The files in this directory were collected and modified by Scott B. Marovich, Hewlett-Packard Laboratories in February, 2004, who wrote the following notes. These are mostly "troff(1)" representations of SNOBOL-related documents written or collected by Ralph E. Griswold while the latter was at The University of Arizona after leaving Bell Telephone Laboratories. At Arizona, they were stored at the URL, http://www.cs.arizona.edu/icon/ftp/contrib/s4/ and were archived on 6/16/1998, shortly before they disappeared due to a disc head crash. Copies of some documents were subsequently saved by Phil Budne, then a graduate student of Griswold's, at the URL, ftp://ftp.ultimate.com/snobol/doc/{arizona,pb}/* They use the MS macro package (whether it's the original Berkeley UNIX version or an Arizona-modified version isn't clear), plus private macros of Griswold's that are apparently lost. Budne modified document S4D58 in order to format it using "groff(1)", and he tried to supply close equivalents of some of the lost macros. I didn't have a working "groff(1)", but I had a rich Device Independent Troff environment with a PostScript output capability, based upon Adobe Systems' old TranScript software and a licensed PA-RISC port of the A.T.&T. Documenter's Work Bench (DWB), once sold by Hewlett-Packard Company for HP-UX. Troff has long supported a "CW" (constant width) font in which to display source program text, but I lacked the "M", "MB" and "P1" fonts used by Griswold, and Troff can't map them to alternate fonts like "groff(1)" can. So, I hacked the source-document files more or less severely: 1. I created "defs.roff", "letterhead", and "reg.mac" files, which represent my best guess at Griswold's originals. "defs.roff" and "reg.mac" are actually linked together because (I think) one originally did a ".so" of the other one. I edited the ".so" directives in all source-document files to refer to these using relative paths. I created all "make(1)" files except Phil Budne's. These are intended for an ordinary A.T.&T. UNIX "make(1)" command, not "gmake(1)". 2. All references (everywhere!) to the "H", "M", and "MB" fonts are changed to "CW". In cases where SNOBOL program text is being presented, uses of the "B" font are also changed to "CW". A "\*M" -> "\fM" string translation is similarly replaced by "\f(CW". Use of the "fat line" drawing character in the "P1" font is replaced by use of Device Independent Troff's "\D'l...'" graphics line- drawing mechanism, where 8 parallel lines are drawn 1u thick and 2u apart; following TranScript's convention which defines a Troff "basic unit" as 1/576 inch, but assuming that your printer renders only 300 dots/inch, this yields an apparently solid line 8 pixels thick. Where possible I tried to concatenate "R" and "CW" text by introducing a ".CP" macro, which is a simplified version of the one originally used with the A.T.&T. UNIX System III "cw(1)" preprocessor; this lets a document be easily formatted using either "nroff(1)" or "troff(1)". But some documents use things like embedded "\fM" sequences pervasively, so I massively changed them to use "\f(CW" instead (i.e., when all else fails, use "brute force"!). 3. All printed dates are "hard wired" to each document's historical publication date, instead of using the "current" date that the MS macros and "?roff(1)" formatters print by default. I used "touch(1)" to insure that a document file's time stamp matches the document's publication date. (Griswold's bibliography of SNOBOL documents is very helpful for such internal-consistency checking.) 4. In document S4D58, Phil Budne made several minor changes that I preserved: adding The University of Arizona's name and address, and numbering the pages. Phil used a SNOBOL program to expand the "{cl,d,p,s}box()" macros in Griswold's source document, but I used a short file of M4 macro definitions because that seemed simpler. 5. I found and fixed a (very) few minor formatting bugs in some documents, which seemed to be originally due to Griswold. I hope he won't mind! Here's a brief list of files beneath this directory: defs.roff, reg.mac My attempt to reconstruct what were apparently Ralph Griswold's private overriding "troff(1)" macro definitions, now lost, but once stored in the files, "/usr/ralph/docs/{defs.roff,reg.mac}". letterhead, seal.ps My attempt to reconstruct "troff(1)" commands producing a facsimile University of Arizona Computer Science Department letterhead, which is probably what was once stored in the file, "/usr/pub/letterhead". These files are based upon originals supplied by Peter J. Downey, now Head of the Department of Computer Science at The University of Arizona (see directory "letterhead.downey/*"), for "groff(1)" with "psroff(1)". I modified them for Device Independent Troff with "psdit(1)". letterhead.downey/* Original letter head files supplied by Peter J. Downey. correg/* Troff files for corrections to SNOBOL4 books. s4b/* Troff files for a few "SNOBOL4 Information Bulletin"s. s4d/s4d43/* Troff files for SNOBOL4 document S4D43. s4d/s4d58/* Troff files for SNOBOL4 document S4D58, as edited by me. s4d/s4d58.budne/* Phil Budne's modified (for "groff(1)") and augmented version of the "s4d58/*" files. Comments from his FTP site, written 12/19/2000: "Includes a SNOBOL4 program to generate TBL commands for boxes, and a Makefile which runs SNOBOL4, TBL, groff (GNU n/troff) to format generate PostScript output, and using ghostscript, a .pdf file "Output is not identical to the original memo, but is very close; "The title page differs either due to Arizona local changes to the 'ms' macro package, or later drift at Bell Labs. "Line and page breaks are different, due to font differences differences in box size, or interline spacing. "I have made NO editorial changes. "If you redistribute modified versions of these files it would be a good idea to include information on their provenance and change history." s4d/* Troff files for various other S4Ds. s4n/* Troff files for various S4Ns.