Hi Kieren,
He’s **extremely** interested in the possible hooks from Lilypond to algorithmic music generation, and the interweaving of music with text (e.g., LaTeX). I’m hoping to get him hooked on the ‘Pond, beyond the high-energy sales pitch I gave him in person.
This is great! Some people have already done great work in incorporating LilyPond and LaTeX. On the topic of Abjad, there is coincidentally a 5-day workshop on using it (led by the creators) that starts today at Stanford:
I've recently done a full musical score entirely in TeX without LilyPond or any other language/interface (although, it is more of a non-traditional "graphic" score that doesn't utilize western notation). So, with the right project, that's certainly a possibility by itself. TeX is extremely flexible/extendable due to all of the packages available, including lilyglyphs:
... and musicexamples:
... both written by Urs Liska (who will likely have more add to this conversation). There are a range of examples of the LilyPond engraving capabilities here:
And openlilylib in general:
... is a resource for lots of new LilyPond-specific tools, in addition to the LilyPond Snippet Repository:
On the note of using algorithmic composition directly in LilyPond source code, I can't comment from direct experience. But if the composer is fluent/comfortable in Scheme (or functional programming in general), I believe that LilyPond would be a more than suitable arena for implementing those techniques, or at least experimenting before "giving up*, or even before moving on to more advanced workflows like Abjad which really do require some time (and a lot of confidence in committing that time) to understand.
HTH,
Jeffery