On Thu, 12 Sep 2013 at 04:06PM -0700, William Stein wrote: > Did anybody write (or consider writing) the other approach to this, > which is a tex file > that gets processed directly by sage to make another tex file, which > is then compiled. > This is trickier for tools, but recently Harald pointed out to me that > this is exactly what they do with R, and it works really well -- with > something like that you have a lot of power... but it causes havoc for > things like inverse/forward search, which bugs me.
I agree with Volker. I have thought about it, but you have to parse a TeX file -- and my undrstanding is that the only thing around that can really parse arbitrary TeX files is...the tex binary. TeX is a programming language that can to some extent alter its syntax on the fly, which makes parsing challenging/nightmarish. That's part of why I believe that Rob Beezer's idea of writing his books in XML is actually really smart. Because his book is in a perfectly parseable format right from the start, he gets access to lots of great tools, and can still turn his XML into TeX and typeset as normal. (No, I'm not advocating everyone stop writing TeX and convert to some XML dialect!) Dan -- --- Dan Drake ----- www.math.wisc.edu/~ddrake/ -------
signature.asc
Description: Digital signature
