Support latest THREE.js

[pehrlich]

Hi

It appears as this project is currently unmaintained:

• The bundled version of THREE is r53. The current latest is 65.
• Running the examples in the readme gives the following output in chrome: WebGL: INVALID_OPERATION: getUniformLocation: program not linked (Could this be my error? It looks like an out-of-date software issue.)
• Upgrading to the latest THREE (and tQuery, etc) causes further errors: Uncaught TypeError: Object [object Object] has no method 'addSelf'. Maybe this is the fix.
• (edit) There are 20 open issues (6 of them PRs), and 1 closed issue.

So two big questions before I go any further:

Does has anyone made an up-to-date version, or know how much work is likely to be involved? New to three.js here.

If I make an updated version and open a Pull, will it be merged?

Best, --Peter

[unconed]

MathBox continues to work for all my uses, so it is most likely a problem on your end.

Upgrading to the latest Three.js will not do much for this branch of MathBox. For one, you'd have to upgrade the tQuery too, and you'd have to rewrite all the vector/math code in MathBox as the three.js API has changed since then. With no tangible end-user benefit for MathBox I can think of.

I'm currently rewriting Mathbox pretty much from scratch with everything learned from using this branch, which is why this repo sits mostly untended.

[pehrlich]

great! I'll stay tuned.

I did get some amount of conversion work done yesterday, so I may publish if it proves functional later.

Thanks for a fantastic library~

[jasongrout]

@unconed: are you pushing your rewrite anywhere public? I'd love to see what you are doing, especially as I'm working on rewriting Sage's 3d graphics to use three.js and IPython widgets.

[unconed]

@jasongrout I don't have a fully functional set up yet, just parts, which I'll publish when ready. Some of that has started trickling out. For example, threestrap is my way of divorcing mathbox from tquery, and the geometry pipeline from my Dark Sunrise demo was a prototype for feeding data from live expressions into the GPU more efficiently. Also working on shadergraph 2, better attributes, and a proper sequencing animator.

I can't put dates on when this will be ready though.

[jasongrout]

@unconed: cool. I'll keep a lookout for any updates. Thanks for the cool stuff that you are doing! My multivariable calculus students were really appreciative of the rectangular/polar demos I tweaked from the Mathbox demos (http://artsci.drake.edu/grout/mathbox/examples/polar.html; http://artsci.drake.edu/grout/mathbox/examples/polar2.html)

[HLueckhoff]

@unconed: thanks so much for this great work. Unfortunately dependencies on Three r53 and tquery become more and more of a deal breaker for me as the mathbox bits I am using are just one part of a bigger Three application. It shouldn't surprise anyone that great software like Mathbox is used in ways not originally intended or foreseen by the authors. Here is my question: How is Mathbox2 doing? Any ETA? Or is there any chance to get access to an early alpha? For my current project I would require only pretty basic functionality around surface object. Best, Hermann

[wdyang]

Great work indeed @unconed! Same question here. Can we get access to early alpha? We are using it in experimental work. So we can deal with instability of alpha.

Thanks!

Weidong