Re: Proposed CG agenda changes

Hi Peter,

Much of my "pushback" was based on my perception of large past differences
> of opinion between myself and Michael Good as to what constitutes a
> "specification" (and what constitutes an acceptable level of conformance to
> one).  Is it intended that this initial specification of MusicXML 3.1 meet
> the criteria in my earlier posting, repeated below?  Or will something less
> be considered acceptable?
>

First, I think we should set aside past differences of opinion. Although
our starting point is MusicXML, this is a new group with a different
mission and different governance. Our stated purpose from the outset is to
create a specification that can be defined, checked and tested in the way
that you propose. The only question is: what is the best time to do this?

As you know, the Chairs first proposed that we actually begin with a spec
for MusicXML as it exists on the ground today. (MusicXML 3.1 has been
effectively downsized to the inclusion of new SMuFL code points, so for
specification purposes it's approximately the same as 3.0.)

But in our discussion, as we gauged the demand for addressing fundamental
issues within MusicXML, we came to feel that those changes would be deep
enough in nature that we should seek to understand them first, beginning
with use cases and proceeding to solutions for those cases we choose to
address. Identifying and fixing these basics will make a clean
specification much easier to achieve. I feel we are quite likely to either
remove or substantially alter the very constructs and approaches in
MusicXML that make a clean spec most problematic today. Do we want to spec
things that are going to be reworked to this extent?  And my personal
opinion (perhaps not shared by the other co-chairs) is that crucial
portions of any 3.x spec would likely either be throwaway, or to turn out
to be almost impossible to nail down, burning lots of cycles in the process.

This is not to say that we should not carefully attempt to specify certain
aspects of MusicXML as we go down this path.  We will have to do that. But
right now, I think that doing this comprehensively for MusicXML 3.x will
subtract energy from the bigger goal of building a much more solid
foundation for what comes next.

In any case, more discussion on the priority of a spec for 3.x is very
welcome.

.            .       .    .  . ...Joe

*Joe Berkovitz*
President

*Noteflight LLC*
49R Day Street / Somerville, MA 02144 / USA
phone: +1 978 314 6271
www.noteflight.com
"Your music, everywhere"