Re: Multi-column distinctness.

Hello,

> FWIW Horiguchi-san is one of the few people who actually took time to
> review

I personally think such kind of things should not to be counted
in judging this issue:)

> the multivariate stats patch, and I don't quite see this patch
> as conflicting with the multivariate one.
> 
> It implements a small subset of the (much larger) multivariate stats
> patch, and reusing it within my patch should not be a big
> deal. Actually, this type of statistics was proposed by Horiguchi-san
> himself, and the multivariate patch does not implement it yet
> (although I intend to address that soon).

Year. I think we can deal Tomas's patch as natural enhancement of
this patch and it is not so bad as a vangard(?).

> So no conflict here. Of course, we need to be a bit careful to make it
> compatible (especially the syntax part).

I agree with it. They don't make bad conflict but we might stuc
by wrong decision in API, like syntax will.

regards,

-- 
Kyotaro Horiguchi
NTT Open Source Software Center