0x09

My friend and ex-adviser Jim Mahoney sent me this response to my previous article on dance notation. One of these days I will post some articles in response. For now, I'll just share this, since it's neat:

Date: Fri, 30 Nov 2007 15:32:49 -0500
From: Jim Mahoney
To: ian
Subject: Re: notation

Hi Ian.



Several comments :

First,

 I think an analogy to TeX would be helpful.  Who before
 Knuth would have argued that we needed a Turing-complete language
 for document typesetting?

Second, after

 "1. Describe any movement of any object in any time format."
 you say "This is pretty self-explanatory".  I disagree.  There's
 a *lot* of explaining needed here: what form is this description
 going to take, and what sorts of objects are possible?  Are easy?

 I want a general "movement notation" that could describe
 dragons fighting in mid-air, an eight-person contra
 dance, or a wink.  That's an awfully big range.

 Moreover, if it describes walking around in my house, should
 in contain a description of the house?  Allow a description
 of the house?  Should/could it contain any meta-information
 about how the motion should be displayed (camera positions,
 lighting) or the figures rendered (textures etc)?

 What spacial resolution should be possible/easy?

 If all of this is possible then one of the big challanges
 is going to be make it easy.

Third,

 I think we need some possible examples of what this
 language might look like.  Perhaps some examples
 along the lines of existing languages (python,
 supercollider, etc)

Fourth,

 if the paper is meant to define the task, then it needs
 to be clear that part of the task is to create the tools
 that will enable this language to create things like
 videos of the dancers.  Ideally (though a harder problem)
 are also tools to make it easier to input the dance motions.

To accomplish all this, I think this language needs to be
tied to to two other technologies :

   (a) a low-level "character" and "animation" description,
       probably x3D, and

   (b) a symantic-web-like naming scheme

so that the language can say things along the lines of

   define dancer "man"        http://..../person.x3d
   define motion "walk man"   http://..../simple_walk.bvh

and then define more procedures using these primitives.

To continue my TeX analogy, without dvi2ps (compilation)
and global fonts (low-level globally named primitives),
the language itself isn't interesting.

I think one of the real challenges in a system like this
will be to figure out how to manage and implement parameter
passing to these low level animation descriptions
(i.e. "quickly" or "left") and adapting them so that
they're not completely tied to one specific movement
for one specific model, as they typically are now.

All for now,

Jim M