Partially filled measures

[rpatters1]

If a measure is only partially filled, is it understood to have a spacer that fills it, or is it an error?

[adrianholovaty]

I'm actually inclined to go with a third option, that it's permissible and left for consuming applications to deal with per their own standards. A consuming app might want to reject the document, or fill in spacers, or ask the user what they want to do.

The MNX spec doesn't say anything about this, and I'm not 100% sure we should take a stance. Some notation software deliberately doesn't enforce rhythmic stability within measures, while other software is militant about it. Seems to me it's a better approach to allow for both possibilities instead of limiting it to one.

Are you asking this in context of your MNX validator? If so, perhaps that could be a user preference?

[rpatters1]

In that case, is it an error to overfill the measure?

One of the reasons I want to include this in my validator is that it may flag errors in the encoding program, which is my primary motivation for including it. Overfilling is definitely an error for my encoder at least. But underfilling is actually possible in normal operation. I was trying to decide if I should encode an extra spacer at the end. (The underfilling situation is seemingly legit for encoding v2 passages in Finale that only exist in fragments of the measure.)

[samuelbradshaw]

Maybe there could be a property on the measure that specifies that it's a partial measure? Then it would be valid to have a partial measure with that property, but invalid otherwise. I think Finale uses partial measures when there's a system break in the middle of the measure (maybe other programs too). There are also pickup measures at the beginning of a song and pickdown measures at the end.

[rpatters1]

In the v1/v2 situation with Finale, only certain voices need to be partial. I'm thinking the sequence should have the option rather than the measure. (Or in addition to the measure.)