Ok, see my responses bellow. I am interested if there are some users who actually do use offline wikipedia and how would they like this new approach.
On Fri, Jan 23, 2015 at 1:34 PM, Federico Leva (Nemo) <[email protected]> wrote: > You don't need a new application for this, you just need some upgrades to > the ZIM generation pipeline. How about sending patches for > https://sourceforge.net/p/kiwix/other/ci/master/tree/mwoffliner to generate > category-based ZIM files and the like? > Which is far harder than creating a new application that wouldn't use this ZIM format. So I don't think so. > Customisation possibilities used to be better here before > https://www.mediawiki.org/wiki/Offline_content_generator was deployed on > Wikimedia projects, as ZIM files of few hundreds pages were very easy to > make on wiki. Help with the OCG component is also appreciated. > > Finally, icnremental ZIM updates are an area of ongoing work. You can > probably help here as well: https://phabricator.wikimedia.org/T49406#492033 Ok these are all "planned to be implemented" solutions that do not work now, and maybe will never work. I don't really see any option to use either of these. ZIM format is completely incompatible with what I have proposed. It's one time-baked file that is supposed to be read only, not heavily modified. SQLite format is far better for this design and already exist and works. My goal is to create application that would be naturally extremely small and super easy to use by users who have no understanding of computers and who don't want to download 40gb files in any way. Even if there was a wiki interface that would allow people bake their own ZIM files, I would find it very complex. If you wanted to add 1 page to your collection, wiki would have to generate a new file and you would have to download it again. Using API instead to retrieve data would be more simple. _______________________________________________ Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
