CSL-M extensions

[fbennett]

I've updated the descriptions of CSL-M schema extensions in the processor docs. I think everything is now covered. If you hit anything in the tests that is missing or inadequately described, just give a shout.

[cormacrelf]

Hey, thanks for doing that! It looks great. A couple of questions from skimming just now:

• What is the purpose of cs:law-module? Are legal types disabled until listed in one of these, or something?
• How do you use cs:alternative and cs:alternative-text?
• Could you give an example where disambiguate="check-ambiguity-and-backreference" gives different behaviour to disambiguate="true"? I admit I haven't searched the test suite, but I feel like there's some valuable insight into citeproc-js' disambiguation scheme lurking here.

[fbennett]

• cs:law-module is an element under cs:info that declares the modular types covered by the module. When the processor walks up an item's jurisdiction hierarchy in search of a set of macros to apply, a module that declares specific types will match only for those types (otherwise it tries the next, until it hits the top level, falling back to us if there is still no match.)
• The cs:alternative and cs:alternative-text elements were added fairly recently, to address the needs of a Japanese style that requires translations and republications to be cited together with their originals. The style guide is in Japanese, but if you search the in-text refs section for = and look through the bibliography examples you can see the challenges. For a worked example, you could check out the Japan Sociological Society style itself, which uses cs:alternative and clears all of the examples in the guide. There is also a user tutorial by Jonathan Lewis. (All in all, it's not a pretty solution, but it's also a very hard problem for CSL in its current state.)
• The check-ambiguity-and-backreference option on the disambiguate condition may not be relevant anymore, I need to check. I remember the issue---a user complained that a note style with note-targeted backreferencing was adding titles unnecessarily---and that including the note number when casting the subsequent form of the site (for "ambig" grouping purposes) solved the problem. I think, though, that this is now the default behavior when the by-cite disambiguation method is used. I'll ... need to check sometime.