Gustavo Goretkin

I only noticed you asked me to open it on `SymbolicUtils.jl` when I went to follow-up on the the discourse thread. Go ahead and close here / open there /...

I think even if simplification rules are not applied automatically/immediately (which seems like a good idea!), I think there's still a question of what good does using the type `Number`...

> I think that is not very un-Julian, except in this case it would be annoying to end up with a 0 for Real but not for AbstractFloat... I would...

I was going to write this out as a comment, but then thought this information might be useful to anyone else stumbling upon that repository, so I added something to...

> we should probably add a specialized method for Statistics.mean. > The alternative is to consider converting the various interpolation libraries to express their internal computation via affine rather than...

I agree with making the distinction systematically, but I don't think that addresses how to define `mean` on affine quantities. Related to making this systematic distinction is "torsos", which I...

I want to bring attention to my rough and partial attempt for Unitful linear algebra: https://github.com/goretkin/UnitfulLinearAlgebra.jl

related: https://github.com/PainterQubits/Unitful.jl/issues/38

Hm, I see your point, but that's also, I'd say, the point of collapsing. `OffsetArray`s nest in a collapsing way. `SubArray`s nest in a collapsing way (in many cases, though...

Making the tests too specific will cause them to fail when they maybe should not on changes like https://github.com/JuliaArrays/OffsetArrays.jl/pull/119 On the other hand, having specific tests are a way to...