Re: MusicXML and GuileV2

[David Kastrup]

Han-Wen Nienhuys <address@hidden> writes:

> On Thu, Aug 30, 2012 at 2:04 PM, David Kastrup <address@hidden> wrote:
>
>> This would also seem to form an excellent starting point for having
>> LilyPond interpret MusicXML directly, bypassing converters as well as
>> the ly parser.  While this would not offer the full tweaking power and
>> expressivity of ly, it would certainly be a good way of interfacing in a
>
> While sxml is certainly nice, I wouldn't overstate its usefulness.

That was what "starting point" was about.

> The big problem with musicxml is that it has a data model that is
> completely different from anything inside lilypond, something that
> sxml does not help a lot with.

But sxml is susceptible to things like Scheme syntax transforms and the
(ice-9 match) module.  The Scheme layer is quite powerful for juggling
XML if it is in suitable form, and sxml seems designed for that.  I'd
like to have the same kind of power for music expressions, but that
requires quite a bit of schemeification before it will be effective.

At any rate, not having to go through the ly parser for MusicXML
interpretation would be seriously useful.  It would make it possible to
use LilyPond in server/rendering tasks with security or resource
implications, similar to PDF as a page description language as opposed
to PostScript as a general programming language with graphical output.

-- 
David Kastrup