My first book, the MIX book, was created on a computer system, using RUNOFF. RUNOFF was a popular markup language. It would do pagination, and page numbering, right margin justification, and so on. I wrote the book out long-hand, and then typed it into the computer, with RUNOFF commands for formatting. I could use this to print the book on a line-printer, where I could read it and mark it up for editting purposes.
My wife (now ex-wife) helped a lot in doing the editting.
The last print-out I had was from July 1977 and is 4 inches thick, weighs 14 pounds.
After the writing was done, I got a contract with a publisher, Academic Press, and could print copies, send them to the publisher, then incorporate their changes in my computer files, and send them new copies. Eventually, when the book was considered ready to print, I sent them a computer tape of the book, with mark-up information for fonts and such, expecting them to typeset from this tape. Instead, they printed out the contents of the tape, and then re-typed the entire book for the typesetter. I could tell since when they sent me the page proofs, there were a large number of errors in the book. This was a highly technical book, and, of course, the typesetters did not know the subject matter. The whole purpose of sending the magnetic tape was to avoid this sort of situation. Instead there were long days of going over the page proofs word by word, trying to find all the errors that had been introduced and fixing them.
For the next book, my Petri net book, technology had advanced. I moved to a Unix system with a different mark-up language -- troff -- which had the ability to work with math symbols. As with RUNOFF, troff could print to a line printer, but it could also use a new device: the daisy wheel printer. A daisy wheel printer was like a typewriter, but the actual letters were on a disk which would spin to position the correct letter to be typed
The advantage was that it was relatively quick to change the print wheel, and so you could change from a normal wheel to a wheel with symbols, allowing it to print Greek and math characters.
The printer could print on sprocket fed paper. I remember writing a program that would go thru a troff output file and replace all the Greek/math characters with a space. Then I would print that file. Then I changed the daisy wheel from the normal wheel to the symbol wheel. I would then re-position the paper back to the beginning and print a second file which had spaces for all the regular characters and contained only the Greek/math characters. The printer would space down, and over, and print the math symbol in the space that had been left from the first printing. With the sprocket feed for moving the paper, the registration was pretty repeatable, so things ended up printing, mostly, where they were supposed to be.
During the development of this book, I spent a year at MIT. MIT had a special graphics printer -- an early version of what are now common laser printers. I was able to get troff to drive that laser printer and get a version of the book that was almost as good as type-setting.
Based on my experience with the MIX book, when I was looking for a publisher for the Petri net book, I specifically wanted to avoid doing manual typesetting again. Prentice-Hall was willing to work with me on that, and found a printer in the Washington D.C. area that would take troff and use it to drive a photo-typesetter. I sent them a tape of the book, and that went really well.
When I got back to the University of Texas, there were no laser printers, but I became aware that the student newspaper, the Daily Texan, used a photo-typesetter for their work. And it was run by a PDP-11/70, which ran Unix. So I wrote a driver for troff to run their photo-typesetter, and was able to use it to do my third book myself, directly from my troff-based files to photo-typeset pages. The typeset output was pasted onto model sheets.
The fourth book, the OS book, brought all of this together. We wrote the book in SCRIBE, so that we could concentrate on the content and not the format. Then I translated the SCRIBE files to troff, and added a set of troff macros, to implement the book design from the publisher. By that time, the department had a laser printer, so we used the troff to drive the laser printer to get everything as we wanted it -- we had a devil of a time with widows and orphans! Once everything was ready, we ran the troff files on the photo-typesetter at the Daily Texan, and produced the page proofs to turn over to the publisher. We did the writing, editing, and typesetting all ourselves.
