Viewing a history listing
Hello, Houston. We had a problem. I translated ~500 core messages. Yesterday I've got message "Language code sjn is not valid.". What happened?
The language code: Sjn is actually not supported by translatewiki. We recently fixed a long standing bug where translators were allowed to translate to unsupported & invalid language codes. While sjn is valid, it is not supported by translatewiki.
I'm informing one of our staff who will review if support for this can be added to translatewiki.
Hey, I hope you haven't forgotten about my problem. I would like to fill out the application required to add a new language section on Wikipedia in Sindarin (sjn). I have previously translated 500 core posts and hope to continue translating in the future.
The line for which constructed languages in scope for translatewiki.net is a bit fuzzy. We have Esperanto and Volapük but not Klingon for example. Is Sindarin being used as a general purpose language?
Well it is valid in ISO 639, in BCP 47, in CLDR... And there are also languages supported in this wiki not only for translating the MediaWiki UI. As well there are MediaWiki-based wikis in Wikimedia that also exist (at least in Incubator, sometimes as subprojects of another project using another main language for the UI, e.g. dedicated sections/categories in international Wikisource and Wikimedia Commons, or translations of labels and terms in Wikidata, or translated terms and wordlists in Wiktionary), still without having a supported MediaWiki UI working in at least one wiki.
But generally they exist because they have an active and stable enough sponsoring community that develop and maintain them.
For languages in development, the first need is to create this community in some existing stable space, create the communication areas and link them, collect sources and references. Then comes the time to start developping an UI for testing a beta version of a wiki or any other software translated here (or in other translation tools hosted elsewhere with some open licencing terms). That small community will make their choices, will decide how to unite their efforts. At some points, there will be enough cross-reference that will allow to start porting the Mediawiki UI and support it in a stable way, with a reasonnable way to cooperate and decide what to do and how to review that work: it will be time for this test wiki to see if it can support the existing efforts (there may be various technical issues to discusss and solve, but here this is not the case for thelangauge code). When this beta wiki or site is working quite well, Wikimedia will support these efforts.
But Translatewiki.net does not depend only on Wikimedia decisions (via its Language comity supporting the Incubator) and as well developers of MediaWiki are not following only the goals of Wikimedia (because MediaWiki may also be used in other independant wikis, including proprietary ones: for example if a government or institution decides to support a language, and wants to develop a wiki for it, using MediaWiki, it will also start working on such adapation and will coordinate with MediaWiki developers to make the needed changes, and if this is released, rapidly Wikimedia will also accept to start testing some beta wikis, and later decide to open it with their own independant projects).
So the most important remains: developing the community. When the community is small, it needs to have its efforts best supported by hosting their project inside another existing project where they can easily find contacts and volunteers (not always speeking other languages with the same level of expertise). The wiki is then not the only goal, the first goal is social (once you have resolved the most critical blocking factor, notably the encoding support for its script: no problem here for Sindarin, as long as it starts with the Latin script because other artistic scripts are still not supported in Unicode)
If there's still no script support, the best that can be done is to first collect a material collection of text sources (e.g. facsimiles of books, and papers written by linguists, protographs of artistic objects...), and then provide that to Unicode/ISO/IEC to create or improve an encoding proposal with enough references (after solving copyright and licencing problems that many artistic creations will have to face for about one century after their initial publication, unless the authors decide to open these sources to the public with open licences). After this encoding step, you need to convince designers and foundries to develop an open font for it, and then make it available from a reliably maintained source. At this step the script will be allowed to convert many documents and build a reference linguistic library that a community can agree on to build their project.
It is not easy, it takes long (many years) to convince and recruit new volunteers and sponsors proposing to help (financially or by offering services). And you have also to accept that this help will also come from people or organizations that don't necessarily have an expertise on your language and may initially make some errors (that you intend to fix later once you've successfully build your community joining its efforts). Finding a local governement or linguistic departement in some educational institution (including public libraries) is generally decisive to accelerate that project, because commercial organizations (including "social networks") often don't find any interest to support this effort.
Hello, User:Verdy p and thank you for your feedback. In fact, all the problems you listed are not, and displaying the wiki interface as a tengvar is not relevant. The community exists, it is huge and helps to translate any text into Sindarin without difficulty. In addition, today there are neural networks that can significantly reduce the time it takes to translate text. Previously, I started translating the interface for the wikipedia project in Sindarin, and translated over 500 lines in August last year, until I encountered a problem whose name is "Language code sjn is not valid." I don't think it's such a complicated problem that it can take six months to solve.
Using "neural networks" or other automated translation tools is exactly what must not be done here. Translations here are meant to be made, reviewed and corrected by humans which are also alone to be able to do that consistantly according to a stable terminology and accurate context of use. It is human translation done here that can be used by automated tools to propose some hints, not the reverse.
And my reply was relevant: even if Sindarin has several scripts (and most Sindarin-addicts people use in fact the Latin script), the original tengvar script (or its numerous derivatives) are not encoded in Unicode. This means that Sindarin for now can only be written by non-speakers, using a foreign Latin transliteration, and not a real translation. And scripts was only the small part of my reply (for compleness).
All the reast remains valid and notably this summarizing sentence: So the most important remains: developing the community. Because this is what is really needed to develop a common terminology and stabilize it with an accepted orthography. As long as it does not exist, it is illusory to start here, and all that can be done is individual developments made on independant competing publications, that will need to find their audience, and may be some comparitive tests and many evolutions to reach some minimum concensus. And this minimum should not jsut be done only for MediaWiki on the web, there are necessarily more contents on other medias and people that are not technicians at all. It is important then to adapt to what people use today to reach them, i.e. firet try to adapt to them in order to get contacts. Then only you can propose to adapt their work to a new media or technology and show them the result and discuss with them: if you do things correctly, may be they will start to adopt your solution and finally they will want to join their efforts to create a community wiki (instead of just their legacy tools).
Last edit: 23:46, 19 February 2022
Thank you for your response. I hope you understand that neural network assistance today is not yet able to completely replace manual labor. No matter how well trained a neural network is, it will always remain only an auxiliary tool. Furthermore, I admire your policy of focusing on the quality of translation. Yet I do not share your skepticism about the technology, because the high accuracy of the automated translation produced by neural networks can significantly speed up translation of large amounts of text information, and correct the resulting translation as necessary, during manual revision, which of course no automated tools can replace. But the main thing that makes automatic translation possible is the work of many, many people in the Sindarinist community from around the world. Without them, without their strong input and numerous consultations with linguists, I would not have been able to assemble the entire corpus of literature I used to train the neural networks I use, for which I am immensely grateful to all these people, and do not in any way belittle their personal merits. In addition, corrections to the translated texts and manual translation of texts on new topics would also not have been possible without the community's involvement. It is thanks to all of these people that we are now at this point of contact.
Nevertheless, I do not find your comment about Sindarin's own writings paramount, since Sindarin and most of the authors working with it are also accessible without the direct use of Tengwar or Sarati. And, at the same time, I am puzzled by your expression "true translation. What meaning do you personally put into it? A real translation in Sindarin can be done using the Latin alphabet, and written in Latin alphabet, Sindarin does not lose any of its qualities. This does not make it any less informative, and it makes it even more accessible to casual observers.
The part you find most important is also beyond priorities, because the community you are talking about has been around for a very long time, although, specifically within MediaWiki, it may be sparsely represented. The required terminology in Sindarin exists and is stabilized with accepted spelling. Also, I don't understand your skepticism about some kind of competing publication and evolution. Sindarin is not a language in which it is possible to express a single concept with hundreds of thousands of words composed in different ways. Although Sindarin allows the construction of similar words, just as "galadhui", "tawarui" and "taurui" are words with different meanings, so any other words that will be used in the text will always differ in meaning and meaning. Again, I appreciate your worries about the quality of the translation, but there is no problem with Sindarin in the meanings you mention.
Further, you mention, as I understand it, some mythical fragmentation of Sindarin speakers that does not exist. As a rule, every single group of Sindarinists has several professional linguists who are also in frequent contact with each other, preventing differences in language rules or some words that cannot or should not exist in Sindarin.
Nevertheless, despite my bewilderment at your remarks, I will still do as you suggest and create a MediaWiki-community, as it might benefit the existing language project in Sindarin. Actually, personally, I have a huge skepticism about the administration of both resources, because in both resources I was allowed to make a successful translation of texts, and then suddenly said "no", at a time when I should have moved on. Now that you can't create new articles in the incubator project, you recommend that I create a MediaWiki community, and looking at that recommendation, I have a vague doubt about whether it's such an interesting form of bullying for people of a certain category who don't speak the same language as the administrators of both resources. Just in case, however, I will keep my ears open and prepare an appeal to the upper management of Wikipedia and a few other important authorities so that they can mediate and witness to the situation and help resolve it in case of possible complications.
Please consider that I'm not a site admin, so I cannot fulfill your request myself.
As long as Special:LanguageStats/sjn or Special:ActiveLanguages/sjn says that the language code [sjn] is not a supported language, I cannot do anything anything else that explaining probable causes that may explain why your request has not been honored by site admins (like Nike and Abijeet Patro that replied above).
And as a consequence the Portal:Sjn correctly displays the orange banner at top (asking you to contact this support page), and so it cannot link to a working translation interface.
Sorry for that, but we've not seen any reply from any one of them since the begining, to explain why your request has stalled and what you can do. But I'm not the right person you'll need to convince here.
I just said that involving more people here to support your request could be more convincing and decisive. But for now you're still the only person asking for it. So it's up to you to contact other peoples you know that will accept to work on this site with you. And then they'll need create an account here (for translating into any other supported language. The same will be true for Klingon [tlh] and other artificial/artistic languages, and then can register themselves in the list of candidate translators in the Portal page and in the babel box of their user page, showing their interest and knowledge of that language. It's not at all a problem of language code, given that [sjn] is already valid in ISO 639-3 and for use in BCP 47 as a primary language subtag and also not a problem for the choice of valid ISO 15924 script code to use by default or as a supplementary BCP 47 subtag (for Latin, or Cirth, or Tengwar).
How is it used as a general purpose language? Are there, for example, any other websites or apps in which the whole user interface is in this language?