posledni zmeny v TeX/MF

[Petr Olsak]

Vazeni priznivci TeXu,

protoze jsem prave pred chvili zminil, ze Knuth udelal v lete nejake zmeny
v TeX/MF, prikladam pro uplnost informaci, ktera presneji tyto zmeny
popisuje. Vyhrabal jsem to ze sveho e-miloveho archivu.  Vsimnete si, ze
skutecne Knuth nezmenil kod TeXu.

Petr Olsak

---------- Forwarded message ----------
Date: Mon, 10 Aug 1998 17:02:15 -0400 (EDT)
From: Barbara Beeton <bnb at ams.org>
To: tex-implementors at ams.org
Subject: note from Don Knuth (fwd)

dek has just finished the 1998 cycle of updates to tex and friends.
the new versions should be appearing at labrea.stanford.edu later
today.  ctan maintainers -- please *don't* post them yet.  don would
like to make sure everything is solid first.

as some of you know, the 1995 details got buried on my desk before
the details (and reward checks) were distributed.  i will try to
take care of everything, old and new, at the same time; since the
paper version for this round came to probably nearly a hundred
pages, it will take at least a month after i receive the annotated
version to transcribe the notes into the files and get them ready for
posting.  i will send a private message around as soon as i learn the
identity of the check recipients asking for current postal addresses.

one more piece of news -- phyllis winkler, who has been don's secretary
for as long as i've known either of them, is retiring on september 30.
i generally don't identify her as the source of messages from don, but
she won't be in that position much longer.  i know that at least some
of you have met her, and you might want to wish her a happy retirement.

i will be leaving on friday for the tug meeting.  even if the papers
with don's notes arrive here before i leave, i haven't time even to
look at them before i go.  but do look at the new errata and bugs
lists, and if you are an active implementor, try out the new code,
and report any problems at once, either through tex-implementors
or directly to me.  please put "tex 3.141592" in the subject -- i
may not be able to read my mail while i'm away, and there will be a
*lot* to look at when i get back; if messages about the new version
are properly identified, i will try to attend to them first, and
forward them to to don at once.
							-- bb
			--------------------

Dear Barbara,

Later today I will update the master TeX/MF/CM sources at labrea.stanford.edu,
based on the 1998 round of corrections. (I will also update my web pages
accordingly; the next round will be 2002, then 2007, 2013, 2021, etc.)
I hope a few people will be able to verify that the new stuff makes sense,
before the files are too widely distributed. I tried to be careful and
check everything carefully, but if I was only 99.9% accurate I made
at least one mistake! I suppose the new sources shouldn't go into CTAN
for another month or so.

The purpose of the present note is to summarize the most important
differences between the 1995 versions and the 1998 versions. Please
pass this information to whoever you think will benefit from it.
Meanwhile Phyllis is mailing you a hardcopy of all the things you
emailed to me, with my detailed handwritten notes on each point;
we also are sending all the related reward checks to you, since you
know best how to reach all the people involved. (The big-ticket item
this time is the check for $327.68 to Ulrik Vieth, who found the
bug in METAFONT/METAPOST.)

+ Changes to The TeXbook

The most important TeXbook corrections appear in the files errata.nine
(changes between the 1992 and 1996 printings, mostly made already in 1995)
and errata.tex. I didn't list insignificant changes in errata.tex;
for example, four of the entries in the index had the page numbers
out of order. A dozen or so trivial things like that have been corrected
in texbook.tex and not mentioned elsewhere, so people who think they have
found a bug (especially a possibly trivial one) should always check
texbook.tex before reporting it.

I'd say the most significant new changes are still pretty minor, namely
corrections to the answers to some of the tougher exercises and
refinements to plain.tex (see below). I did make an embarrassing TeXnical
error on the bottom of page A233 that nobody noticed until now!

+ Changes to plain.tex

The fmtversion is now 3.1415926, two steps ahead of the version of TeX
itself.  I corrected the definition of \AA so that it will work correctly
in caps-and-small-caps fonts. I improved the definitions of \d and \b
accents so that the slant-dependent shifting of fmtversion 3.141592 now uses
a much more useful and better motivated auxiliary \sh at ft macro. I corrected
the definition of \c so that \c{\bf c} would not turn on boldface. I
corrected the plain output routine so that the page number position in
a footline would not depend on the current setting of \lineskiplimit.
I introduced a new control sequence \Orb, which typesets the big circle used
in \copyright.  The \rightarrowfill and \leftarrowfill operators became
more robust (already in fmtversion 3.141592).

+ Changes to TeX: The Program

A few corrections were made to the comments, but the version remains 3.14159.

+ Changes to The METAFONTbook

Numerous nitpicky corrections are recorded in file errata.tex.
The only really important change is the correction to the syntax of
path expressions on page 129 (repeated on page 213).

+ Changes to plain.mf

A cosmetic change only; base version number remains 2.71.

+ Changes to METAFONT: The Program

Here we move to version 2.7182, correcting the bug involving unprintable
strings of length 1. I also made a small improvement to the routine for
Pythagorean addition. I'm told that the web2c change files already fix the bug,
in a better way than I did, so users will not notice any difference
between 2.718 and 2.7182 if the web2c folks didn't introduce an identifying
version number. But their change files should be modified to apply to the
new mf.web, and the version number should certainly be upgraded too.

+ Changes to Computer Modern Typefaces

These are, of course, documented completely in the file cm85.bug, as
well as in the changes to Volume E in errata.tex.
Most of the corrections simply make the programs more robust in the
presence of weirder combinations of parameters. Also, each of the
standard files now begins with a prominent notice that it should
not be modified in any way.

I claimed at the TUG Q&A in Florida that the Computer Modern character
shapes would never change again, but that turned out to be false. (Once
again, of course, I am sure that the changes just made will be the last!
And I can be a lot surer of that now because I've just finished typesetting
nearly 3000 pages of The Art of Computer Programming and Digital
Typography; also because the next round of corrections will be in 2002,
when I will surely not be in a mood to change ANYTHING!)

A few characters have changed in nontrivial ways: The lowercase beta
and omega have been tuned up to look better --- I did this for the
new editions of TAOCP. The uppercase C and G in sans-serif fonts don't
overshoot so much as before; this also was for TAOCP, after I saw words
like EXERCISES in sans-serif bold. And the position of the dots on the i's
in sans-serif fi and ffi ligatures have descended to the normal position
for i dots (this affects only the fonts cmss[1-9]* and cmssi[1-9]*, not
cmssb* or cmssd* or cmssq*). The division sign in cmtex10 doesn't have
such big dots.

All other changes are insignificant at high resolutions, and wouldn't
affect "Adobe Type 1" versions of the fonts. At low resolutions, I also
made a few improvements: The \circ and \bullet will look like circles,
not like American footballs, as sub/superscripts at 300dpi.

+ Changes to TeXware

DVItype is now version 1.6; it reports some errors better.

VPtoVF is now version 1.5, fixing a bug with respect to rules of
dimension zero (and also correcting typos/vagaries in the documentation).

+ Changes to MFware

I corrected typos in GFtoPK.web and PKtype.web, but kept the version numbers.

+ Other changes

In logo.mf, I redesigned the S, although it still isn't beautiful.
The S (for METAPOST) now sort of assumes that a T will follow.

++++++++++++++++++++++++++

In summary, I'm pleased that people still care enough about TeX/MF
to understand the details and to help me get them right.
But oh how I wish I hadn't made so many mistakes! The errata.tex file
now prints 16 pages; maybe it is time to publish it in TUGboat
once again, as in olden days.

Also, of course, I cannot thank you enough personally for the
extraordinarily marvelous way you have buffered and organized all these
materials for me to look at.

Best regards, Don

PS: Please tell our friends in the UK that I was wearing the
"1986--1996 TeX Archive of the United Kingdom" T-shirt while
composing this message.