From: l...@neteng.engr.sgi.com (Larry McVoy)
Subject: Groff maintainer?
Date: 1996/11/26
Message-ID: <57e4s2$c8g@fido.asd.sgi.com>#1/1
X-Deja-AN: 200806430
sender: bug-groff-requ...@prep.ai.mit.edu
organization: Silicon Graphics Inc., Mountain View, CA
x-gateway: relay4.UU.NET from bug-groff to gnu.groff.bug; Tue, 26 Nov 1996 03:38:25 EST
reply-to: l...@neteng.engr.sgi.com
newsgroups: gnu.groff.bug
Who's doing groff these days? I have a few hacks to contribute.
--
---
Larry McVoy l...@sgi.com http://reality.sgi.com/lm (415) 933-1804
From: j...@jclark.com (James Clark)
Subject: Re: Groff maintainer?
Date: 1996/11/28
Message-ID: <2.2.32.19961128205233.0096c818@pop.jclark.com>#1/1
X-Deja-AN: 201302168
sender: bug-groff-requ...@prep.ai.mit.edu
x-sender: j...@pop.jclark.com
content-type: text/plain; charset="us-ascii"
x-gateway: relay6.UU.NET from bug-groff to gnu.groff.bug; Thu, 28 Nov 1996 18:44:51 EST
mime-version: 1.0
newsgroups: gnu.groff.bug
x-mailer: Windows Eudora Pro Version 2.2 (32)
At 07:05 26/11/96 GMT, Larry McVoy wrote:
>Who's doing groff these days? I have a few hacks to contribute.
I am still officially the groff maintainer. Unfortunately I'm rather
overwhelmed at the moment, and have virtually no time free to devote to
groff. I also temporarily don't have a Unix machine here. However, I have
been filing away bug reports and have fixed the serious bugs. When I have
some time free, I do plan to do a new release. My excuse is that I have
been working on a free next-generation document formatting system (see
http://www.jclark.com/jade/ for more info). If there is anybody who would
like to take over, I would be delighted to hand over maintenance, especially
the configuration and installation side of things.
James Clark
j...@jclark.com
From: st...@nestor.wlu.ca (Steve Izma)
Subject: groff/SoftQuad-troff diffs
Date: 1997/02/15
Message-ID: <199702151927.OAA00317@nestor.wlu.ca>#1/1
X-Deja-AN: 219012077
Sender: bug-groff-requ...@prep.ai.mit.edu
x-gateway: relay7.UU.NET from bug-groff to gnu.groff.bug; Sat, 15 Feb 1997 13:25:11 EST
Newsgroups: gnu.groff.bug
I have been using various forms of troff since about 1982 (I've
been a commercial/professional typographer since 1971). For the last ten
years I've been using mostly SoftQuad troff running under System
V. About a year ago I started using groff fairly intensely under various
distributions of linux both at home and at work. At work (Wilfrid
Laurier University Press, a scholarly publishing house in
Waterloo, Ontario, Canada), I manage the computing systems that
produce our books and journals. We generally have fairly high
typographic standards and have never found a desktop publishing
system that gives us the kind of control we've had with SQPS.
Since SoftQuad is no longer supporting SQPS, we've been trying
out groff but haven't moved to it except for a few special
projects. I write all our tmacs; we don't use ms, mm, etc.
Clearly for us, groff has great potential and many enhancements
over SQPS. However, many of our problems require discussion with others
who either maintain the software or else have similar experiences with
it. The end of this message is just a list of how the differences
between SQPS and groff affect our work. I'd very much appreciate either
comments on the following problems/differences or else pointers to more
info.
Thanks in advance.
Differences between SoftQuad troff and groff that affect us:
1) I've become very dependent on SoftQuad's trace output for
debugging macros; groff's trace output is minimal, as far as I
can tell -- I usually have to insert comments in macro code to
get reports at various places, but with sqtroff I can follow the
execution of various commands and the values of any variables
used.
2) sqtroff consistently allows square brackets for enclosing not
just long names but also the arguments to backslash commands
(e.g., \w \h \D \l, etc.); pairing and nesting of square brackets
is easier to decipher than the quote marks required by groff.
3) To get morticing/kerning in groff appears to me to require
using the tracking request and keeping track of font changes and
other variables, whereas sqtroff has a simple (?) kerning request
in .swd
4) groff has a definite advantage in handling fractional point
sizes for fonts; however, I usually set number registers to be
used for arguments to .ps and I can't figure out how to make the
number register a fractional number (use 1 register for whole
point sizes and concatenate another register for fractions? seems
awkward).
5) I can't find documentation for the hyphenation-exception
dictionary (seems to be based on Tex, but I can't find info for
that either); SQ dictionaries are compiled and less formulaic,
but I suspect that the groff method is better. For some reason,
all-cap acromyms that aren't English words get hyphenated.
6) sqtroff has a greater number of codes for fixed spaces,
breakable/paddable/unpaddable combinations than groff, as far as
I can tell.
7) SQPS has a method for setting paper origin to compensate for
printers being unable to print to edges.
8) I frequently use .vs 0 in sqtroff for various kinds of
manipulations, but this produces an warning message in groff (a
range message?)
9) In groff a full-column-width line generates a "can't break
line" warning, which I don't think is appropriate since the line
is purposely the full width.
10) sqtroff has an x condition that tests for validity of a
numerical expression; this is often useful to determine if
the argument to a macro is a number.
Thanks again for any help. Despite these differences, I think
that groff is an extremely important program.
--Steve Izma, Wilfrid Laurier University Press
Waterloo, Ont., Canada N2L 3C5
(519) 884-0710 ext. 6125 FAX: (519) 725-1399
st...@nestor.wlu.ca or si...@mach1.wlu.ca
From: j...@jclark.com (James Clark)
Subject: Re: groff/SoftQuad-troff diffs
Date: 1997/02/17
Message-ID: <2.2.32.19970217041425.0096aca4@jclark.com>#1/1
X-Deja-AN: 219299794
Sender: bug-groff-requ...@prep.ai.mit.edu
X-Sender: j...@jclark.com
X-Authentication-Warning: jclark.com: Host j...@lox1.loxinfo.co.th [202.44.203.10] claimed to be durian.jclark.com
Content-Type: text/plain; charset="us-ascii"
x-gateway: relay1.UU.NET from bug-groff to gnu.groff.bug; Sun, 16 Feb 1997 23:23:22 EST
Mime-Version: 1.0
Newsgroups: gnu.groff.bug
X-Mailer: Windows Eudora Pro Version 2.2 (32)
At 14:27 15/02/97 -0500, Steve Izma wrote:
Thanks for your comments. Unfortunately I have very little time to work on
groff nowadays.
>1) I've become very dependent on SoftQuad's trace output for
>debugging macros; groff's trace output is minimal, as far as I
>can tell -- I usually have to insert comments in macro code to
>get reports at various places, but with sqtroff I can follow the
>execution of various commands and the values of any variables
>used.
I think tracing is more appropriately done with macros (by renaming the
to-be-traced macro and then defining a macro with the to-be-traced macro's
name that prints tracing messages and then calls the renamed macro). There
are one or two additions that would be useful for this, eg allowing a macro
to determine whether it was called with "'" rather than ".".
>2) sqtroff consistently allows square brackets for enclosing not
>just long names but also the arguments to backslash commands
>(e.g., \w \h \D \l, etc.); pairing and nesting of square brackets
>is easier to decipher than the quote marks required by groff.
This is not an extension that is compatible with troff. An argument to \w
that starts with [ is legal in both groff and DWB troff and is terminated by
a matching [ (not ]). Both groff and troff allow any character to be used
to delimit arguments to \w. I like many of the SQ extensions, but not this
one: I think it confuses two fundamentally kinds of escape sequence argument.
>3) To get morticing/kerning in groff appears to me to require
>using the tracking request and keeping track of font changes and
>other variables, whereas sqtroff has a simple (?) kerning request
>in .swd
I would have said the groff request was easier. You can just to .tkf once
at the beginning of the document. What are you using .swd for?
>4) groff has a definite advantage in handling fractional point
>sizes for fonts; however, I usually set number registers to be
>used for arguments to .ps and I can't figure out how to make the
>number register a fractional number (use 1 register for whole
>point sizes and concatenate another register for fractions? seems
>awkward).
.nr x 10.25z
.ps \n[x]u
>5) I can't find documentation for the hyphenation-exception
>dictionary (seems to be based on Tex,
It is.
>but I can't find info for
>that either); SQ dictionaries are compiled and less formulaic,
>but I suspect that the groff method is better. For some reason,
>all-cap acromyms that aren't English words get hyphenated.
Try asking in comp.text.tex: there is documentation available somewhere.
>6) sqtroff has a greater number of codes for fixed spaces,
>breakable/paddable/unpaddable combinations than groff, as far as
>I can tell.
I can't comment without more details.
>7) SQPS has a method for setting paper origin to compensate for
>printers being unable to print to edges.
I want 0,0 to be the top-left of the page whatever the size of the
unprintable area is. It would be reasonable to allow the DESC file to
specify the minimum margins.
>8) I frequently use .vs 0 in sqtroff for various kinds of
>manipulations, but this produces an warning message in groff (a
>range message?)
I think groff is just following Unix troff here.
>9) In groff a full-column-width line generates a "can't break
>line" warning, which I don't think is appropriate since the line
>is purposely the full width.
An example, please.
>10) sqtroff has an x condition that tests for validity of a
>numerical expression; this is often useful to determine if
>the argument to a macro is a number.
Since you can control warnings about validity of numeric expressions, I
don't see that this is necessary: you can just use the macro argument with
.nr and see whether the register has been set.
James
|