I can tell you that as a programmer with a music background, lilypond has been a godsend. As programming is my day job, many of my notation projects are long-term things that won't get touched for weeks. I have a backlog of 40-50 piano improvisations I want to notate, plus some pop/jazz originals for which I write various different versions of arrangements / lead sheets, not to mention a collection of altered lead sheets for my jazz group. And much older classical compositions from college, and a full-orchestra film score I did once.

With my git repo, I can see diffs over time, I can "compile" different versions of the scores, and by far and away most important, I have long term backwards compatibility. It was just last month I realized I missed porting an old flute duet from Sibelius to Lilypond, and I could no longer open Sibelius due to my version not running on my hardware anymore. Putting me in the position of having to buy an upgrade just to see it. Luckily I found one exported pdf, and that was enough. Now it's in Lilypond and I'm safe.

I have a workflow where I do my composing in Musescore, export to musicxml & convert with musicxml2ly to Lilypond, and finally do the typesetting there.

I basically throw away everything in the generated .ly files except the notes. But saves a lot of effort regardless.

I never found a a Lilypond -> Musescore conversion that worked really well. I did write a very simply tool based on python-ly to at least get the 'notes' to musicxml so that I can import something into Musescore.

> I basically throw away everything in the generated .ly files except the notes. But saves a lot of effort regardless.

That seems like a good workflow to me for the Musescore→LilyPond direction, and speaks to the incompatibility in approaches that make a more seamless integration less likely: LilyPond, as a typesetting engine, is essentially incompatible with a WYSIWYG workflow.

I did a bunch of engraving (probably around 100 works) when I ran a music publishing business. Many of the scores were organ and/or choral works, which are quite complex to engrave (especially compared to solo instrumental pieces). My workflow was to get the notes, words, articulations, etc. into LilyPond with no formatting tweaks, see what it gave me, and then tweak from there. But the defaults provided by the engine were almost always a very good starting point.

I would think that the other direction would be much easier, though, since that's basically the same workflow as copying plain text into Microsoft Word (albeit music notation is more complex than plain text).