Before the advent of LaTeX and other computerised word processing/typesetting tools, how would you have gone about writing a mathematics paper? Would the author have set it out on paper with a typewriter/handwriting and then sent it to a printers shop? Was there some other way? How would you ensure all the symbols used were available at the printers?
Magazines specializing in scientific papers would certainly have fonts for Greek letters and special mathematic symbols. Sending a paper into Physical Review would be no problem.
For writing your own papers or making your own illustrations, there were special printers that would make labels, or there were drafting letter set guides that you could use pantograph-style to trace Greek and other letters from a guide using an ink pen.
I did three theses without benefit of computers, and either hand-lettering, using such Letraset guides, or photocopying Greek characters. Someone looked at one of my theses years later and asked what software I’d used.
“Rapidograph”, I replied.
Occasionally, authors just typed double-spaced to leave room for superscripts and hand-drawn ilde’s and such, and left gaps for the characters like Greek letters that weren’t on their typewriter so they could fill them in later by hand. “Knots and Links” by Dale Rolfsen is a classic example of that: excellent book, cheap typesetting.
My father used to write the paper by hand using ink of different colors and various kinds of underlining to indicate special characters to the typesetter. The typesetter would then send him typeset proofs for him to correct.
As far as ensuring that needed symbols were available – the typesetters that were used specialized in scientific/mathematical publications and always had standard character sets available.
Is that available for Mac?
I kid… I’ve got a set of Rapidograph pens (actually, I think they’re K&E - please don’t make me spell that out from memory!) somewhere in the garage.
The old IBM Selectric typewriters also had interchangeable heads. One was available with symbols on it. You’d stop typing, change heads, and type the symbol; then change back to the regular head.
Yessir, those were the days, back when REAL typewriters had balls. 
Find some old papers from the 60’s and 70’s on the ACM Digital Library. While some look almost TeX-like, there are plenty that have been scanned in from typewriter-typed documents with hand-drawn figures and equations. You can also find old math monographs in the library that are typewrtier-typed. The ones I’ve seen have typeset equations but often have hand-drawn figures.
I wasn’t in math or science, but during my first go-round through grad school in the early 80s, we had
WYLBUR, and I assume it must have had sequences for equations and numerals as well as other typesetting features. You had to go to one of the campus computer centers to use it, and it was all greenscreen based.
While I just missed the pre-LaTeX era, I know a senior mathematical physicist who claims that the first time he came across TeX was when he was the external on a thesis that had been written using it: he was so gobsmacked with it, compared to what he was used to preprints and theses looking like, that he was prepared to pass the thesis on the basis of the typesetting alone.
Yeah, I have never heard of LaTeX before this thread. And I’ve been writing engineering reports since 1985.
I started publishing just as computers became common (1980s). However in the seventies there was another method apart from those mentioned.
In some cases the formulae or diagram was carefully drawn by hand and photographed. Alternatively Letraset transfers were used (anyone else remember it? http://www.pkfont.co.uk/images/Instant-Lettering-Sheets01.gif )
I started doing typesetting in the late '60s, when computerized typesetting was in its infancy. Everything was punched onto paper tape, and there was no monitor. What you saw were the last 20 characters/codes you typed, shown in LEDs, marching across a little band above the keyboard; everything else was in your head. As far as equations were concerned, the biggest problem was the lack of “reverse leading,” i.e. the baseline of a given character could not be higher than any preceding character in that line. So equations were set top-to-bottom, rather than left-to-right. We had to jump through countless mental hoops to get each character in the right place.
My thesis was written in 1962 and I typed it myself. Everything was double spaced. I rolled the platen up and down a half like to do sub and superscripts. I put in the Greek letters and other symbols by hand (on the original and all five carbons) and so on. Sometimes I overstruck–typed an o, backspaced and typed a / or - to get a phi or theta. The Forschungstitut fuer Math at the ETH in Zuerich had what was essentially two typewriters with a single double-sized platen. To type a symbol, you hit a special shift key that moved to the symbol side, typed the symbol and then shifted back. It was very slow. Later on, I experienced (although I didn’t use) two other systems. One, you could change balls on a Selectric and two, you had “Typits”, single characters on single keys. You inserted the key in the correct place with one hand and hit any key with the other. The inserts had complete sets of Greek letters and many many symbols. Typists could very fast with that.
Tex, which I discovered in the early 80s, was a revalation. I learned it right away although it was several years before I had a machine that could complie it. Even before that I had published a book that used a beta version of Latex. The editor had a computer that could compile it. Since then I have typed all my own papers.
However, it is probable that it is not the typesetting but the editing that has made the bigger difference in how I work. I used to have to prepare a manuscript that the typist could read and any editing thereafter was a recipe for disaster.
Now I use Latex (more accurately a mixture of plain and Latex) exclusively.
TeX was written by Donald Knuth. . .a god in the world of Computer Programming.
That article says he planned to finish it in 1978 and it wasn’t “frozen” until 1989. I read a story about it once referring to how even Knuth was susceptible to “software time”.
Some interesting stuff there about it’s creation, and what came before it.