Seems I am on a downvote streak lately so I'll keep going; why would you not use (Xe)LaTeX? If I was too lazy to use that I would probably use something like bookdown.
Well, you could, at least for the print part, if that was what you were comfortable working in. LaTex is sort of notorious for being designer-unfriendly (something like ConText might be better if you want more design flexibility). But if you don't actually know anything about book design or typesetting (like most self-published authors) then at least it might keep you from making a true horrorshow by using a common stylesheet.
If you want to do epub/kindle publishing, then you need to be able to output that into the necessary html files those formats are constructed out of. This is a little chore, but a manageable one.
I am using LaTeX, but I'm almost certain it will be a pain when I go to the eBook step, since they all have their own formats and I'm not aware of any tools that can render the math formulas in my book faithfully on all of them.
This is a problem. In epub, you can use MathML, but support on reading systems is spotty. You could embed something like MathJax, but that would also depend on there being scripting support (not all systems have it). On the Kindle, there is none--no Latex, no MathML, no scripting, nothing for math support at all. Your only choice is graphics. SVG or high-res PNG works best. Though on a Kindle, there's no "S" in the SVG (i.e, the Kindle can't actually scale SVG). Occasionally we do a math-heavy econ book here, and it is always a compromise.
What is your opinion on publishing a math-heavy e-book (like, equations on every page, inline and offset)? I don't think I will be able to do better than having a physical book paired with a pay-to-download PDF.
I guess it comes down to what the market is, how much work it would take to do the conversions, and what compromises you're willing to take. Inline math is easier to deal with, usually it's just a matter of making sure the characters display correctly. Complex multi-line math is where things start to fall down. If this is a specialized technical reference that is of interest to 500 people, it might not be worth a lot of work to sell 20 copies on Kindle.
Thanks. I didn't know about it, but I think this makes Mezzano a failed project anyway. Sure it's good as a "proof of concept" to demonstrate CL's potential, but if I haven't heard about it, how can you expect an ordinary user to know, let alone, use it?
BTW, I'm strongly against the idea of creating new OS's just because we _can_. By end-user app, I mean something along the lines of a browser, fully capable back/front-end framework, essential industrial software, etc.
