Jn the mid 1990s I set up my doctoral dissertation, which was on digital circuit simulation at the anolog level, using nroff and eqn. It took awhile to get the hang of it, but once I did I could work pretty quickly. One of the nice things about the nroff-eqn package was that you could exert very fine control over the typesetting. I've also done a lot of mathematical writing using MSoft Word and its Equation 3.0 package. It has a WYSIWYG interface, which is OK, but its great weakness is that you can't cut/copy and paste to extract pieces of equations for use in others, so you have to re-create symbols from scratch at every occurrence. After these experiences, learning Latex, which I have done since participating in MO, was easy.