It varies by sub-area, but artificial intelligence is also my area, and I don't see that dominance anymore, especially in the more interdisciplinary areas (anything that overlaps with HCI, psychology, cognitive science, etc.). I also don't see the stigma anymore among younger researchers; I detect that attitude from older people mostly, and some of the "harder core than thou" people in math-heavy sub-areas, but there's a bigger mix of preferences among people under 35 who work in less math-y areas. I use TeX myself when it's my choice, but I've collaborated on Word papers as well, if I wasn't the primary author/instigator, and it seems common/expected these days. Especially if someone from industry has been the instigator (e.g. on DARPA-contract type research), or if it's interdisciplinary with someone not from CS/math, they've preferred Word.

AAAI, IJCAI, and AAMAS now provide both options, and my informal observation is that more Word papers are being submitted than used to be the case, especially but not exclusively when it comes to authors from industry. CHI recently officially deprecated LaTeX as a supported option, but still provides the old (no longer maintained) stylesheets as a courtesy. Several universities (e.g. Georgia Tech) have also stopped officially supporting LaTeX stylesheets for theses and moved to Word as the only official option, though they do distribute student-edited LaTeX stylesheets as a courtesy. I assume that one's because nobody in the IT department knows how to edit the stylesheets. The unofficial GT thesis stylesheet is a hilarious example of copy/paste cruft, too, with bits taken from 20-year old U. Texas stylesheets and various other places.

It depends on your particular flavor of CS. The less-theoryish flavors, like games or graphics, definitely do seem to be accepting of Word, and I've been to a couple conferences where Word was more prevalent.

Personally, I can't stand Word as an editing environment, because it doesn't offer a way of writing without worrying about formatting at the same time (seriously, if it had a markup view, I'd probably switch), and citation support is still a huge pain in the ass. However, LaTeX has many, many more pains in the ass for most people.

This kind of has been my experience as well in Statistics/EE/Machine Learning. While, IEEE/ACM have word style files, I have seen very few submissions that use word. In fact, I know reviewers who have had a significant bias (which whether it is right or not exists) against word submissions.

I agree that markup/stylesheets is what most strongly needs to be improved. I've only just started experimenting with it, but Biber, positioned as a BibTeX replacement, does really seem to be an improvement for the citation piece of the puzzle; its stylesheets look like something a mortal might be able to edit. It can also be used to generate HTML output, which has long been pretty crufty with BibTeX (I already have paper lists in BibTeX, so I should be able to generate things like my web page's publication list, and be able to customize the generated HTML, maybe even have some sorting/grouping options, all of which Biber can do).

If I had to predict, I would guess that something like HTML+CSS, or possibly a Markdown-ish input language with some kind of CSS-ish stylesheet, will eventually overtake TeX in significant areas, once the PDF renderers get good enough. Maybe someone will even find a way to pipe it into TeX as a renderer, as you suggest. A plus of that workflow is that it also makes it easy to produce good-looking HTML versions of articles, which is getting more important. And the "math in HTML" question is (finally) converging on a constellation of acceptable solutions. Advanced typography in HTML is slowly creeping forward as well.

Thank you for informing me about the existence of Pandoc.

At MIT Press they've moved to Word with the eXtyles plugin as their preferred manuscript/intermediary format, and InDesign for final layout. I'm not sure what precisely eXtyles does, but it appears to produce some kind of XML-based workflow that lets them retarget books to both InDesign and ebook formats, as well as process them through various scripts (e.g. for indexing).

They appear to still accept TeX manuscripts as a "not preferred" option, with their preferred approach to those being to convert them to Word+eXtyles ASAP, via some kind of processor they have. It looks like they still offer a full-TeX pipeline if authors insist, but it's deprecated and not done for most books anymore (and they won't produce ebook versions of your book if you choose this one).

More info on:

* The preferred pipeline: http://mitpress.mit.edu/authors/guidelines/monographs.asp#MS...

* The TeX options (I assume they put this information only in a Word .doc, when it's just plain text, out of a desire to taunt potential submitters of TeX manuscripts): http://mitpress.mit.edu/authors/guidelines/texscenarios.doc

LyX is ok for what it is. It certainly gets some use, even in publishing.

But at least for me a big part of the appeal to TeX is that it isn't a (%*#% GUI. It's text. I can edit in an efficient plain text editor, "refactor" styling, etc. I've even at times generated TeX code programmatically (It's often less painful for stuff like receipts than writing a PDF generator by hand, usually looks better too.)

So, I'd like a better input language, but I _don't_ want to fundamentally change the way it works. Just don't make me use a textual-replacement macro language.