Just a Crazy Idea

[Rangarajan]

For software development, we have requirement gathering tools, UML-like modelling tools, Design Patterns, IDEs for programming, and even software Testing tools. When it comes to music composition, if you are manually composing music, there are tools such as "Finale" for notating, or you can fall back on the ubiquitous "paper" for expressing your composition. On the other hand, if you use algorithmic composition, we have many tools such as Opusmodus, Open Music, SuperCollider, Max/MSP, and so on. I think what is missing, and what I feel is important, is a way to represent intermediate stages of composition, which is really the creative part. Just as in software development we say the design is more important than the final code, we should be able to independently study, understand and critique the compositional design of a musician. In fact, it might even be possible to automatically generate executable code from this intermediate design representation (just as in simple cases, UML design can be converted to code in any language). So what I am really proposing is to come up with a "Music Design Notation" for algorithmic composition. When I listen to some of Stephane's pieces, they sound good, but when I look at his code, his "design" insight is not obvious. This is true for any non-trivial program. On the other hand, if there was a MDN representation of his work, it might be easier to understand what is going on. Whether your final piece is rendered on paper, or Finale, or in Opusmodus, there is a great benefit to documenting the "higher" level design of the composition. Then we can even talk about "Design Patterns" in musical composition in a programming language-neutral way.

 

For those of you who do not come from software background, I encourage you to read the works of Christopher Alexander, especially, "The Timeless Way of Building" (Oxford University Press, 1979).

 

I have had this on my mind for a long time, so thought I will post it to see what others think.

 

Regards,

Rangarajan

[opmo]

Dear Rangarajan,

at the beginning of Opusmodus development we have played with the idea to use some kind of UML design. I wonder how would you imagine such a design and how the conversion would be represented. Have you worked in the past on similar project. I would be happy to elaborate more on this.

 

I just both the "The Timeless Way of Building", thank you for the info.

[Rangarajan]

At the moment, I am blank on this. My limitation is that I do not have a music background, but I am reading a lot these days, so I am confident something will come up soon (hopefully). It would certainly help if people like you (both s/w and music expertise) and other musicians in our forum start to "introspect" on the creative process each goes through when composing, and then jot down notes, ideas, or whatever that comes to mind at that time. To me these are valuable artefacts that could be formalised eventually.

 

When I get some inspiration, I will share in this forum.

 

-Rangarajan

[Stephane Boussuge]

Hi,

 

when i compose with opmo, i use not so many different strategies , it is kind of patterns i use and reuse very often.

For example, i like to use a vector for describe and drive some aspects of the piece like rhythmical density or instrumental density, harmonic tension/relaxation....

i have some strategies for every aspect like rhythm, harmony, density, dynamics etc...i combine some of them in every composition  and i 'm trying to refine them and invent some news from composition to composition.

 

I don't know nothing about UML but if it can describe the thinking of a composer more accuratly, it is pedagogically very interesting !

 

SB.