[MEI-L] note/@fermata values

[Craig Sapp]

Hi Perry,

You were mentioning part of this to me last week, but I will
fill everyone's inbox with a feature request for note/@fermata.

Currently the note element's attribute called "fermata" can
be one of two values: "above" and "below".  It would be useful
to add a third option which indicates that a fermata is present
without specifying the placement.  Such as "auto" or "true".

In Humdrum data, the fermata's placement is not encoded.
So the translator to MEI is currently forced to make up a
placement (which is easy enough, and I will probably automatically
decide the placement even if you add this new attribute value),
but this is conceptually bad if you are following the
selective-encoding model.

The automatic placement rule I use: if the fermata is on a note in
the first (i.e., top) layer of a staff, it is placed above the staff.
If the note is in the second layer, the fermata would go below
the staff.  For higher layers, it would be difficult to predict, but
they would usually fall under my next attribute value request:

Examine this Bach chorale:
    http://kern.ccarh.org/cgi-bin/ksdata?file=chor010.krn&l=users/craig/classical/bach/371chorales&format=pdf

Notice that there is another type of fermata which is invisible:  the
alto and tenor voices do not display fermatas on any notes, yet
there are implicit fermatas for those voices.  It can be difficult for
a computer program to decide where these implicit fermatas
are located.  For example, look at the last bar of the example chorale.
The alto's implicit fermata is on beat 3 of the measure, linked to
the fermata under the bass note, rather than on the first beat linked
to the soprano note.

It would be useful to have an attribute value for fermatas which
conveys this information, such as "invis" or "implied".

-=+Craig