[comp.text] tex82 or ctex2.9?

ken@gatech.edu (Ken Seefried iii) (07/05/88)

Which is better to use, the original TeX82 in pascal from Stanford, or
ctex2.9?

By better, i mean things like speed, public domain driver support, stability
(bugs per square inch ;'}), etc.


	ken seefried iii	...!{akgua, allegra, amd, harpo, hplabs, 
	ken@gatech.edu		inhp4, masscomp, rlgvax, sb1, uf-cgrl, 
	ccastks@gitvm1.bitnet	unmvax, ut-ngp, ut-sally}!gatech!ken

	soon to be open: ...!gatech!spooge!ken (finally ;'})

ken@cs.rochester.edu (Ken Yap) (07/05/88)

|Which is better to use, the original TeX82 in pascal from Stanford, or
|ctex2.9?
|
|By better, i mean things like speed, public domain driver support, stability
|(bugs per square inch ;'}), etc.

All TeXs generate the same DVI format so a driver will work equally
well (or badly) with any TeX.

Unix TeX now comes with WEB to C so Pascal is no longer needed. It
passes the trip test on the most common Unix boxes. It is about 20%
faster than the Pascal version.  The advantage of this version is that
it can track bug fixes in the original WEB code as and when Knuth
announces them.

Ctex 2.9 is claimed to pass the trip test and I could be convinced it
does on the original machine, a PC. However I couldn't get it to pass
trip on my Sun.

Whenever TeX is ported to a new machine, it must be validated by the
trip test, because the test not only exercises the particular
translation but also the compiler system.

	Ken

mark@motcoh.UUCP (Mark Mills) (07/05/88)

In article <11071@sol.ARPA> ken@cs.rochester.edu (Ken Yap) writes:
>|Which is better to use, the original TeX82 in pascal from Stanford, or
>|ctex2.9?

What's the best way to obtain ctex2.9?


-- 
Live: Mark Mills, Motorola Microcomputer Division Inc.
Uucp: mark@motcoh.UUCP
USPS: 6563 Worthington-Galena Rd.  Worthington,  OH. 43085
Tel:  (614) 846-9460  	Disclaimer: My views are personal.  

rusty@velveeta.berkeley.edu (rusty wright) (07/06/88)

The requirement for a tex port to pass the trip test cannot be
overemphasized.  It may not be important to everyone, but I derive
great comfort knowing that I can upgrade to the latest version by just
ftp'ing to stanford to get the new .web file and just changing the
version number in the .ch file and uttering "make".

One thing that Ken did forget to mention is that with the web2c setup
you not only get tex82 in C but you also get Metafont and a lot (not
all [yet]) of the TeX and Metafont support programs in C.

--------------------------------------
	rusty c. wright
	rusty@cartan.berkeley.edu ucbvax!cartan!rusty

kwok@iris.ucdavis.edu (Conrad Kwok) (07/07/88)

In article <11071@sol.ARPA> ken@cs.rochester.edu (Ken Yap) writes:
>
>Unix TeX now comes with WEB to C so Pascal is no longer needed. It
>passes the trip test on the most common Unix boxes. It is about 20%
>faster than the Pascal version.  The advantage of this version is that
>it can track bug fixes in the original WEB code as and when Knuth
>announces them.

I compile common tex 2.9 using "gcc -O". That is 40% faster and
20% smaller than the PASCAL version of tex. I didn't try it on
the WEB2C version. But I think the result would be similar.

Another advantages of using common TeX 2.9 is that they have a
BIG and BIGG version. BIGG version has a memory size about five
times of TeX.

--Conrad

internet: kwok@iris.ucdavis.edu
csnet   : kwok@ucd.csnet 
csnet   : kwok%iris.ucdavis.edu@csnet.relay
uucp    : {ucbvax, uunet, ... }!ucdavis!iris!kwok

ken@cs.rochester.edu (Ken Yap(probably)) (07/07/88)

|I compile common tex 2.9 using "gcc -O". That is 40% faster and
|20% smaller than the PASCAL version of tex. I didn't try it on
|the WEB2C version. But I think the result would be similar.

I think I would take correctness over any small gain in speed.

|Another advantages of using common TeX 2.9 is that they have a
|BIG and BIGG version. BIGG version has a memory size about five
|times of TeX.

I have a changefile (originally from Pierre Mackay) for a big version
of TeX in C, with 256kb of memorywords and it can easily be edited for
more. Without preloading, which wins very little in speed, according to
my tests, the big-tex text (code) size is actually *smaller* than the
normal version because most short/long conversions have been
eliminated.  The virtex binary is only 180k. The bss area is allocated
from VM anyway.  It runs just as fast as the normal version.

Tests indicate no reason not to install just one, non-preloaded, big
version of TeX.  No more out of memory messages (until the users get
overambitious again, anyway).  No need for separate tex, latex and
slitex binaries either. There is just initex and virtex now.

	Ken

PS: Don't ask me for the changefile. It is (will be) in the
distribution.