[tex-k] note from Don Knuth

Maggie McLoughlin mam at theory.stanford.edu
Mon Jul 11 03:38:11 CEST 2005

Dear folks,

I tried using gftodvi today on my teTeX, for the first time in awhile.
It complains of (well, more precisely, it apologizes for) not being able
to make diagonal lines of slant 1/4.

So I tried "gftodvi --kpathsea_debug=-1 foo.2602gf" in order to see
what was going on. Hmm: that option isn't understood. I thought
all our utilities used kpathsea?

Then I figured, I must not have the slant font. I successfully made
slantlj4. Still got the same complaint/apology.

So I read the manual (texinfo) for gftodvi, according to the old adage
"If nothing else works, read the instructions."
Aha. I needed to say
  slantfont slantlj4;
at the top of my source file test.mf. (I was using the ztest program.)

OK, that worked.

But when I distributed TeX/MF in the 80s, and when I wrote
Appendix H of the MFbook, we always had a slant font installed
by default; for example, if the normal printer for hardcopy MF proofs
was a Laserjet 4 (in those days it was a Canon of course), there
would be a symbolic link like slant.600pk -> slantlj4.600pk.
The GFtoDVI program would try for a font named "slant" if no
special slantfont had been specified. Therefore my test programs
like 3test and 6test didn't need to worry about slanted lines.

Maybe, if people have an hour or so to spare one of these years,
one could make the process of getting hardcopy proofs a bit
more friendly. Maybe even have a little writeup that includes
a small walk-through of using mfw with rtest? (We used to distribute
a sample file test.mf that had a lowercase "thorn" character;
I don't know if that still is done.)

Hello to all from cool-and-sunny Stanford! -- Don

More information about the tex-k mailing list