On Fri, Jun 16, 2006 at 04:09:55PM -0400, Hans-Christoph Steiner wrote:
On Jun 16, 2006, at 7:08 AM, Tim Blechmann wrote:
and/or object-oriented languages, techniques, debugging tools, etc. Compare that to the amount of human-hours spent on visual programming languages. Its miniscule in comparison. So given that, I think that dataflow languages stand up quite well.
I have a feeling that dataflow languages will also be a really nice fit on massively parallel systems with thousands or even millions of processors. Imagine if every subpatch or data path in Puredata for example could be executed on a different CPU. I realise this is not technically possible due to Pd's design, but is interesting to think about. In any case, I think Pd would probably not be the best tool for general purpose dataflow programming since it's so geared towards making noise and art. Maybe something like Gridflow with a different scheduler?
Chris.
------------------- chris@mccormick.cx http://mccormick.cx