Spatial validation rule 21350 too stringent

[luciansmith]

Rule 21350 only applies when the mapped compartments are the same dimensions, not if they are different (cf sbmlteam/libsbml#275).

The new rule should be:

"The value of the attribute \token{spatial:unitSize} on a \CompartmentMapping must have a value between 0 and 1, inclusive, when the dimensions of the referenced compartments have the same dimensions."

[luciansmith]

I'm inclined to make this change before submitting to JIB. @fbergmann , opinions?

[fbergmann]

If we change it, since really it is a best guidance rule, we additionally should change it from a must to a should. This is also what came out in discussions with @jcschaff. I'm happy for it to change in time for the JIB thing. it does not change anything form the spec, other than this validation rule.

[luciansmith]

Sounds good to me! Changed from 'must' to 'should', and downgraded it to a warning instead of an error.