Write it in Lyx and export to Latex. Pleasant editing with professional-looking result. Best of both worlds, and the journal won't know.

Pandoc is a great solution for converting Markdown into LaTeX and you can add arbitrary LaTeX code at the beginning and end of the output.

Personally, I've never encountered a journal that wanted a LaTeX file. Always the PDF.

I've had journals ask for both. However they claim to only use the LaTeX file as reference in case there are issues in typesetting. In any case, it's still necessary to produce a LaTeX file in order to use the LaTeX template.

Even the ones that will accept a PDF tend to require a PDF generated using their very specific template and style.

Do you mean for review, or even when they've accepted and ask for a camera-ready version? I always assumed all publishers need the source so they can do the small tweaks they do.

[OTOH, I've seen claims many publishers already have a process for publishing Word submissions — something involving monks spending their lives typesetting everything from scratch? — and so when they do get a nice LaTeX, they throw the source out re-use the above process! Perhaps that was only about the bibliography though.]

I've seen some Springer (LNCS) submissions, and those definitely included the .tex source.

arXiv requires the source, mainly so they can re-compile them in the future with better tools [https://arxiv.org/help/faq/whytex]. Unusually, they make sources available for anyone to download.

Earlier this year, I needed to upload the LaTeX file and it needed to be written with the journal's template. In order to specifically fit into the journal's guidelines, the template was sent to some editing contractor where some minor changes where made to the source, e.g. "eq." to "Eq.".

As far as I know, a lot of the journals have their own automated build systems for the manuscripts to perfectly fit into their designs.

Nevertheless, I've also encountered PDF uploads. But even then, the PDF will be automatically checked for consistency.

I believe most (all?) Elsevier journals require the LaTeX source, so you upload the .tex, .bib and any images etc. to their EES wepage, which compiles the PDF there.

One of the problems with accepting Markdown is that your work will still likely be formatted as a PDF and if you submit Markdown, you don't really know what that final output is going to look like. At least not without a lot of extra tooling on the part of the publisher.

That's also true about Word docx submissions, too. After acceptance, you approve their proof.

Even if your journal accepts PDFs, michaelmior's comment raises an important issue that a LaTeX style template cannot be imported into most (or all?) of these Markdown editors. This would be a great feature to add, IMO!

The issue with that is that a LaTeX template is turing-complete. Doing a meaningful import of arbitrary journal styles without re-implementing TeX is a very, very, very hard problem.

> Even if your journal accepts PDFs, michaelmior's comment raises an important issue that a LaTeX style template cannot be imported into most (or all?) of these Markdown editors.

Perhaps it would be easier to go the other way -- to modify a Markdown parser to output a simple subset of Latex that can fit into most common academic templates.