datavis

meow 2015-12-22T14:44:48.000761Z

@kephale: Have you looked at vectorz for any of your simulation work? Check out this implementation of Conway's Game of Life: https://github.com/mikera/core.matrix/blob/0cf43fe3b3a1992aec2fbe041ae0ab5ed34d02b5/src/test/clojure/clojure/core/matrix/demo/life.clj

2015-12-22T18:48:29.000763Z

@meow: yeah VCG looks pretty solid, the thought with meshlab (stemming from meshlabjs) was that, i believe, VCG has been compiled to JS

2015-12-22T18:48:41.000764Z

so it might be usable via clojurescript interop

2015-12-22T18:49:01.000765Z

yeah I’ve looked at vectors, and mikera is in this channel too

2015-12-22T18:49:51.000766Z

i haven’t tried it out in practice since i got a bit committed to my codebase before core.matrix came out, but i’m likely to start porting soon, especially if core.matrix gets the JS support that it has listed on the todo list

2015-12-22T18:50:48.000767Z

FWIW, i swapped over to ojAlgo as the matrix library because EJML (which vectorz is based on) doesnt has some tradeoffs relative to the matrix size you’re dealing with

meow 2015-12-22T20:55:29.000768Z

@kephale: I'm going to play around with core.matrix and vectorz just to get my feet wet. Assuming I write all my mesh ops against core.matrix protocols should leave the door open for alternative matrix implementations.

meow 2015-12-22T20:56:37.000769Z

The underlying matrix stuff I really don't want to have to bother with - just want to work with a higher level api.

2015-12-22T22:37:07.000770Z

@meow: yeah, that sounds like the way to go. writing against matrix implementations is probably going to be way better for speed, the only thing that drives me nuts in those cases is how (generally with just cause) 3D renderers use specific data structures for 3D vectors, so there is a conversion overhead there, but for static renders it might not even be noticeable