Helmut Kopka's interpretation of the TDS

Paul A Vojta twg-tds@mail.tug.org
Mon, 25 Nov 1996 22:53:47 -0500


> Date: Mon, 25 Nov 1996 10:18:17 GMT
> From: Sebastian Rahtz <s.rahtz@elsevier.co.uk>
> To: twg-tds@tug.cs.umb.edu
> Subject: Re: Helmut Kopka's interpretation of the TDS
> 
> Paul A. Vojta writes:
>  > That is, as far as I can tell, the benefits to users are all negative:
>  > 
>  >    o	TeX and its support programs become bigger and slower.
> i don't see why; we know they are not slower,

Hold it right there.  Here are timings for a typical paper of mine, done on
a 486-33 using Linux 2.0.25:

4.730u 0.240s 0:05.29 93.9% 0+0k 0+0io 92pf+0w
4.700u 0.300s 0:05.36 93.2% 0+0k 0+0io 92pf+0w

But, if I do:

	touch texmf.cnf
	mkdir tex fonts
	find /usr/local/tex/texmf/tex -type f -exec ln \{\} tex \;
	find /usr/local/tex/texmf/fonts/tfm -name \*.tfm -exec ln \{\} fonts \;
	setenv	TEXMFCNF	.
	setenv	TEXFORMATS	/usr/local/tex/texmf/web2c
	setenv	TEXINPUTS	.:tex
	setenv	TEXFONTS	fonts
 
then the timings become:

4.520u 0.260s 0:05.18 92.2% 0+0k 0+0io 92pf+0w
4.510u 0.150s 0:04.98 93.5% 0+0k 0+0io 92pf+0w

This is not a noticeable difference, but it is measurable, and the differences
add up after running tex a few thousand times.

> and the extra size is surely trivial?

Maybe.  It could be a problem under MS-DOS (they have a 640k limit on main
RAM; I don't know to what extent emtex and others use extended or expanded
memory).

> if i want to look at plain.tex, i do "vi `kpsewhich tex plain.tex` ":-}

Interesting.  I'll try to remember that.  But, how about:

grep \\\\somemacro /usr/custom/tex/inputs/{plain.tex,ams*}

--Paul Vojta, vojta@math.berkeley.edu