David Chiang

I don't know (but could take a couple of guesses) why Numpy matches from the right, but I definitely think it's good that Numpy chooses a single direction. As for...

Sorry, I mean a 0d array (or a scalar) with an array with 3 or more dimensions.

Thanks! The last case still has two different behaviors (skcuda likes it, numpy doesn't). I'll try a version and you can decide whether it's appropriate?

Sorry, you're right! I got confused by lines 885-888, which cover the case `len(a_gpu.shape) != 1`. These lines may cause surprising behavior in cases like these: ``` x23 = pycuda.gpuarray.to_gpu(numpy.random.uniform(0.,...

@lebedov, please see pull request #125. I think all three of the shortcomings mentioned there could be fixed, if this kind of generality is something you want.

This all sounds great. But we need to set it up so that it looks the same either way.

Also related if any pub chairs are still using it: https://github.com/yz-joey/ACLPUB/issues/7

XeLaTeX has a major disadvantage, which is that arXiv does not support it. So I don't think it can be made the default (yet). But I definitely agree with making...

Definitely agree with using the geometry package, which is very, very standard. It should also replace all uses of fullpage. I suspect not many people are using memoir together with...

It looks like @mjpost is the owner. Also, it appears that I can edit it -- I hope it's not world-editable -- is it?