TeXhax Digest Tuesday, March 29, 1988 Volume 88 : Issue 30
[SCORE.STANFORD.EDU]<TEX.TEXHAX>TEXHAX30.88
Editor: Malcolm Brown
Today's Topics:
SLiTeX on UNIX
Tib notes
LaTeX naming inconsistency
alternate page numbering style in LaTex
SLiTeX on UNIX
Resetting \textheight in LaTeX style-option files
help: need "proof" environment
Wanted: Tektronix 41xx Preview for VAX/VMS
!TeX on 3b2/400's
Re: Lowercase greek character in bold (TeXhax V88#25)
Re: Varying the Numbering Style in Enumerate Environments.
tex for at&t 3b1.
Metafont & IBM3812
Re: PATGEN
Re: !TeX on 3b2/400's
re: VMS and Apple LaserWriters
More Puzzles
CMBase/MF
Wanted: GNU Emacs BibTeX mode; BibTeX 0.99c
----------------------------------------------------------------------
Date: Thu, 17 Mar 88 13:11:41 EST
From: Charles R. Martin <crm@summanulla.mc.duke.edu>
Subject: SLiTeX on UNIX
Could someone clue me in on the state of SLiTeX for UNIX, and what must
be done to install it? I'd understood that it was available with this
newest release, but the documentation seems to suggest otherwise.
Thanks.
Charlie Martin
(...mcnc!duke!crm, crm@cs.duke.edu)
------------------------------
Date: Thu, 17 Mar 88 09:34:53 ]Io
From: James Alexander <alex@eneevax.umd.edu>
Subject: Tib notes
Steve Kille <steve@Cs.Ucl.AC.UK> writes:
We have noted a slight interaction problem between TiB and LaTeX.
In some of the TiB styles, including the default one, where the
.[] command immediately follows on from Section{}, the first reference
is shifted a few em over the left margin.
Is this a known problem? I would appreciate a fix.
It seems the \section command does something to the indentation. The
quickest way to work around it is to force TeX in and out of horizontal mode,
e.g.
\section{References}
\noindent\par
.[]
......
(see the discussions on \noindent and \par in the TeXbook).
John Sotos <SOTOS@SUMEX-AIM.Stanford.EDU> asks for a reference style
... the study of Smith et al.5
with the '5' superscripted a la a footnote.
kumard@cs.Buffalo.EDU (Deepak Kumar) asks for a reference style
[Nerd 1987] J K Nerd ...etc.
Tib is designed to make (a wide variety of) formats easy to implement.
The package contains examples and the documentation contains
instructions. In particular, there are superscript format files
included and several name-date formats. Those, possibly with minor
changes, should work.
------------------------------
Date: Thu, 17 Mar 88 13:49:18 CST
From: William LeFebvre <phil@rice.edu>
Subject: LaTeX naming inconsistency
To change the default separation between two columns in an array
environment, one should change \arraycolsep? Right.
To change the default separation between two columns in a tabular
environment, one should change \tabcolsep? Right.
To stretch out the default separation between two rows (i.e.: lines) in
an array environment, one should change \arraystretch? Right!
To stretch out the default separation between two rows (i.e.: lines) in
a tabular environment, one should change \tabstretch? WRONG!!!
One should change \arraystretch!
Why?
I imagine that there is a very good technical reason why this is, but
from a user-interface standpoint, it is blatantly inconsistent.
William LeFebvre
Department of Computer Science
Rice University
<phil@Rice.edu>
------------------------------
Date: Thu, 17 Mar 88 11:12:04 EST
From: Vince Pugliese <apollo%ecf.toronto.edu@RELAY.CS.NET>
Subject: alternate page numbering style in LaTex
we are interested in occasionally numbering pages in reports
using the convention:
section_# - page_# (e.g. 5-3 , the 3rd page of section 5)
or section_# . page_# (e.g. 5.3 , the 3rd page of section 5)
does anyone on the net know how to set do this in LaTex? our guess
is that one would have to play with the page counters to do this.
any suggestions (macros!) or advice much appreciated!
thanks in advance,
vince pugliese
apollo@ecf.toronto.edu
------------------------------
Date: Thu, 17 Mar 88 22:15:25 PST
From: mackay@june.cs.washington.edu (Pierre MacKay)
Subject: SLiTeX on UNIX
SLiTeX is a special subset of LaTeX, and has been part of the UnixTeX
distribution tape for 18 months, more or less. The macros and fonts
can also be got from the Archives on SCORE
------------------------------
Date: 18-MAR-1988 09:29:35 GMT
From: CCZDGR%VAXA.NOTT.AC.UK@forsythe.stanford.edu
Subject: Resetting \textheight in LaTeX style-option files
I'm trying to produce a LaTeX style-option file that would be appropriate for
theses on A4 paper. I'm thinking mostly in terms of 12pt, and hope that
Nottingham may follow Oxford (TeXhax V88, #02) in accepting LaTeX's standard
linespacing.
I started by looking at John Pavel's A4.STY in the LaTeX style collection.
It includes the following lines:
% Must preserve "(\textheight - \topskip) divides \baselineskip".
% \topskip always appears to be 10pt.
% So only add/subtract multiples of \baselineskip if you want shorter pages.
\topmargin 0 pt % Nominal distance from top of paper to top of page
\ifcase \@ptsize
% mods for 10 pt (\baselineskip=12pt)
\textheight 646 pt % Height of text (including footnotes and figures)
\or % mods for 11 pt (\baselineskip=13.6pt)
\textheight 649.2 pt % Height of text (including footnotes and figures
\or % mods for 12 pt (\baselineskip=15pt)
\textheight 655 pt % Height of text (including footnotes and figures)
\fi
However, in issue 101 of TeXhax vol. 87, Leslie Lamport says that
\textheight needs to be set to accomodate an integral number of lines.
Also, in REP12.DOC,
\textheight is 536.5pt
which doesn't seem consistent with John Pavel's assumptions [since
(536.5 - 10) isn't an exact multiple of 15].
Now 536.5 is an exact multiple of 14.5 (normalsize in REP12.DOC). Is the
convention in fact that, at 12pt, \textheight should be a multiple of 14.5?
(E.g. to implement Lamport's suggestion in V87 #101, one just adds 3 or 4
times 14.5 to 536.5.)
If so, could the comments in A4.STY be corrected?
David Rhead
------------------------------
Date: Fri, 18 Mar 88 15:48:02 EST
From: fred@ms.uky.edu
Subject: help: need "proof" environment
I am writing a mathematical document using LaTeX and need an
environment for proofs. Problem: since my proofs are set in roman type I
want to put a "box" at the end to show where the proof finishes and
ordinary text begins. My current attempt looks like this:
\newenvironment{proof}{\begin{trivlist}\item[\hskip\labelsep{\sc Proof.}]}%
{\nolinebreak\nopagebreak\hfill\rule{1ex}{2ex}\end{trivlist}}
Trouble is, if my proof ends with a displayed formula, the rule box goes
on its own line and I get this big blank space before the next chunk of
text. Even worse, if the last line of the proof is *full* of text, the
box goes at the *beginning* of the next line (I want it just inside the
right-hand margin). Worse still, in a couple of places the box goes at
the top of the next page. The \nolinebreak and \nopagebreak seem to make
no difference.
Anybody got a better way to do this?
Thanks in advance,
Martin Peim
------------------------------
Date: Sat, 19 Mar 88 00:51 GMT
From: Peter Flynn UCC <CBTS8001%IRUCCVAX.UCC.IE@forsythe.stanford.edu>
Subject: Wanted: Tektronix 41xx Preview for VAX/VMS
Has anyone got or heard of such a beastie? Would it mean another 32000
VAX blocks of disk for font files? Or is there any chance it would do a
nasty piece of interpolation and run with my 300dpi LN03 font files?
Peter Flynn <cbts8001@vax1.ucc.ie> | <cbts8001@iruccvax.bitnet>
------------------------------
Date: Fri, 18 Mar 88 20:46:01 -0500 (EST)
From: Michael Meyer <mm8s+@andrew.cmu.edu>
Subject: !TeX on 3b2/400's
Amongst other machines, I have 4 3b2/400's (still running sysV.3.0), with the
fpcc compiler (the one that supposedly generates code that uses the FP
accelerator). Off and on I've been trying to get TeX and a postscript device
driver working in this environment.
Well, I've now succeeded, and it was not too hard. I used Tim Morgan's
tex-to-C program to convert TeX 2.9 into C and then compiled it on the 3b2. (I
did the Tex-to-C conversion on a Vax where I had the weave and tangle programs
-- then copied the C files to the 3B2). In order to get it to compile I had to
make two minor changes to the C code.
1) I needed to break up a long switch in tex8.c (the compiler complained of a
switch table overflow) into two smaller switches
2) In extra.c sprintf() was declared as an ext char . I changed this to an
ext int (as in stdio.h).
I then had a working Tex with no extra fuss. This is really a triumph for TeX.
I am amazed at how easy the port was. I'd previously had problems with other
versions of Tex in C on the 3B2.
The next task was to obtain a working dv2ps. This was only slightly harder. I
started with a version of dvi2ps which uses the gf font files ( It claims to be
VERSION 2.10gf, though my comments probably apply to most other versions of
dvi2ps).
In dvi2ps.c I had to change the instances of index and rindex into strchar and
strrchar (a #define does nicely here). There was also a documented use of >>
on signed integers which had to be modified for the 3B2 (the code assumes >>
sign extends, whereas the 3b2 does not sign extend. K&R condone either
action).
In findfile.c instead of including <sys/dir.h> I used ndir.h from the Gnu emacs
distribution.
There were also a couple of places where things were delcared as int instead of
chars (or vice versa), but again the port was quite painless.
If anyone wants more details, feel free to send me mail, although I think I
have described jsut about everything I had to change.
Regards,
--Mike Meyer
Michael.Meyer@andrew.cmu.edu
------------------------------
From: F3WL%MCGILLC.BITNET@forsythe.stanford.edu
Date: FRI MAR 18, 1988 21.08.56
Subject: Re: Lowercase greek character in bold (TeXhax V88#25)
% I was wondering if anybody knows of an available font or of a way
% to get lowercase Greek letters to print in bold face.
There is a font "ambi10" which has the same character map as "ammi10".
DVILASER/HP has supplied the font at magstep 0, half and 1. In order
to get the lowercase Greek letters to print in bold face, one would
have to play with \fam and \mathchardef as described in the TeXBook
(Chap. 17, especially Ex. 17.20). I have come up with the following:
% define the font and a new \fam
\font\tenbfi=ambi10
\newfam\fambfm
% set up the fonts for the new family
\textfont\fambfm\tenbfi\scriptfont\fambfm\sevenbf
\scriptscriptfont\fambfm\fivebf
% redefine the lowercase greek characters be of class 7 (variable fam.)
\mathchardef\alpha="710B
\mathchardef\beta="710C
\mathchardef\gamma="710D
\mathchardef\delta="710E
\mathchardef\epsilon="710F
% and so on for others
% make a definition to call
\def\bfm{\fam=\fambfm}
% an example here
$$\alpha{\bfm \alpha}\beta{\bfm \beta}\gamma{\bfm \gamma}
\delta{\bfm \delta}\epsilon{\bfm \epsilon\epsilon}$$
\bye
Note that the above requires redefining all the greek characters
and as noted in the TeXbook, if someone types {\rm \alpha},
spurious results occur. Also, only normal text size is possible.
Therefore, user must be careful.
Wai Hung Leung
BITNET: F3WL@MCGILLC
B8AV@MCGILLC
------------------------------
Subject: Re: Varying the Numbering Style in Enumerate Environments.
Date: Sat, 19 Mar 88 00:30:10 -0800
From: Alastair Milne <milne@ICS.UCI.EDU>
I don't recall whether I ever got this to work, but I assumed that the
\begin{enumerate} instruction did all the defining of the number-printing
commands. It may therefore be necessary to \renewcommand{\theenumi}...
*AFTER* the \begin{enumerate}, rather than at the head of the document, or
in the preamble.
I know I've tried this, but I can't remember the result. Good luck.
Alastair Milne
------------------------------
Date: Fri, 18 Mar 88 22:54:50 EST
Subject: tex for at&t 3b1.
From: scott@golem.mef.unicus.com (Scott Thurlow)
I have recently heard rumors of a version of tex for the at&t 3b1. Does
anyone know anything about this?
Thanks.
Scott.
Scott Thurlow Golem Services
internet: scott@Golem.Mef.Unicus.COM
uucp: {utcsri!unicus,uunet!gwusun}!golser!scott
------------------------------
Date: Sat, 19 Mar 88 13:53 CST
From: <GHARRIS%AUDUCVAX.BITNET@forsythe.stanford.edu>
Subject: Metafont & IBM3812
I'm trying to generate GF files for our IBM 3812. I made a version of metafont
with the cm base loaded and a local file with mode_def settings from the
TUGboat:
blacker:=.4;
fillin:=-0.2;
o_correction:=0.4;
(G. Neumann (who has a fine 3812 driver) recommends other settings, but
his settings seem dark on our printer). The above works ok (although the
results don't seem to be as `crisp' as our PCTeX epson .pk files) until
I get to magstep3. Then I get the following from metafont:
%%%%%%%%%%%%%%%%%%%%%%%%%%
This is METAFONT, VAX/VMS Version 1.0.0 (preloaded base=cm 88.3.19) 19 MAR
1988 11:49
**&cm \mode=ibmulf; mag=1.728; input CMITT10
(DUA3:[SYS0.TEX.CM]CMITT10.MF;1 (DUA3:[SYS0.TEX.CM]TEXTIT.MF;1
(DUA3:[SYS0.TEX.CM]ITALL.MF;1 [97] (etc.)
% ...
(DUA3:[SYS0.TEX.CM]ROMANU.MF;1 [65] [66] [67]
[68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82]
[83] [84] [85] [86] [87] [88] [89] [90]) (DUA3:[SYS0.TEX.CM]GREEKU.MF;1
[0] [1] [2] [3]
! bad pos.
<to be read again>
;
pos->...XPR3)<=currentbreadth:errmessage"bad pos";
fi.fi(x(SUFFIX2)r-x(SUFFIX...
l.97 pickup crisp.nib; pos1(shaved_stem,90)
; pos2(hair,0);
?
! Emergency stop.
<to be read again>
;
% ...
Font metrics written on AU$DEPARTMENT:[JBTEST]CMITT10.TFM;1
Output written on AU$DEPARTMENT:[JBTEST]CMITT10.415GF;4 (62 characters, 8260 by
tes).
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
It seems that if I use fillin:=-0.1; blacker:=.2; o_corrections:=0.6;
then I don't get the problem. Can someone help me? Does anyone have any
other mode_def recommendations for the 3812?
--Darrel Hankerson, bitnet address: hank@auducvax
------------------------------
Date: Sun, 20 Mar 88 19:04:46 GMT
From: CET1%phoenix.cambridge.ac.uk@NSS.Cs.Ucl.AC.UK
Subject: Re: PATGEN
In TeXhax #26, Nico Poppelier asks what PATGEN.WEB is a part of. The
awnser is F. M. Liang's thesis, published as Stanford report number
STAN-CS-83-977, entitled ``Word Hy-phen-a-tion by Com-put-er'' (which
completely ****ed its entries in our library keyword index!), dated
August 1983. Pages 1-44 are the body of the thesis, the PATGEN listing
is the (first) appendix. It's all good stuff.
If you don't have access to a library getting these reports, then I
suppose you should write to Stanford, hoping that it is still in print.
Can anyone from there say?
Chris Thompson
JANET: cet1@uk.ac.cam.phx
ARPA: cet1%phx.cam.ac.uk@nss.cs.ucl.ac.uk
------------------------------
Date: Sun, 20 Mar 88 17:06:34 -0500 (EST)
From: Michael Meyer <mm8s+@andrew.cmu.edu>
Subject: Re: !TeX on 3b2/400's
Barbara Beeton asked the obvious question about the trip test for TeX(?) on
3b2's.
Here are the differences (in unix diff format) between my trip.log's and the
trip.log's in ps:<tex.test> on score.
There are a few differences in trip.log (mainly in the last decimal place of
some floating point numbers). I'll leave it to someone more knowledgeable than
I to declare if this really is a TeX or not.
Regards,
--Mike Meyer
Michael.Meyer@andrew.cmu.edu
First the differences in tripin.log. The < refers to my version. The > to
SCORE's version.
1,2c1,2
< This is TeX, C Version 2.9 (INITEX) 20 MAR 1988 16:39
< **trip
---
> This is TeX, Version 2.9 (INITEX) 24 DEC 1987 00:05
> **\input trip
82,83c82,83
< (preloaded format=trip 88.3.20)
< 1170 strings of total length 22663
---
> (preloaded format=trip 87.12.24)
> 1180 strings of total length 22718
Now the same ting for trip.log
1,2c1,2
< This is TeX, C Version 2.9 (preloaded format=trip 88.3.20) 20 MAR 1988
16:42< **&trip trip
---
> This is TeX, Version 2.9 (preloaded format=trip 87.12.24) 24 DEC 1987 00:23
> ** &trip trip
1280c1280
< \hbox(0.0+0.0)x2.0, glue set -1.99254
---
> \hbox(0.0+0.0)x2.0, glue set -1.99255
1426c1426
< ..\hbox(0.0+0.0)x2.0, glue set -1.99254
---
> ..\hbox(0.0+0.0)x2.0, glue set -1.99255
3144c3144
< \vbox(10000.0+2.0)x28.0, glue set 9730.58789fil
---
> \vbox(10000.0+2.0)x28.0, glue set 9730.5874fil
4262c4262
< ...\kern -5.50002 (for accent)
---
> ...\kern -5.50003 (for accent)
4265c4265
< ...\kern 2.50002 (for accent)
---
> ...\kern 2.50003 (for accent)
6989,6990c6989,6990
< 45 strings out of 1827
< 249 string characters out of 9317
---
> 43 strings out of 1817
> 229 string characters out of 9262
6992c6992
< 358 multiletter control sequences out of 3000
---
> 358 multiletter control sequences out of 2100
6995c6995
< 5i,7n,9p,111b,38s stack positions out of 200i,40n,60p,500b,600s
---
> 5i,7n,9p,113b,38s stack positions out of 200i,40n,60p,500b,600s
------------------------------
Date: Mon, 21 Mar 88 00:48:38 EST
From: jonradel%icecream.Princeton.EDU@Princeton.EDU (Jon Radel)
Subject: re: VMS and Apple LaserWriters
The most recent VAX SIG tapes from DECUS (the DEC users group) are
listed as having VMS printer symbionts for the LaserWriter. Since I
won't get my hands on these tapes until next fiscal year (9 days and
counting :-) ), I don't know what features they actually have, though
I'd imagine that message handling is one of them.
--Jon Radel
jonradel@icecream.princeton.edu
------------------------------
Date: Sun, 20 Mar 88 21:22 PST
From: Don Hosek <DHOSEK%HMCVAX.BITNET@forsythe.stanford.edu>
Subject: More Puzzles
I'm sending some more puzzles for all of you TeXnicians out there.
The first puzzle is:
The Singapore tourism board (or some such entity) has ads in the
Los Angeles Times in which two or three paragraphs are typeset with
the typeface alternating between roman and italics with every
character (spaces don't count). To illustrate, let's pretend that an
uppercase letter represents italic type and lowercase represents
roman. Then the phrase "come to Singapore" would be typeset as: "cOmE
tO sInGaPoRe". All punctuation is typeset in roman. The text typeset
in this manner may span several paragraphs and could conceivably
include TeX control sequences. The goal is to find an elegant way to
typeset text in the style of this ad. [[Disclaimer: The editor has no
association with the Singapore tourism board. In fact, he's never
been further west than Santa Monica]]
The sole response to this last TeX puzzle came from Dominik Wujastyk,
who wrote, "I have the answer to the SiNgApOrE problem: shoot the
designer." I'll leave this one open (and send it to TeXhax for a
wider distribution).
The next puzzle should be another interesting one: in prosified
editions of The Bible (such as The New Jerusalem Bible), if a verse
division occurs in the middle of a line, a centered dot is placed
where it occurs to indicate the location of the division. However,
if the verse division occurs between lines, no dot is printed. (Verse
numbers are indicated in marginal notes; I won't ask puzzle solvers
to supply a marginal note macro unless they *really* want to write
one, since Donald Knuth published one in TUGboat 8(3) and LaTeX has
its own built in.)
(Aren't you lucky to have someone like me around looking for weird
typography?)
-dh
------------------------------
Date: Mon 21 Mar 88 01:15:21-PST
From: Tomas G. Rokicki <ROKICKI@Sushi.Stanford.EDU>
Subject: CMBase/MF
Here's something I just noticed . . . don't include CMBase if you
don't need it. For instance, mf &cmbase logo10 will seem to work,
but the resulting characters will have a horizontal escapement of
zero pixels somehow. Oops! -tom
------------------------------
Date: 21 MAR 88 14:12-
From: KIESEL%DBNINF5.BITNET@CUNYVM.CUNY.EDU
Subject: Wanted: GNU Emacs BibTeX mode; BibTeX 0.99c
Hi there,
recently I read about a GNU Emacs BibTeX mode. I'm working a lot with
both GNU Emacs and BibTeX, so I'm *very* interested. I asked Ken and Oren,
but none of them heard of it. So I'm posting in the world!
I'm only on Bitnet (i.e. NO ftp), so either a direct mail or a usable
Bitnet server address would be nessesary.
so long
Norbert
P.S.: I'm also looking for the new BibTeX 0.99c using the same distribution
as above! Norbert
Norbert Kiesel @ Universitaet Bonn
Institut fuer Informatik
Abteilung II - Betriebssysteme
Roemerstr. 164
D-5300 Bonn 1
bitnet: kiesel@dbninf5.bitnet
uucp : ...!mcvax!unido!bnu!kiesel
------------------------------
End of TeXhax Digest
**************************
-------