On 04.07.24 19:00, Carl Sorensen wrote:
The Alain and Rachmaninov only have one fewer beam, so the beam count between the groups is not appropriate for the lengths of the subdivided groups, according to the Gould rules. Personally, I think the Gould rules are correct, but two of the music publishers in your example do not agree with me. And they should probably have MUCH more credibility than I do.
I think this is a typical case where the modern rules are “better” in that they allow for even more clarity in all situations, even if they are even more complex than these. And the rules followed in those Alain and Rachmaninov examples were good for the time and internally consistent; they don’t have any advantages per se over the modern rules, except that those composers used them to convey their ideas and changing the graphical appearance of the music will cause some change in how the musician perceives it. Alain and Rachmaninov didn’t have need for more thorough subdivision.
Reger did: in his op. 73, even the first edition of 1904 already uses the full tool set of subdivision, over wide areas of the piece (starting, in this first edition, on the third page of music). https://imslp.org/wiki/File:Reger-op73.pdf
LilyPond should by default use the full modern subdivision, but allow limiting it.
Best, Simon
