On Fri, 29 Sep 2006, Hans Hagen wrote:
gnwiii@gmail.com wrote:
I don't agree that lack of a current manual is a big problem for potential users of ConteXt. In my experience, the biggest problem is with the 3rd party tools (perl, ghostscript, ruby) on Win32 and legacy commercial unix (where ruby is not provided and the system perl will be a very old version).
actually, one could run context using "pdftex --fmt ...... " but it's less convenient;
technically i could generate most of the tuo file directly (although index sorting always has to rely on an external prog)
some two pass data (cross ref etc) can be loaded before an aux file is written, but toc info cannot, and tex itself cannot rename a temp file afterwards; because context can create tocs at any level any time, this means that there will always be a need for an separate read and write file (actuallym this is also true for more two pass data since some data structures may be defined anywhere in teh document which also leads to async loading)
in luatex, i will probably write the index sorter in lua as well as support a different two pass info model as well which means that in principle one could provide an embedded kind of texexec funtionality; of course multiple runs still have to be managed by some external script then but this can be simple since tex itself can signal the need for that
Does this mean that in future (distant future) it will be possible to simple pass cont-en.fmt file along with the source file and write %&cont-en in the tex file, and someone just having plain tex can run tex filename (possibly multiple times) to get the output? If so, it may make things like submissions to some journals and online archives like arxiv very easy. Aditya