🚀 go-pugleaf

dave@cs.arizona.edu (Dave Schaumann) writes:


>I am interested in a language called `TRAC'.  Unfortunately, the only language
>referece I have is a photocopy out of a textbook with no title, author, or
>bibliography.  There is a footnote that says "TRAC is [...] being maintained
>by the Rockford Research Institute Inc., Cambridge, Mass."

>If someone knows of relevent articles or ftp sites, I'd be obliged if you could
>pass them on.

The textbook is probably Peter Wegner's "Programming Languages, Information
Structures, and Machine Organization", McGraw Hill, 1968.

TRAC is the brainchild of Calvin Mooers, who was active in the early days
of the PC. I haven't heard anything about him in years. He was noted for
espousing an unusually strict view of copyrights as applied to programs.
I suspect that he is dead and with him the Rockford Research Institute, but
this is pure speculation on my part.

There is a TRAC spec, published by the Rockford Research Institute in 1972.
I have seen a copy.

Some other sources of information are Ted Nelson's "Dream Machines", which
was republished by the Microsoft Press a few years ago. This is not a
particularly good reference, but it is something.

There is a version of Emacs for DOS machines called "Freemacs". Freemacs
uses a built-in programming language called Mint in much the same way
that Gnu Emacs uses a variant of Lisp. Mint IMHO is very similar to TRAC.
The Freemacs documentation includes a reasonably complete description of 
the language.  I know that Freemacs, until recently at least, could be 
found on an archive server at Clarkson. Getting a copy of Freemacs may be
your best bet.

Fred

i found a description of TRAC(tm) in the second
edition of _macro processors_ by a.j.  cole
(cambridge university press, ISBN 0-521-28560-7).
it is nowhere near a specification, but it does
give the flavor of the language, and you could
probably implement a similar language with the
description provided.

references cited by cole include:

	mooers, c.n. & deutsch, l.p. _TRAC, a text handling language_,
	pp 229-46 proc. a.c.m. 20th nat'l conference, 1965.

	mooers, c.n., _TRAC, a procedure-describing langauge for
	the reactive typewriter_; pp. 215-19 comm. a.c.m. vol. 9,
	no. 3, 1966.

	mooers, c.n., _reply to letter "language protection by trade-
	mark ill advised" by b.a. galler_, comm. a.c.m. vol. 11,
	no. 3, 1968.

i'm fairly certain that TRAC(tm) is indeed a
trademark.  mooers is (in)famous for his
insistence that the trademark symbol accompany use
of the language's name.

sam76 is/was a similar language; it was described
in a couple early issues (~1977) of dr. dobb's
journal.  sam76 may have had a brief revival on
the pc in the early 80's, but i am not certain of
this.


-- 

landon dyer
apple computer newton team
not the views of apple computer

I fiddled with TRAC in the mid-sixties.  I thought it was way cool.
I don't remember much, except that it stood for "Text Reckoning and
Compiling," it was a VERY pure macro string-replacmeent environment,
it had some of the flavor of things like the UNIX shells and the
C preprocessor and STAGE2.  I think the main superiority was that
it carefully distinguished between levels of quoting and expansion
in a way that most macro processors don't.  I believe \(expression)
would evaluate the expression and insert the resulting text, while
\\(expression) would evaluate it, RESCAN it, and insert it, or
some such.

I would suggest that maybe Peter Deutsch might know.  I seem to have
an email address for him:

   deutsch@SMLI.Eng.Sun.COM

which puzzles me, actually, since I thought he was working at PARCPlace...

-- 
Daniel P. B. Smith
dpbsmith@world.std.com

TRAC was first documented in an article in the CACM in about 1965 by Cal
Mooers and Peter Deutsch.  He later printed up some defining documents
under the imprint of Rockford Research, Mooers' little company.  Rockford
Research is long gone, but maybe Mooers still has copies. I've been
meaning to get back in touch with him, since he lives about three blocks
from here.

Trac is an interesting language for several reasons.  Technically, it is
the only macro-generator I've ever seen that makes a clean distinction
between macro scanning and expansion on the one hand and I/O on the other.
On the legal front, Trac was the first language where someone used
trademark law to try to regulate the quality of implementations by not
letting people use the trademark unless the implementation was up to
snuff.  This predated the DOD's use of the Ada trademark by about 20
years.  (The first trademarked language name was Fortran, but IBM gave up
the trademark before 1960 and never made the use of the name an issue.)

There were lots of Trac intepreters over the years.  I worked on two of
them with Peter Eichenberger before 1970.  One was for the PDP-10,
originally with the RESISTORS, and later at Yale, where as an
undergraduate I used it to prototype an APL to Basic compiler in 1972
(which I later learned was at the time probably state-of-the-art) and to
write what may have been the first outlining word processor which I used
to write my PhD thesis.  We also did a PDP-11 Trac processor which Peter
spiffed up as a summer student at Western Electric and may have been given
to DECUS.  Other notable Trac processors include Deutch's original PDP-1
version, one that Mooers' did for the PDP-10, and Claude Kagan's greatly
expanded version for the PDP-8, PDP-10, and later Z-80 (they shared
considerable source code, using a lot of macros.)  The Z-80 version turned
into sam76 when Kagan had a falling out with Mooers.  Kagan usually used
his middle names Anselme Roichel on sam76 documentation to avoid giving
the erroneous impression that the sam76 work was related to his day job at
Western Electric research.

Preston Briggs at Rice (preston@rice.edu) whipped up a little Trac
interpreter about a year ago which he would probably be happy to give
away.

Regards,
John Levine, comp.compilers moderator
johnl@iecc.cambridge.ma.us or {spdcc|ima|world}!iecc!johnl

[btw, "MINT" stands for "Mint Is Not Trac".]

I can recall a textbook with nice programming programming projects,
one of them being a TRAC interpreter.  The language description is
sufficiently precise to get a feel for trac but it is by far not
a reference manual.

	C. WETHERELL:
	ETUDES FOR PROGRAMMERS
	PRENTICE-HALL
	ENGLEWOOD CLIFFS
	0-13-291807-2
	1978

Russ Nelson's - or was it Nelson Russ? - mint implementation and its
documentation and tutorials coming with his Freemacs are *very* well
done.  I can only support Fred's suggestions, TRAC didn't look different
at all to me.

Martin Neitzel