Path: utzoo!utgpu!water!watmath!clyde!bellcore!rutgers!ucsd!nosc!helios.ee.lbl.gov!lll-tis!ames!hc!beta!hwe From: hwe@beta.UUCP (Skip Egdorf) Newsgroups: comp.arch Subject: Re: PDP-10 user I/O (really: third-party operating systems) Summary: RE: Really old codes -> troff Message-ID: <19840@beta.UUCP> Date: 4 Jun 88 02:46:47 GMT References: <3327@phri.UUCP> <125@daitc.ARPA> <23004@bu-cs.BU.EDU> <3330@phri.UUCP> Organization: Los Alamos Natl Lab, Los Alamos, N.M. Lines: 34 In article <3330@phri.UUCP>, roy@phri.UUCP (Roy Smith) writes: > In article <23004@bu-cs.BU.EDU> bzs@bu-cs.BU.EDU (Barry Shein) writes: > > Depending on exactly where you start counting from Unix is getting > > close to 20 years old > > More amazing is that troff (if you count its roots back to the old > RUNOFF programs on TOPS-10) is probably even older than that. Actually, TOPS-10 RUNOFF and Unix's ROFF are both siblings on the family tree of the MIT roff program by J. Saltzer. Roff was the stylistic insperation for troff which was followed by nroff (for preview of troff on 'N'ormal terminals). Also of like age is vi / ex / ed / qed ... back to where early teco split from the same family, leading to ITS teco, leading to emacs... They are all related to some extent. Perhaps someday enough of this folklore will be lost to where computer anthropologists (You don't think they are alive??) will be holding learned conferances about troff Robustus and ed Habilis and the relationship between the two. This is getting a bit away from 'arch'itecture and I would like to draw it back a bit. Some reading this might think that the history of our field is not of particular interest. I disagree, and would recommend that at least at the master's level (if not the undergrad), a history of computers course would be useful. I don't mean just the big historical systems read about in an OS or computer architecture course, but a more 'history' like class including the economics and internal political struggles of the industry. I wonder how many of the undergrads drooling over the power of SunOS 4.0's mmap() recognize Multics in the shadows? Skip Egdorf hwe@lanl.gov