The history of equation tags

Question

In plain TeX, one can add tags to equations by using \eqno and \leqno. Since the development of amsmath package, we can now use \tag and \tag* to assign tags to equations. So, in the period when amsmath had not come out, did TeX users have any other ways to set tags except using \eqno and \leqno?

Answer 1

Apart from some niche formats, before 1992 there was the choice between bare plain TeX, AMS-TeX and LaTeX.

For mathematical writing, AMS-TeX (with its wonderful manual by Mike Spivak) was essentially the only choice. LaTeX provided for automatic numbering and cross-references, but only offered \newtheorem, equation and eqnarray (the latter producing plainly wrong output). People using LaTeX for mathematics often resorted to plain TeX constructs, but these didn't support numbering: not the best framework.

To the contrary, AMS-TeX provided no automatic numbering nor cross-references, but had very flexible mechanisms to typeset theorem-like structures in different styles or alignment displays, to which it wasn't difficult to add tags. I only used AMS-TeX since when I could obtain the AMS fonts for the Macintosh (they came in floppy disks), about 1987, iirc. I knew about LaTeX and tried my hand on it, but it was definitely unsuited for mathematical papers (in my field, at least).

Around 1993 I started to study LaTeX because of the New Font Selection Scheme by Mittelbach and Schöpf that provided support for AMS fonts and so made it possible to port AMS-TeX under the name of amslatex; in order to practice, I produced my lecture notes for a new course using amslatex.

In the meantime I was helping my former advisor in the typesetting of his book (the basic typescript had already been produced in AMS-TeX by someone else) and implemented myself cross-references and automatic numbering.

In 1994 a colleague of mine asked help for typesetting a mathematical physics conference proceedings book, quite a tough task because the papers had very complex and long equations and alignments.

With the release of LaTeX2e that incorporated NFSS2 and sported amsmath the situation improved considerably; the transition was not painless: changing amslatex into amsmath didn't produce many errors, but several constructs of the former didn't really work with the latter yielding funny output. But once I understood what had to be changed (the @-shorthands, mostly), it was easy to fix the typescript.