its trying to communicate to an (optionally started) osc-bridge-script. (you can download this at shaderoo.org (see help section) and start it manually on your localhost to be able to communicate to an OSC-device to adjust parameters - e.g for VJ'ing)
oh, thanks for the comprising debug info...
yes, the first error might not be crucial...
i think i tracked down the reason, its due to some new code-inline-image feature i'm currently testing. some lines too long in the shader code - firefox doesn't seem to like that... i'm on it...
and the socket stuff is for remote osc-connections (you'd have to start a small server-script locally to use that - see shaderoo help-page)
well, actually its a self advection based on rotations only (and therefore automatically divergence free). so there's actually no explicit navier stokes equations or pressure/tension physics involved (implicitly one could argue though...).