1. 6
    1. 4

      I used to solve this (back when I was in academia a decade and change ago) by just using LyX which loses you the ability to have your text editor of choice, but gets you a lot of the nice features of LaTeX with a minimum of “and now I get distracted fiddling with the markup”

      1. 1

        Same, back when I was writing my thesis (in Emacs, of course) LyX was just starting to get viable.

    2. 4

      It may worth emphasize that LaTeX is a typesetting system. Always write in plain text first and only start typesetting it once you feel comfortable with the text.

      1. 5

        That depends a lot on what you’re writing. If you have code snippets, for example, then it’s a lot more effort to go back and annotate them later. I generally write in LaTeX-syntax semantic markup first and then go and write definitions of all of the macros that I’ve used.

      2. 2

        Hard seconding this. I have never written my first or even second draft in LaTeX. Imagine trying to compose a speech by sitting down at a typesetter’s bench with a composing stick and cases of type. Sure, it can be done, but why would you.

    3. 3

      Moving to Pandoc solved a lot of my LaTeX problems for writing basic documents (articles, white papers, etc.). It enabled me to get most TeX code out of the way and just write prose. I can easily drop in TeX when I need it and some light build system stuff makes setting up packages for things like headers and watermarks pretty easy.

    4. 2

      I personally really enjoyed the “write / compile” cycle of LaTeX because it cleanly separated content from presentation. And LaTeX markup is imho less intrusive than for example HTML.

      However I do believe that for some people the WYSIWYG experience of Word, LyX, certain programming environments etc is a very desirable thing, and TeX/LaTeX simply doesn’t work that way (again, I know LyX exists, but I don’t know how good it is for texts as long as say, a thesis). Maybe this is the case for the author.

      I acknowledge that there’s a mismatch between proofreading the typeset text and immediately fixing errors in the source - a quicker feedback loop is probably expected here. Again, I’m oldschool, and having a proofread session where you mark the errors and then fix them in a batch is how I was taught.