Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!uwm.edu!linac!att!att!cbnewsl!npn From: npn@cbnewsl.att.com (nils-peter.nelson) Newsgroups: comp.text Subject: Re: Q: nroff character sizes Summary: Brief ?roff history Keywords: nroff term Message-ID: <1991Jun13.200408.28043@cbnewsl.att.com> Date: 13 Jun 91 20:04:08 GMT References: <2133@ecicrl.ocunix.on.ca> Organization: AT&T Bell Laboratories Lines: 35 Since the parentage of nroff is in question... I have two unfair advantages-- I was around when it was born, and I have access to the on-line data base of all Bell Labs memos ever written. runoff was the original grandaddy. I recall it was written by Salzer at MIT in the 1960's. Generated fixed-width line printer output, but did right justification, vertical spacing, etc. The date on the first nroff memo is 1973. It was intended to take advantage of the Model 37 teletype, which had lower-case letters (a breakthrough!), half-line spaces, reverse line-feed. The first troff memo is dated 1974. It was designed around the features of the C/A/T phototypesetter (no lasers then-- it had a rotating wheel with tiny film images of the characters.) Joe Ossanna wrote both programs. The essential difference between nroff and troff is troff's ability to handle variable-width characters, and to move freely around the page vertically and horizontally in small increments (1/720 inch). Thus, a "fancy" nroff that takes advantage of new printers would, in fact, be troff. Were the original poster to do the modifications to nroff he suggests, he'd get ragged right margins only, since nroff believes all characters are equal width and adjusts accordingly. For those who still want to venture into this, it is very easy to modify the troff font width tables in DWB 3.1 and 3.2 (they are readable ASCII files in a self-evident format) to agree with your favorite printer, call it prt. Then "troff -mm -Tprt file | dnroff" would do what you want (need width tables under /usr/lib/font/devprt).