RE: For your consideration: an MNX thought exercise

Dear Alex, 

 

Nice examples you give, and they do prove a point. But it’s a different one
than you intend to make, I’m afraid. And it is at the core “philosophical”
problem we’re having in this discussion. The discussion about having a
measure object or not is not the core. The core is about the scope of a
standard being written.

 

In my view, your examples actually prove that such examples should not be
taken into account when writing/discussing a standard. 

 

The Pömmelchen example should be clear by itself. It’s a drawing. It’s a
nice find, but the graphical meaning is more important than the musical one.
It’s actually very hard to read, and leaves room open for
interpretation/error. What happens when the right hand “track” joins the
left hand? Do we continue in treble clef, or start playing bass-clef with
the right hand? The end also seems ambiguous to me. So as it is written,
without explanation, it’s not playable as it has no *standard*
interpretation even at this moment. So let’s not take it into account for a
new standard. 

 

The “medium-level” example even more proves the point that this does not
belong in a standards discussion. The piece does not follow normal music
conventions in notation as we have them now. It really is two pieces that
have no relation (and that could/should be written completely separately)
that are forced onto each other. This is made 100% clear by the instructions
in footnotes that show how to interpret things written down. 

Now, I would consider standard notation to already be more flexible than
what realistically can be covered in a computer standard/program which
typically is narrower than “real-life”. So if it doesn’t already fit a more
flexible “standard”,  there’s no use trying to fit it into a standard that
is more ridged in nature. Square peg and round hole comes to mind here. 

 

But the problem is really more fundamental. One could say that the actual
intent of the composers of these two examples was to not follow standard
conventions. They are people that like to experiment, and color outside of
the existing lines. Stockhausen and Satie were mentioned, and those also fit
this description. 

 

You will always have this tension between the norm(al) and border. It cannot
be avoided. You can make a standard as wide and flexible as you want, you
will find people that will not follow it. So the goal as described here by
some to have a standard that has the potential to cover “everything” is a
mirage.

 

I’m all in favor of having a quite narrow scope. It gives focus. It’s easier
to implement, and therefore has a larger chance of actually being used by
enough programs so it achieves its goal of fostering interoperability. 

A too wide scope typically will either fail to define all elements with
sufficient detail to avoid ambiguity, or it will be so large that no one
will actually implement it. 

 

Now, Joe is trying to have his cake and eat it too by defining a wide scope
in general, but using profiles to narrow down the standard to a workable
subset. I’m fine with that. As long as we define the CWMN profile ASAP, and
make it into something useful that is usable by a lot of programs. 

 

CWMN probably covers >95% of the market at this moment (*). So it really is
a very sensible thing to do to limit us to that as a starting point.  

 

(*) It’s probably more like >99%. Let’s walk around in Hall 8 (publishers)
at the Messe, and do a sample of what’s on display. For those that say we
should also do that in Asia to get another perspective, well, I’ve done it
at the Music Fair in Shanghai latest November. Lots of really different
instruments, but all sheet music on display was CWMN. 

 

Regards, 

 

JanR

                

 

 

 

 

 

 

From: notenlektorat [mailto:post@notenlektorat.de] 
Sent: dinsdag 4 april 2017 0:44
To: public-music-notation-contrib@w3.org
Subject: For your consideration: an MNX thought exercise

 

Following the discussion over the last days I kept a steadily growing text
file with notes and questions. But with each new entry the prospect of
discussing all those points in detail over a mailing list got me more and
more frustrated, since this mode of communication is really not my forte.

 

So, I tried a different approach. Based on my understanding of Joe's
proposal I set out to encode as thoroughly as possible this medium-complex
example <http://notenlektorat.de/examp/MR7.jpg>  (the first page of the
seventh movement from Ligeti's Musica Ricercata), building on many
principles that Joe outlined, but at times also going off into a very
different direction – here is what I came up with
<http://notenlektorat.de/examp/An%20MNX%20thought%20exercise.pdf> . I should
say at this point that I do not present this here as a deliberate proposal,
but rather as a jumping off point for further discussion. The whole document
is still more on the pseudocode side of things, and there is of course the
fact that I am merely an autodidact coder, with a good chance that I have
cooked up something that does not make sense.

 

I hope that the document is self-explanatory in many ways. Since Frankfurt
is coming closer and closer, I decided to just get this out instead of
producing a version with some more in-depth commentary (something that I
will gladly provide in case it is requested, outweighing any
well-intentioned ridicule). Still, here are some select points of interest:

 

It will become obvious quickly that I am of the faction that would like to
see the measure to lose the fundamental structural importance that it is
endowed with in MusicXML. However, I would also like to stress that the
structure that I have come up with should be perfectly capable to have any
encoding of the sort presented here rearranged in an equivalent form that
comes close to the measure-centric approach – at least so I believe.
Furthermore, it should be possible to provide a generic algorithm for
converting one form into the other and vice versa (Full disclosure: this is
not much more than a conjecture at this point). It is my conviction that the
presented approach has distinct advantages, but it is in fact also my hope
that attempting to reach a high level of abstraction will actually allow for
safely encoding the same thing by way of different structural
encapsulations. The example at hand is written in a way that makes heavy use
of referencing; with mostly the same tools (other than those for
referencing) it should be possible to write the same content out concretely,
although I suspect that it would be much more complicated. Still, such an
equivalent form that emulates a measure-based structure might be preferable
for some applications.

 

One might also notice that barlines do not feature at all within the whole
thing. This is because barlines are considered implicitly provided elements
of the containers with type "measure", at least within the context of the
notationtype "CWSN". Implicit sub-elements are in fact central to the whole
system. Looking at my proposal for tied notes one might wonder why no actual
elements for ties do appear. However, those are assumed to be implicitly
included in the notated tie chain elements, and a parser would, presumably,
conjure them up from default settings when processing the file. Only if an
implied element does deviate from its default settings will it be explicitly
encoded. There are some such cases within the example, although not with
ties.

 

Lastly – the document uses an extensive colour scheme. I hope to expound the
principles behind this at a later point and to lobby for their inclusion,
but for now you might safely ignore it.

 

I will try to post a few more examples of smaller scale before Frankfurt. I
particularly hope to present next some twenty-seven bars from a Beethoven
string quartet (as an example of a much more conventional score), the
encoding of which focuses on a few aspects that are not prominent in the
Ligeti. I am aware though that today's example alone is quite a whopper to
be studied in some depth within the remaining time. So I will expect no one
to hold their breath for me posting even more. Still, I would be glad if I
have provided some new input to the overall discussion. The only thing that
saddens me is that there won't be enough time before the meeting any more to
attempt a complete manual pseudo-encoding of this
<http://notenlektorat.de/examp/P%C3%B6mmelchen.png> pièce de résistance.

 

Best,

 

Alex