Re: [eigen] on fuzzy comparisons

[Benoît Jacob <jacob@xxxxxxxxxxxxxxx>]

On Thursday 05 June 2008 22:03:03 Benoît Jacob wrote:
> I showed that norm, not L2norm, was the good notion. So in all this
> business with K and L, we are just trying to approximate the matrixNorm
> with L2norm, but matrixNorm remains what we really want. Of L/K becomes
> arbitrarily big, our approximation becomes arbitrarily imprecise.

D'oh, L2norm is the same as norm for vectors, so disregard this paragraph 
please :)

Benoit

Attachment: signature.asc
Description: This is a digitally signed message part.