Topic on Project:Village Pump

Jump to navigation Jump to search

Main page Chinese translation problems

Tranve (talkcontribs)

Hi! Recently I have improved some translations of the main page, and found several problems doing the job:

  1. It seems that there's a hardcoded whitespace after each sentence in the top section, but these whitespaces should not be seen in Chinese.
  2. MediaWiki/zh-hant and MediaWiki/zh-hans don't work with LanguageConverter markups. It formerly use Template:Conversion-zh to function but making the translation quite a mess, and it also doesn't work with this (it use automatic conversion of the zh-hans version), so I removed it.

I'm asking here if anyone can provide a wise solution of these.

Shirayuki (talkcontribs)

Other languages require whitespace after each sentence, so please compromise on whitespace.

Tacsipacsi (talkcontribs)

No, there’s no need to compromise. Just like how you won’t force German translators to write common names in lower case, French to omit the space before colons or Japanese to write dates in MDY format, you shouldn’t force Chinese to have spaces where there shouldn’t be. There are multiple options to solve this:

  • Conditionally insert the space, i.e. {{int:word-separator/{{pagelang}}}} ({{int}} doesn’t seem to work for this use case).
  • Let translators translate whole paragraphs. This is how the Translate extension is designed to work, and it gives translators more freedom (for example, something that’s expressed in two sentences in English, may fit better in one sentence in another language).
Tranve (talkcontribs)

Thanks! I think the second way may be a better solution.

Ciencia Al Poder (talkcontribs)

Translating whole paragraphs wastes translator effort when only a sentence of a large paragraph is edited, and it invalidates a large region of text.

Tacsipacsi (talkcontribs)

If only one sentence of a large paragraph is edited, then only that needs to be retranslated. Yes, the whole paragraph is marked as outdated and gets the pink background, but the translation interface shows a diff, so the translator can see which sentence they need to retranslate. They can ignore the rest (or check if the translation of the rest is correct, but that’s completely optional).

Ciencia Al Poder (talkcontribs)

Well, that's true, but translation uses translation units instead of translating the whole page for the same reason.

Tranve (talkcontribs)

Thank everyone here for your replies!

If using whole paragraphs as a translation unit is really problematic, I think the first solution is also acceptable.

Lakejason0 (talkcontribs)

DO NOT use the LanguageConverter marker (-{}-) on main page please; they would break when directly accessing the Main page instead of Main page/zh. @Tranve

MGChecker (talkcontribs)

I agree that translating whole paragraphs is the correct solution. Otherwise, high-quality translations are generally impossible anyway.

Translating whole paragraphs is not the same as translating the whole page.

Ciencia Al Poder (talkcontribs)

How high-quality translations are impacted with the split of sentences? In all the translations I've made to Spanish, I have found no case where 2 or more sentences in a paragraph needed to be restructured when translated.

Tacsipacsi (talkcontribs)

I run into it every now and then. For example, when the (in English) three-sentence first paragraph of Extension:Cognate/hu was split up into three units, I had no other choice than blanking one of them in Hungarian, as translating it as more than two sentences would result in text that simply doesn’t sound good. In English, short and basic sentences are usual. In Hungarian, it’s very common to use compound sentences. (Maybe it’s because Hungarian sentences tend to have less words than their English equivalents—the subject can often be left out, instead of prepositions and auxiliary verbs, we just put all information in affixes, compound words never contain spaces—except for some cases involving words with foreign origins—, and so on. For example, “I have already logged in to my mail account.” is translated as “Már bejelentkeztem a postafiókomba.”, five words less.)

Ciencia Al Poder (talkcontribs)

Fair enough.

For large paragraphs and bullet points, I think splitting translation units is the correct approach. For example, adding a new sentence to an existing paragraph or a new bullet point to a list, when adding it to its own unit, makes only the new addition untranslated. On the other hand, merging it into the main unit, causes the entire paragraph or list to become untranslated. This is specially harmful when those annoying bots just blindly null-edit those outdated translations, making them current but without including the new content, and nobody will notice even reviewing changes.

Tranve (talkcontribs)

Thank you all for your comments!

However, the problem I actually want to solve is the whitespace after each sentence. If treating the whole paragraph as a translation unit isn't a good approach, is it possible to replace the white space with something like {{int:word-separator/{{pagelang}}}}, just as what Tacsipacsi mentioned above?

UPD: It seems that this problem also exists on lots of pages besides Main Page.

Ciencia Al Poder (talkcontribs)

Manually adding a template or something to conditionally insert white space would make wiki code much harder to read and edit (more than the current status).

I wonder if the software could provide some help here, for example, removing white space automatically when present between adjacent translation units, in languages where it's needed. There may be probably some corner cases where this is unwanted, though.

Reply to "Main page Chinese translation problems"