convert-ly should work in both cases.
When you have a new version and you want to profit from improvements to
automatic engraving, in both cases you have to go through all manual
changes (tweaks) and see whether they are still necessary and whether
they do the right thing. To keep up with best practices, you would even
have to have the whole score in mind and see what can now be done in a
more elegant way (like \omit Clef etc. - the old syntax is not wrong
just more verbose).
The *pros* of the edition-engraver with respect to maintainance:
- You can keep the tweaks separated from the music, so it's easy to find
all of them and go through them
- You have to touch less files depending on your setup.
- Each tweak has a context and measure number so it's easy to find the
effect in the score. But point-and-click helps you in both cases.
The *cons* of the edition-engraver with respect to maintainance:
- You have an external dependency on the edition-engraver code
- You need it for your scores to compile
- If you share your files, the recipient also needs this code
- You need to trust that it will work with future versions of LP or
you need the skills to update it.
This list is most probably not complete. But that's what comes to my
mind. One could call it "independence vs. code purity". And my personal
opinion tends to the latter (i.e. using the edition-engraver).
I once started a visual diff between scores (could be used for comparing
the output of different LilyPond versions, different tweaks, etc.). I
think that would be a nice tool to see the main differences at a glance.
And using the edition engraver, you could easily compare a version with
and without tweaks.
lilypond-user mailing list