Andreas Borgen Longva

I'm currently more of the following opinion, but I can quite possibly be convinced: If users really know what they are doing and want direct LAPACK access, they would probably...

Anyway, I wanted to start on reorganizing all the decompositions (i.e. in terms of splitting into multiple files, no change in API or functionality), but I was unsure whether to...

@vks: You are right, there's a lot of stuff in the pipeline which is not yet on crates.io. In particular, on master there's also `FullPivLu`, `Cholesky`, `HouseholderQr` and in a...

Thanks for this PR! From a brief glance it looks good. I agree about the example. Do you have anything more suitable in mind?

(Just mentioning #172 here so that we get a reverse link)

`nalgebra` heavily relies on inferring storage from dimensions. That is, given a *shape* `(R, C)` where `R: Dim`, `C: Dim`, `nalgebra` infers through the impls on `DefaultAllocator` what the associated...

I think both your suggestions for `nullspace` (or `null_space`? Bikeshedding time?) and `condition_number` are good, and I agree with your assessment that a relative criterion is more useful. If someone...

@geoeo: did you mean to close this? It looks like you've discovered an important deficiency in `nalgebra-lapack` here. I'm otherwise unfortunately not familiar with the codebase of `nalgebra-lapack` and I'm...

I've reopened this because the linked PR (#1106) does not directly address this issue.

Thanks for reporting this @acshi. Can you confirm your version of `nalgebra`, please? I also rely on eigen decompositions, so I also have a personal interest in making sure they...