I jsut stumbled on this analysis of citeproc. Alas, it requires I subscribe to Passport to leave a comment, so I thought I’d instead post a quick update on progress related to citeproc and CSL, since public documentation is rather dated.
My focus has always been on CSL, and the XSLT 2.0 implementation as a solid proof-of-concept. To my mind the promise of CSL is really langauge and document format independent citation styles.
I have been struggling my way to the 1.0 finish line with the schema, trying to finish with some tricky features, and to wherever possible simplify and rationalize the logic so it is easy for styles authors and developers to work with.
When they release the extension sometime in the next few months, expect it to support CSL out of the box for citation style configuration.
Alongside that, Johan Kool jumped in and decided to work out a Python version. With the three of us working on design details of CSLâ€”and each using different languagesâ€”we’ve mananged to make a lot of progress in resolving some of the more difficult problems. We are targetting a pre-1.0 test release sometime in the next couple of weeks, and then a final 1.0 freeze early September.
At that point, hopefully things get more interesting, and I can sit back and watch how others make use of CSL.
- the “info” metadata element uses the same content modeling as that in Atom
- the whole thing is designed on a consistent inheritance model that makes it simple to do the common stuff, and possible to do more complicated customization
- in part as a result, the data field and template models are simpler; constructing GUI editors ought to be easier
- I finally figured out how to support complex internationalization options without making things more complicated for those who don’t need them
- at the request of Matthias Steffens from the RefBase project, we figured out a fairly elegant localization approach