I doubt there is a typo in Translations:Extension:CommonsMetadata/10/en; Isn't commons-inspecific a typo of commons-specific? Where can I find meaning to that terminology? The sentence goes: *Ideally such a system would be as commons-inspecific as possible, with the commons and non-commons part separated.
Is it typo? commons-inspecific
The design goal was to avoid being Commons-specific (ie. have parsing logic that works on any wiki, not just Commons).
In practice that's not really the case, the extension mostly looks for Commons-specific markup, but that markup is kept simple and is documented so any other wiki can easily add it. See Commons:Machine-readable_data.
Hello there, would it possible to extract description data also from another two sources: 1, description of the image in the gallery section of a page, and 2, plain text in the text body of an image (content of the page File:Image.jpg) ?
1. is hard, and we don't want to put much effort into parsing data from wikitext, the plan is to support structured data via Wikibase (Wikidata) instead. As for 2., it would not be very useful without locating out the useful part of the description (there will be license templates, section headers and whatever else intermingled with it) which is again more effort than worth. (If you do need the contents of the file page though, you can get the via the revisions API.)