[r-t] Does a rotation by any other name smell as sweet?

[Graham John]

Tim wrote:

> More importantly, if splicing these two methods became
> popular for some reason, and different composers
> produced various compositions, then without the
> ability to name both methods in the central library,
> composers would have to come up with their own
> notations to specify their compositions, thereby
> leading to a loss of standardization.

And 

> Furthermore, if you include a rotation of a method in
> a composition of spliced even without including the
> original method, if you can't name the rotation, you'd
> have to footnote the composition with something like
> "Where it says to ring method x, you actually ring a
> rotation of method x starting at change no. 6."
> This seems messy and makes it harder to store
> compositions in a machine-readable format.

This is not a problem. What is important is just that the base method
notation is held in the library. Then, if not using complete leads in a
composition, you need to state the start row number and the number of rows
of the method referenced. This is how I have handled this problem
generically in Composition Library. A composition has one or more method
definitions. The method definition is by default a lead of the chosen method
(by reference to the method library name) with a start row of zero and the
length of its lead. The start row and length can be modified, and if
required another definition can be added for the same method where different
starts are required. I also have a start row number for the whole
composition, which deals with a mid lead start when otherwise complete leads
of the method are rung. Having said that, there are very few compositions
that actually require the former flexibility at present as most compositions
just say snap start, or start at the 10th row of a lead.

> I'd echo some previous posts and say let's keep
> the existing rotation ban...

Careful. It is not banning rotations. It is just saying that we don't need
different names for rotations of a method.

> Final point: we allow reversals to be separately
> named - aren't rotations just another way of
> reordering changes in a lead?

For the vast majority of methods, reversals are identical because they have
palindromic symmetry. It is true that reversals of non-palindromic methods
could be expressed by a convention of adding Reverse/Inverse, but as these
terms have already been used for different purposes and there are probably
quite a few reversals already named, it does not seem worth changing this
convention. Furthermore, if we did this, why wouldn't we do the same for
place notation inversions (Reverse Bob is Plain Bob etc). 

Graham