Inventing XML for music

I got an interesting message in response to "Learn how to invent XML languages, then do so". Michael Good of Recordare LLC wrote:

I enjoyed your response to Tim Bray's piece on inventing XML languages.

I hope that when listing important XML vocabularies in the future, you will consider including the MusicXML language for music notation:

For 20 years people had tried to invent a better format than MIDI for exchanging music notation between applications. MIDI was not designed for this purpose, even though it was used this way. It could do the bare basics but not much more, and was really inadequate to the task. The two major attempts, NIFF and SMDL, failed in attracting any significant industry support (in SMDL's case, any industry support whatsoever). MusicXML is the first such language to succeed.

MusicXML is supported by over 50 applications, including the market leaders in music notation editors (Finale and Sibelius) and all the major players in music scanning. It has been adopted by commercial and open source projects; by industry developers, hobbyists, and academic researchers; by established products and innovative new applications. Consumers can finally exchange digital sheet music files between applications, and the barriers to entry for innovative new applications in the market has been dramatically lowered (e.g. see the entry of MuseBook, OrganMuse, Notion, and musicRAIN into the market).

I'm sure there are other examples where XML has successfully (not just potentially) broken down barriers between document interchange in specialized fields. But for now, MusicXML is the most dramatic success story I know. I do want to better understand how MusicXML measures up to XML vocabularies in other industries in terms of adoption rate. If you have pointers to other work in this area, anything you could send on would be most appreciated.

This just underscores my first reaction whenever I hear someone discouraging people from inventing XML formats—how can that be the product of anything but the narrowest world view? XML's strength is in providing a syntactic framework that can be used across innumerable domains. There is no reason why a musical XML format should not be as important as , say, Docbook. I don't know anything about XML formats in the music field, but I'm certainly happy to see that there is room in the XML universe for MusicXML as well as Country Dance (Folk Dancing) animation language, to grab another example plucked from

I was curious, so I browsed the landscape a bit for Music and XML. There were some interesting nuggets in the usual sea of noise in this SlashDot article on MusicXML, including this comment:

[Don't forget] the archival value of MusicXML -- [people] criticize it for "re-inventing the wheel," but they're only looking at the value for music composers and consumers.

The true value of MusicXML is as a universally understood format for describing musical scores digitally. The music libraries of the future aren't going to be made of paper, don't you think?

This speaks very sensibly to the overall value proposition of XML. The universal syntax allows data formats to evolve that enhance the longevity of stored data. Longevity that comes from transparency. (OK, so you have to have long-lived storage media as well, but that's a different topic). Such longevity is further enhanced by (once again) the closeness of the expression to the domain model.

There are other XML specifications in the area:

Actually, just go straight to the indefatigable Robin Cover on the topic.

[Uche Ogbuji]

via Copia