Data models for virtual distribution of musical scores

With the rise of portal sites for image-file download of musical scores, the electronic distribution of musical scores, an activity in progress of 15 years, has attracted increased attention. We describe several data models for score distribution and emphasize the one used in the MuseData archive of musical data (1984-). This archive (containing substantial quantities of the music of Bach, Handel, Haydn, Mozart, Beethoven, Corelli, and Vivaldi) employs a "solar" model of supersets of musical data. From these files, score printing, sound generation, and logical information for musical analysis can be generated. The design, efficiencies, and shortcomings of this \´write-once, convert-many" system are compared with models emphasizing generic musical-data interchange, and models in which music must be re-encoded for each new application. Finally, these considerations are examined in light of social practices and economic models traditionally associated with the print publication of music in Europe and the U.S.