Fri Oct 14 17:43:36 2005
1) see examples and http://blogs.bl0rg.net/netzstaub/archives/000831.html for how to proceed to tie visual frames to lisp code (example is only dealing with avi as container and manipulating frames - we need data from frames in streams) - also suggestion of isis - macros?
but also that frame/container format is not so relevant (but does need to be resolved) as we can always use tools to convert formats - need address such base issues (avi, dv otherwise). automated conversion of files (dv whatever) uploaded say with kino to desired format - they reside in directory (or series of) which fm01 uses as working directory dissecting these to clips (again this can be abstracted out).
language commands, enter clips, scan new clips - suggested dirs for generated content
this should be what we have
2) at same time rewriting script? to resemble a probable language or the language and even experiment with Lisp code to re-generate/by way of a code stepper or parser? and if code stepper then we could bootstrap macros and function definitions - spiralling upwards bootstrapping affair to increasingly rewrite itself and functions towards linguistic fm01
2.5) experiments with examples and model and our own rough code/text generation system. then start to think of mobile interpreters
and then to tie to streams ????
also some kind of modularisation of system to fit with description or [where we are now]?
3) and in parallel with above two efforts commence recording of scenes (audio or overdub base question) and start working across all 3 aspects of project