Re: [complib-talk] Storing compositions on the database

[Don M. <df...@ri...>]

On Mon, Apr 27, 2009 at 6:12 PM, Graham John <gr...@ch...> wrote:
> In a printed composition, it seems that you have to work out was is intended
> from the given coursehead. Is it necessary to enter every coursehead for
> this type of composition to define it? This doesn't seem ideal. Is there a
> better way?

I think you either have to supply some indication of how to recognize
a course head, or specify the course's length. The canonical example
of the former is, of course, to write out the course head, though in
principal you could often supply a simpler rule, such as "tenor comes
home" or "back three bells come home in some arrangement". But,
particularly given the modern fashion of more interesting back bell
positions and even things like calling little bell runs in Stedman a
course end, this gets harder, and the whole course end may really be
required.

The alternative of specifying the length of the course is, mercifully,
becoming more common. Anyone that has tried to enter into proving
software a composition with unusual courses and back bell positions
understands how painful it can be having to work things out backwards
from a desired course end. Yes, if you're calling it, the discipline
of writing it all out by hand is probably a good one, but it works
against trying to enter something quickly to prove or format it when
you have no intention of calling it yourself.

I think working from course lengths instead of course ends has a
further advantage: it keeps you entirely in the realm of changes,
rather than getting down into the rows they generate. This is where
things like repeating blocks fall naturally, and it seems best to stay
away from the actual rows when defining a composition to the extent
one can.  Sometimes, of course, you have no choice but to operate at
the level of the actual rows; Dixon's, for example.

This problem doesn't just arise in things with numbered calling
positions. Consider something like Smith's 23, when it is notated as
calls at W and H, with the observation bell changing each part. Or
compositions on higher numbers with unusual back bell positions that
are using M,W,H with respect to whatever bell was last in last place
at a course end.

Anyway, I think in most cases, at least for single methods, it is best
to specify the length of every course when using numbered calling
positions; or, rather, to specify a default course length, and then
make explicit any deviations from it.



-- 
Don Morrison <df...@ri...>
"The ships hung in the air in exactly the same way that bricks don't."
                -- Douglas Adams, _The Hitchhiker's Guild to the Galaxy_