From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <005501c079a8$257e3a70$62356887@HWTPC> From: "Howard Trickey" To: <9fans@cse.psu.edu> References: <20010108171937.8549919A1C@mail.cse.psu.edu> Subject: Re: [9fans] off topic: troff book MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Date: Mon, 8 Jan 2001 14:21:05 -0500 Topicbox-Message-UUID: 484447aa-eac9-11e9-9e20-41e7f4b1d025 From: "James A. Robinson" > He started working on what he calls a 'proto TeX' in 1977, and I believe > the PDP-11 showed up at Stanford in 1976. He states in the introduction > of TeX: The Program that he had many comments from Howard Trickey, who > later ported TeX to Unix at Stanford. I expect he could answer whether > or not Troff was known to folks at Stanford I didn't get to Stanford until 1980, well after the start of Tex. Stanford had a bunch of Unix machines by then, and I'm sure, even at the time when Knuth started Tex. So, though I don't know for sure, I find it hard to believe that Knuth wasn't told about troff as soon as he made it known within the department that he was going to work on typesetting. Knuth didn't work on Unix, however. He worked exclusively on the 'SAIL' system (running on a DEC-20), a rather strange system that had, among other things, a special terminal with a TV display, a keyboard with a lot of special characters (e.g., an 'alpha' key), and its own visual editor. There was no C compiler for SAIL. All I know about the start of Tex was that, as someone else noted, he was shown a galley for a reprinting of one of his Art of Computer Programming books, done with some 'new' computer typesetting system, and it offended his sensibilites. He really wanted his books to look like 'old-time' math textbooks, with their Modern-style fonts, and he was told by the publisher that he couldn't get it. I think he was also offended by the bad spacing in math formulas in the galley he was shown. So his motivation in doing Tex was to get (a) something that ran on SAIL; (b) something that used fonts like the books of his youth; (c) something that did a good job of spacing in formulas. Troff at the time would have needed a lot of modification to meet all of these requirements, and he vastly underestimated the amount of time that the Tex/Metafont project would take him. Knuth is a very parsimonious person, and I think many of the not-so-great choices in Tex were due to that parsimony. For instance, he hated wasting time, so he had Tex run with a chatty interface that let him edit the input stream on the fly as errors were encountered, so that he wouldn't have to start a run over. In implementation, he was forever squeezing things in memory (admittedly, the computers in 1977 didn't have a lot of memory to work with), so Tex77 only allowed single letter names for fonts. Etc. Knuth is also a perfectionist: thus, he would prefer to have Tex fail utterly with something so ugly that you have to fix it rather than allow it to output a paragraph with too much interword spacing. - Howard Trickey