Claire Foster

I believe I may have just encountered the same underlying problem. High level situation: I was trying to use `group` on a `FlexTable` containing missing data. A bit reduced (specific...

Yep, it seems to fall off a cliff at 8 elements. In my real problem I had a table with 17 columns with many missing values so julia slowly used...

Do you foresee any difficulties just swapping out the internal representation for a `Dict{Symbol,Any}`? I was having a quick look but I don't know much about the tables ecosystem yet.

I... I'm not sure what to say to about that It's clearly a semantic abuse, but is it worse than the `getindex` hack in base for typed comprehensions? ~~Presumably `t[(:age,...

> Presumably `t[(:age, :name)]` would make more sense No, just ignore that :-) My head is still in `DataFrames` land.

Beautiful summary, thanks. Those algebraic properties seem about right. In mathematics there's some precedent for expanding the set of things which operators act on into a union of several unrelated-seeming...

> just so long as what we implement here and in Rotations obeys some obvious rules Sure, agreed. From that point of view we can just use `similar_type`, document that...

Should we be trying to address this in some practical way for version 1.0? It's easy to revert #763 to make `one()` work "better". But things get subtle when trying...

> Can we use `@GC.preserve` on both `v` and `val` and use something like the isbitstype path? We basically want to copy a pointer... so long as the garbage collection...

> For the "fail hard" option, wouldn't it be better to disallow even creating `MArray`s with non-isbits elements? If so, I think defining `setindex!` here makes sense if it is...