David van Leeuwen

Hi, OK, well, I'll have a quick dig at it then, but maybe concentrate on the type-stability of some other functions as well, and then release this for 0.6. Further...

I don't seem to be able to able to convince Julia what the resulting type of `broadcast_c` is. I've spelled it all out in 41d7c45, but that is not enough....

(Truly, I didn't close this issue, Github did that for me.) Thanks for the detailed proposal. I can see why the type-compiler (is this called "inference"?) has problems deducing the...

Hello @nalimilan I don't get anywhere with type programming, see c299c71 . Even the simplest functionality seems to require a `typeof()` which appears to be a no-go for type inference,...

Small progress in 891fa55, but can't get `dictstyperecursive()` to be type stable.

Probably. I don't know what the Julia approach for printing is, really. I've tried to make normal showing in the REPL work, and very similar to dense Arrays. But I...

Definitively a reasonable request. I can foresee a potential problem when you give integer dimensions to the names: ```julia julia> n = NamedArray(rand(2,2), (["a", "b"], ["one", "two"]), (2, 1)) ```...

I think it might have been a while before the code got a release number, I think it is in 0.9.5 now

I'm not sure if it is related to #58, but it definitely is something I wasn't aware of an that needs some attention. The type stability issue still hasn't been...

It looks like the culprit is `similar(n::NamedArray)`, which has to build the index dicts. It doesn't know the key types in advance, since the dimensions of the similar array may...