Sections
I suppose this it is a result of experience. The bigger the portions are to translate, the less likely it is that they get translated or alternatively the longer it takes to translate them. Take this as a good example for a horrifying message in terms of lenght. I have translated about 1850 of the 1923 messages for EOL within two or tree weeks. This was several weeks ago. So far I did not feel like sparing the hour needed to do a decent translation for it. Probably EOL will have to wait another month. If they had split up the message, the translation would already be at their end. However, I am speculating here, but I am sure that reality is not so far off. Cheers
This works if messages are actually autonomous, not if they're highly connected (currently, even sentences are split if they're on multiple lines, as in the example). Usually translators, at least on Meta, will translate a page completely, or they'll translate half each (or similar portions), and the next translator proofreads, and so on; this way, they have to make dozens of edits and to remember very carefully how they've translated the previous strings or they'll make all sorts of errors. The first time I saw (here) a page to be translated this way, in many pieces, I gave up.
I understand that this is the way how the extension detects that a particular portion of the page needs to be (re)translated, the progress/status of the translation and so on, right? But perhaps this could be disabled on a page by page basis (it's not always needed). Ideally, you would have the possibility to edit the page as a whole (or as if you were editing it as a whole) while retaining those features.
Why would you want to use page translation feature at all if you are not going to split the page up?
You underestimate it, it's very useful in any case :-p ; why do you think that's the essential feature?? I really can't understand.
For instance, it manages the list of existing translations for a page automatically, creates a list of translation requests, it marks them as outdated when the source is changed, AFAICS (will) allow review/proofreading with appropriate tagging, can allow automatic selection of translation to be shown to the user, hopefully can be developed to fetch automatically translations from where they're performed to where they're needed (protected namespaces on Meta or even other wikis like m:wmf:?)... the current translation system on Meta is a nightmare, you know? Also, perhaps it's possible to develop some interface to edit the translation as whole as you can do with the source.
And just to be clear, I'm not saying the splitting should be disabled entirely, but that there should be alternative ways to deal with split pages/paragraphs/sentences and that the splitting should be controlled on a per page basis (think also of the old translations which have to be moved to the new system)... at least, that's my confused feeling.
(I'm being serious here, not intended as a joke)
You appear to have gotten so stuck in the horrific translation process that Wikimedia's meta uses, that you're not seeing the extremely high barrier to entry and the lack of overview anymore. Take a step back and compare... And most importantly: use it.
P.s. This does not count as using it :) (2 edits)
What do you mean with "lack of overview"?
I don't think I'm stuck with Meta's system; I do know it, but I've never been very involved in it, and I've used the Translate extension way more and before. I may use the same argument: I think that the page translation may be too stuck with some concepts from the system messages translation features it derives from, and that's why I'm trying to provide some feedback that may make you see it from a different perspective and help us all to understand and explain better.
I do want to use it as broadly as possible, that's why I'm trying to understand precisely how it works and what's the best way to use it...