Dirk Laurie recently emailed me about inability to use mtrnuml=1 to represent a whole note time unit, (normally printed as "1" in denominator). I also discovered an email from him dated 10/98 on the same subject . I thought we had discussed it further at that time but couldn't find any more in my or the list's archives. In my current development version I initially fixed it to act the way he wanted, but I've just realized this creates backward incompatibilities. The table below, based on my experiments and coming with no guarantees, shows how I think Version 2.1 interprets 4 different inputs for mtrdenl and mtrdnp in 3 possible circumstances (new "m" is the syntax with slashes). -------------------------------------------- Circumstance -- input -- mtrdenl -- mtrdnp -------------------------------------------- setup -- 1 -- 16th -- "1" setup -- o -- error -- error setup -- 0 -- whole -- error setup -- 16 -- 16th -- "16" -------------------------------------------- old "m" -- 1 -- error -- error old "m" -- o -- 16th -- "1" old "m" -- 0 -- whole -- error old "m" -- 16 -- 16th -- "16" -------------------------------------------- new "m" -- 1 -- 16th -- "1" new "m" -- o -- error -- error new "m" -- 0 -- whole -- error new "m" -- 16 -- 16th -- "16" -------------------------------------------- Please don't ask how it got this way, I'm not sure I can reconstruct it. The good news is that every possible combination (16th or whole for mtrdenl, "1" or "16" for mtrdenp) is covered under any circumstance, and that "16" always represents a 16th note. The bad news is the inconsistencies, and the fact there's no way to remedy them without losing some backward compatibility. For that reason alone I'm tempted to leave things the same as in 2.1 . Dirk wants mtrdenl=1 to represent a whole note. On the other side, since that currently is interpreted as a 16th in setup and new-"m", changing it could cause incompatibility. Also, I don't see how to make "1" represent a whole note in old-"m", because when a "1" is encountered, the following digit is automatically appended. That's part of why new-"m" was installed. My proposal is to leave it as is, except that I could be fairly easily convinced to allow mtrdnp=0 ; that would make 0 always represent a whole note. Convincing me to make mtrdenl=1 equate to a whole will take some stronger arguments, but I'm willing to listen now that the considerations are on the table. --Don Simons
