[tei-council] Fwd: Some more on notatedMusic to the council

James Cummings James.Cummings at oucs.ox.ac.uk
Wed Jun 22 14:44:58 EDT 2011

More comments from Raffaele...

-------- Original Message --------
Subject: Some more on notatedMusic to the council
Date: Wed, 22 Jun 2011 19:17:56 +0100
From: Raffaele Viglianti <raffaele.viglianti at kcl.ac.uk>
To: James Cummings <james.cummings at oucs.ox.ac.uk>

Dear TEI council,

I think there is a misunderstanding between the proposed content model
for notatedMusic and the semantic weight that the music SIG wants to
impose on the element.

We proposed a fairly basic content model that makes use of <ptr>s and
<graphic> for external representations and <desc> for a human-readable
description only to minimize the creation of new elements.

However, what really matters to us is having an element that
specifically marks the presence of music notation in text. This is very
different, conceptually, from marking the presence of something that is
not text; if a generic element for grouping pointers to alternative
representation is a real necessity in TEI, then it should live *inside*
<notatedMusic> and not *instead* of <notatedMusic>.

For the same reason, I would still argue for membership to model.global.
As a figure can exist between textual divisions, so can music notation.
This is because notated music exists at a higher level of abstraction
than graphic or other binary objects.  Just as <graphic> points to an
instantiation of its <figure> parent, the <ptr> elements point to an
instantiation of their <notatedMusic> parent.

If the generic use of a mime-typed <ptr> is not welcome, then we need to
add *new* qualifying elements that have a similar collocation to
<graphic> or <formula>: i.e. <audio> or <video>.

To stress the difference in level between <graphic> and <notatedMusic>,
it may be helpful to remind ourselves that <graphic> only contains
descriptive elements and cannot, semantically, be extended to contain
textual data, only a <figure> should be extended in that sense.
<notatedMusic>, instead, uses <ptr>s as a way to defer the
representation of the textual data, but one may as well embed the
notation directly inside the element. And this is exactly what we plan
to do with a TEI-with-MEI customization (and this would work as well
with any other non-binary representation of music notation, i.e.
musicXML, lilypond, mup, humdrum, whatever).

To summarize:

<notatedMusic> : <figure>
<ptr>, <audio>?, <video>? : <graphic>

Is it necessary to extend -at this stage- our proposed model to include
the new elements <audio> and <video> so that they can also be used
elsewhere? Or is it sufficient to use mime-typed <ptr>s to keep things
simple? I am happy to follow both paths.


Raffaele Viglianti
PhD Candidate and PG Research Assistant
Department of Digital Humanities
King's College London

More information about the tei-council mailing list