On Fri, Sep 08, 2006 at 10:22:16PM +0200, Benjamin Mesing wrote: > sub.card(tag) / full.card(tag) * sqrt[full.card(tag) / > full.packageCount()] > Why do you divide by full.packageCount() inside the square root? As long > as the full collection stays the same for all tags (which I suppose it > does) this will not change the relation between packages. This is, > because the equation can be rewritten to: > [ sub.card(tag) / full.card(tag) * sqrt(full.card(tag)) ] / > sqrt(full.packageCount()) > and the additional factor would be meaningless for the '<' '=' '>' > relations (as long as it is positive anyways...).
You are absolutely right. And it allows me to simplify even further:
sub.card(tag) / full.card(tag) * sqrt(full.card(tag)) ==
== sub.card(tag) / sqrt(full.card(tag))
With such a sorter expression, it's faster to compute, and I might even
be able to properly formalise why it works one day :)
> Very good, I think a very useful feature is, to be able to enter a new
> search expression to get a new tag set. This allows for an incremental
> search using tags. I think I will steal some ideas there :-)
Go on, by all means :)
Ciao,
Enrico
--
GPG key: 1024D/797EBFAB 2000-12-05 Enrico Zini <[EMAIL PROTECTED]>
signature.asc
Description: Digital signature
_______________________________________________ Debtags-devel mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/debtags-devel
