🚀 go-pugleaf

In article <id.IIWO.Y05@ferranti.com> peter@ferranti.com (Peter da Silva) writes:
>In article <3395@ecicrl.ocunix.on.ca> clewis@ferret.ocunix.on.ca (Chris Lewis) writes:
>> That much?  Eeep!

>I don't see a 60% savings that much of a big deal. Personally, I don't bother
>even turning on optimization in psroff. I'm more interested in groff for the
>font support... and you can't do much about that except convert "\f" requests
>into some magic using your bell-system-logo overstrike + strategic "'fp"
>requests.

FYI: For psroff with the perl manual:

PL10 Optimization off:       1531033
PL10 Optimization on:         573074
PL11 "normal" Optimization:   554333 (shortened some strings)
PL11 "hyper" Optimization:    303104 (line optimization and some wierd bits)

The last is actually *less* than groff with kerning on.  And would probably
be another 40K smaller with a bit more effort.

Which isn't bad since the man page itself is 190K (big man page!).

[Not that these numbers are particularly important.  There are a number of
factors involved.  Who cares anyways?  As long as the output looks good it
doesn't matter how big it is (unless you're in the habit of posting the
output of your troff - Hi Michael ;-).  At least psroff doesn't introduce
errors of its own - several of the ditroff backends are lazy and can
produce some seriously awful output depending on what the input looked
like.  Tpscript and xtroff are cases in point.]

>Actually, if you could make the 4-position font limit vanish I'd blow off
>groff completely. I don't *WANT* to install G++.

I wish I could.  I haven't come up with a way yet.  Hell, if I can get
CAT troff to do pic, there must be a way ;-)

>> Postscript's space width is 1/3rd
>> narrower (12/432 @ 6 point versus 9/432 @ 6 point))

>Doesn't that depend on the font?

Yes.  Rephrase: CAT troff hardcodes space width as 12/432's at 6 point
in spite of the fact that there's a width table entry for space.  Thus,
CAT troff ignores the width table entry for space.  PostScript Times-Roman
actually has 9/432's spacing at 6 point.  Other fonts have different
values.  Eg: BookmanLight is 12/432 at 6 point.

Fortunately, the new optimizer doesn't care whether CAT troff reads it or not.
-- 
Chris Lewis; clewis@ferret.ocunix.on.ca; Phone: Canada 613 832-0541
Psroff 3.0 info: psroff-request@ferret.ocunix.on.ca
Ferret list: ferret-request@ferret.ocunix.on.ca