Justus Sagemüller

Results 101 comments of Justus Sagemüller

After repeatedly rolling the thoughts in my head, I like none of the options (for achieving the tensor products of hmatrix-vectors to be implemented as hmatrix-matrices) enough to justify the...

Yet another option, before I settle on one: add ```haskell class (...) => TensorSpace v where ... type StaticDimension v :: Maybe Nat ``` This would be allowed to be...

Some heads-up on the (glacial-pace) progress with regards to this PR: - Added two classes, [`DimensionAware`](https://github.com/leftaroundabout/linearmap-family/blob/ac257d0b2f332172586af8d5693137f666f64512/Math/VectorSpace/DimensionAware.hs#L84) and [`Dimensional`](https://github.com/leftaroundabout/linearmap-family/blob/ac257d0b2f332172586af8d5693137f666f64512/Math/VectorSpace/DimensionAware.hs#L95). The latter expresses essentially just “space `v` has dimension `n`”, so those...

Absolutely. That seems to be a hack I wrote whilst in a hurry to get some paper compiling, it is not sane by any means. Your work on type-checked labels...

Thanks for the input. I'll have a look when I have more time.

Oh, I seem to to have been completely forgotten about this issue... Could you give some a bit more concrete / complete example of what you want this for? The...

Your impression is largely correct. In fact, my first idea before I implemented the `Agent` classes was to tap into `proc` syntax, but I couldn't figure out a good way...

(This would still require a type class that might get pretty hairy.)

...Actually your `Pair` pattern synonym already looks like it would do most of the job. Would it be enough to just have concrete `Pair` / `Triple` / ... synonyms and...

Really, there are two bigger issues underlying here: 1. Always carrying around the constraints for objects _and_ pairs of them is a pain. I'm not sure what `HasCon` / `OpCon`does...