if they think it's relevant. /Johan (WMF) (talk) 11:29, 10 May 2016 (UTC)[]

Problem statement[edit]

The Wikimedia movement is an international movement supporting communities in more than 270 languages. No it is not. The movement is the people using hundreds of languages. The actual number is irrelevant. In my opinion we should aim to support most languages (excluding pig latin etc.) so that they have equal footing from tooling side. We should foster all these languages to grow communities who can for example do translations, but also understand that those communities will always be of different size and activity.

The statement of the problems is not clear. Here are some thoughts:

unclear source material
Clear source material is easier to translate – and this is something we haven't put much dedicated effort outside interface messages.
not enough is translatable
I definitely agree with this one. E.g. the Wikimedia Shop.
translation is not easy enough
I definitely agree with this one. But there is no mention of e.g. putting resources to visual page translation, fixing the translation memory, better integrating machine translation (which I would consider as ways to work on this) in this document.
proliferation of translation tools and places
I think we have very good tools tailored for their use cases. Some convergence can and will happen, but I don't see this as the main issue.
translation coverage varies a lot by language
I think this is just a fact we need to accept. There is only so much we can do to encourage people to translate.

--Nikerabbit (talk) 09:08, 10 May 2016 (UTC)[]

The way I've envisioned this process is that we start working on the things which wouldn't require engineering resources. /Johan (WMF) (talk) 11:34, 10 May 2016 (UTC)[]