Plan for the future

[addshore]

Mainly inviting @samwilson for a chat here :)

I had a look at the dpeendenats of https://packagist.org/packages/addwiki/mediawiki-api-base/dependents?order_by=downloads And it looks like:

a) a few users still use it, including you b) People don't tend to use the other packages that I maintain as part of this mess

I would propose

1. I sunset this "addwiki" experiment as a whole
2. We focus on mediawiki-api-base more (instead of the rest of this mess), which would proably lead to being able to keep the library more up to date, and up to speed with the mediawiki apis etc?

Thoughts?

[samwilson]

Yeah that sounds good.

I had a look at the dpeendenats of https://packagist.org/packages/addwiki/mediawiki-api-base/dependents?order_by=downloads

Dependent packages is one measure, but I suspect most usage of this is in projects directly so those numbers aren't the full picture, see also the stats (where it's obvious that V3 hasn't really taken off). Wikimate is probably the next most popular similar thing (judging by this list) with 13k downloads (vs 140k for mediawiki-api-base).

I think retiring this and focussing on mediawiki-api-base and some of mediawiki-api would be good. Perhaps a rename as well, to make it super clear? Something generic like MediaWikiClient or MwPhp.

It'd be somewhat similar to the JS Mwn maybe? Aim to provide simple ways to do common things (e.g. request continuation, category traversal, stuff that is more than just a wrapper to the API calls). Which it already is, of course, I don't mean that there'd be anything particularly new.