Re: Notations in Scope

Thanks, James -- I think this is a very useful dialogue; I'll try to
respond below.


> *"A Performable Notation is one in which one or more synchronous events
> can be mapped to one or more graphic objects in the score."*
>
> Note that nothing is said about the alignment of the graphic objects. In
> CWMN, the symbols are aligned vertically of course, but that's a special
> case.
>
> Also, nothing is said about the relative complexities of the event(s) and
> the graphic(s).
>
> Suggestions for improvements to this definition would be much appreciated.
>

I agree that events are more useful to talk about than instants, but I
think that just as the term "graphic objects" is very open-ended, so is
"synchronous events". And while SVG is a nice existing standard covering
graphic objects, it is far more open ended than MIDI. MIDI is in my view
not much better than CWMN (and is arguably worse) in the range of musical
events that it can represent.


> I think you may be confusing the (humanly readable, perceptible)
> instantiations of the score (on screen or paper, and in performed time)
> with the encoding of it in a computer file. I see no point in having an
> encoding for a music notation that cannot be instantiated.
>

I agree that music notation needs to be able to be instantiated spatially
and aurally to be meaningful as "music notation". But that does not mean
that an encoding should be forced to supply that instantiation in all of
its rigor. An encoding should be able to supply enough information to
generate such instantiations, and such encodings are very useful.

Part of the "special case" of CWMN (and part of what makes it so
significant in our culture) is that it is possible to create many possible
instantiations of the same notational data, to suit the needs of the end
user. There are many stories in our collections that testify to this
requirement.

In Frankfurt, Werner and I were talking about text and letters of the
alphabet as an analogy. We do not need to know how a "T" is shaped as a
graphic in some particular font, in order to know that it is a "T" and to
be able to instantiate a rendering of it (perhaps in various fonts
depending on the needs of the reader -- I need a pretty large font these
days!). In fact, the encoding representing this email does not include a
graphic of the letter "T" -- just a numerical Unicode code point -- yet you
are looking at a graphic of it right now.

I'm not the best metaphysics expert in the CG, I'm sure... but let me
explain what I mean by "semantic". Perhaps we'll find a better word in the
process. I hope we don't need to resurrect Wittgenstein and get him to join
the CG.

When I say "semantic musical data", I mean musical data that is sufficient
to support a *space of possible instantiations*. It is abstract musical
data, in the sense that the abstract CWMN construct "eighth note with pitch
C4" refers to a class of many possible visual realizations... different
sizes, fonts, stem directions, etc. Same with aural realizations. Yet all
these visual and aural instances are constrained by the nature of CWMN to
things recognizable by readers (or listeners) as "eighth note C4".

When a notational system is codified and stable -- as with CWMN -- that
space of possible instantiations is large and flexible. This flexibility is
desirable, especially for publishers and musicians who want to work with
music on many devices in many formats. Thus, it's simple and it's useful to
have an encoding that supplies this stripped-down, more abstract data.

The same is true of neumes: there is a limited vocabulary of glyphs, and it
is possible to step back from the "graphics" and encode these as sequences
of abstract symbols which could conceivably be rendered using different
fonts, different spacing, and so on. Monks didn't reproduce the exact
geometry of their source documents when making a copy!

When a notational system is more unique -- since I greatly admire
Stockhausen, let's say, like the notation shown here:
http://www.sonoloco.com/rev/stockhausen/91.html -- isn't it pretty much
limited to the instantiation provided by the composer and their
collaborators? (Yes, the composer can supply a set of rules on how that
instantiation would be transformed for use on different media and display
sizes... but that problem is very far from being solved for pure graphic
standards like SVG.) In these cases, an encoding is going to have to be
much more visually literal, and the types of musical events will be much
more complex and ramified. A standards group cannot know what concepts to
abstract, in order to allow some "stripped down" representation.

Now... ideally I would like a single standard that could embrace both these
worlds. I don't know that this is possible yet, but these are some of my
reasons for thinking that SVG + MIDI is  not a good starting point: the
lack of abstraction in such a system will be an enormous problem when
working with CWMN.

We may need to look at ways to allow alternative approaches to encoding --
one approach that is based on existing notational schemas and abstractions,
and another that begins with literal (but completely general) graphical
expression. I favor pursuing these ideas in parallel, rather than trying to
adopt one over the other. Ways of bridging these approaches may then become
more apparent. Naturally, I look to you to share more of your thinking
about a pure-graphical approach.

...Joe