🚀 go-pugleaf

connolly@convex.com (Dan Connolly) writes:
>I am interested in a text formatting system written in PostScript.
>It would read data from the input stream (similar to the way
>the image operator, or any number of image decompression operators
>read data from the input stream). Then it would arrange the text
>on successive pages and render the text.

This falls into the category people like to call "Theoretically Possible,
but..."  It is possible to format text in PostScript, but it's really not
very well suited to the task.

PostScript is interpreted, and although that's not necessarily a damning
feature, it makes the efficiency problems a lot harder.  I think the two
basic problems here are late binding and latent types.  You can fiddle with
"bind" and //, and interpreters try to avoid re-evaluating binding where
possible, but it's still an issue...and latent types do require a lot of
extra code for run-time type checking.

PostScript string handling is weak and clumsy.  It turns out that for-
matters spend a lot of their overall processing time in line breaking,
hyphenation, and justification.

Also, you probably want the formatter to run on your host machine, so that
it can find hyphenation tables/dictionaries, include files of macros and
common text, etc.  This can be done, but it's not the model you suggested.

>i.e. what I'm looking for is troff or TeX or something, but written
>in PostScript...

At this point, most of us are probably asking "why PostScript???"  That is,
why does the implementation language matter?  It seems like PostScript has
no particular strengths, and several serious weaknesses.
-- 
Dick Dunn    rcd@raven.eklektix.com   -or-   raven!rcd    Boulder, Colorado
   ...Never offend with style when you can offend with substance.