Synthesis of general programming language (Python) with tailored language (PureData/MaxMSP/ChucK)

Question

I am learning Python because it appeals to me as a mathematician but also has many useful libraries for scientific computing, image processing, web apps, etc etc.

It is frustrating to me that for certain of my interests (eletronic music or installation art) there are very specific programming languages which seem better suited to these purposes, such as Max/MSP, PureData, and ChucK -- all quite fascinating.

My question is, how should one approach these different languages? Should I simply learn Python and manage the others by using plugins and Python interpreters in them? Are there good tools for integrating the languages, or is the proper way simply to learn all of them?

Answer 1

I would say learn them all. While it's true that many languages can do many things, specialised languages are usually more expressive and easier to use for a particular task. Case-in-point is while most languages allow shell interaction and process control very few are as well suited to the task as bash scripts.

Plugins and libraries can bridge the gap between general and specialised languages but in my experience this is not always without drawbacks - be they speed, stability or complexity. It isn't uncommon to have to compile additional libraries or apply patches or use untrusted and poorly supported modules. It also isn't uncommon that the resulting interface is still harder to use than the original language.

I know about 15 languages well and a few of those very well. I do not use my prefered languages when another is more suitable.

Answer 2

It's perfectly possible to build good interfaces from Python to such specialized languages: one example in point is RPy, which lets you drive R (for statistics) from Python (for all sort of general-purpose stuff).

Of course, one has to be competent in both languages - and such bridges, unfortunately, will not already exist for every given pair of one general purpose language and one specialized one. "Learning all of them", if you want to use all of them, remains the royal road!

Answer 3

Python would be a great language to learn, since it works well with a lot of other languages. It makes a great general purpose language as well as a "glue" language. Spend time learning the languages you are interested in, and keep Python knowledge around for it's flexibility and power. I don't think I would recommend trying to learn them all unless you really have the time.

You may interested to know that PureData has a python extension.

Answer 4

This thread is a little old, but I wanted to point out that the majority of the mature audio development environments e.g. supercollider/max-msp/pure data can be controlled via open sound control. You can google up a better description of OSC, but suffice it to say that it allows you to send control data to synths built in these environments similar to how MIDI works, but way more extensive. This does not solve the problem of actually building synths in python per se but it allows you to "drive" these other environments without having to know the ins and outs of the language.