Re: Re: Heirloom Troff for NetBSD (was: Removing ARCNET stuffs)

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

>> [TeX] has, however, nothing like troff's nroff.
> This could be rather easily made with a dvi driver (i.e. taking the
> dvi formatted output and extracting the characters and including
> rendering sequences depending on the type of fonts).

But you'd have to re-typeset everything using only monospaced fonts,
and you'd have the issue that line justification and such are done with
glue that can take on fractional-character widths.  I've been pondering
trying to do character-cell typesetting with TeX and haven't come up
with anything entirely satisfactory.  It might work to set the page
width to 80sp and use a font with character metrics specifying 1sp
widths - it would be an interesting stress test, if nothing else. :-)
Another possibility would be to set up ragged-right with no stretch or
shrink in the interword spacing - I'd have to think about it and would
probably need to swap the details back into my brain; it's been a long
time since I used TeX.  And no matter what you do, there are hrules and
vrules, and you have to depend on the page not doing anything like
hskip half a character width or drop into math mode.

> This can be derived from dvitype(1).

dvitype is a relatively poor starting point; it works well enough for
purposes that can tolerate suboptimal output (eg, a line that comes out
overly long because it happens to be heavy on narrow characters) but it
is no replacement for something that, like nroff, is aware of
character-cell restrictions.

> What was appalling with TeX is the way it had been packaged.  [...]

Agreed...but irrelevant to this discussion, and also true of a great
many other things than just TeX.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B