Problems of representation in musical computing

Abstract

One of the advantages which projects in musical computation have over computing projects in other humanities disciplines is the previous existence of a rather good notational system for the description of the subject matter at hand. Music notation is fairly compact, is visually suggestive of the material it represents, and is easily read and written. More remarkably, the musician is able to perform music in "real time" from the musical score. Music notation as we have it today has been evolving for upwards of 3000 years and assumed a form prototypical of its current form some 350 years ago. It was then logical, when certain applications in musical computing arose, to attempt to devise an adequate encoding language which represented musical notation. A number of such languages have been designed, reviewed in books by Lincoln (1970), Brook (1970), and Heckmann (1967), and some of them are still used today. One such language is known as DARMS (Digital Alternate Representation of Musical Scores), which was devised by Professor Stefan Bauer-Mengelberg of the IBM Corporation, to form the input stage of a musical printing process. The history of its development is discussed by Erickson (1975), and its syntax, described in Erickson (1976) and in Erickson and Wolff (in press), will provide examples to follow. The virtues of music notation are predicated upon certain information-processing characteristics of the human perceptual and cognitive apparatus, which exceed those of state-of-the-art computer programs. It is these capacities which permit the high level of complexity found in musical notation. To adequately interpret muscial scores, we must make syntactic and semantic decisions in processing serial information, based on broad left and/or right contexts. We are required to resolve formally ambiguous cases, and to be conversant in a number of "dialects" of musical notation