Preferred way to translate a large text that's part of the software
Ah ok. Yes, it would be straightforward to use page translation, with the content of each p tag as translation unit. If we don't add anything else to the translatable page, then the translation pages will only have the expected HTML as content.
I would just go for it and put them up as subpages of Translating:Lib.reviews. We don't have a good way to import the existing translations but we could also just freeze the translation of the unneeded languages until we figure it out.
I've imported the two pages now:
I've also marked the page content with <translate> tags. I'm unsure about the best way to mark up the page and whether these docs are still accurate; other pages here like How to translate appear to use a whole tag for the entire page, so I haven't manually segmented it beyond excluding the <languages/> tag at the beginning. Please let me know if that works. If so, can you mark the pages for translation?
This is acceptable, but I thought you would want to preserve your own HTML (for instance the IDs): it would be ok to paste your HTML as is and add a translate tag inside each p tag, I think.
Excellent point re: the IDs; it's indeed best to keep these stable across translations. Beyond that I think I should be able to transform the MediaWiki output (mostly by stripping any extraneous stuff I don't need automatically) on import. I've added the IDs and excluded the relevant HTML from translation. Does that look reasonable to you?