On 5/1/07, Andrew Lazarus <andrew@pillette.com> wrote:
> Let me just thank the list, especially for the references. (I found
> similar papers myself with Google: and to think I have a university
> library alumni card and barely need it any more!)
>
> I'll write again on the sorts of results I get.
Looking forward to hearing about them. I have worked with such dataset
problems, but never attempted to apply them to a relational database
such as PostgreSQL. If you want speed, nothing beats in-memory vectors
on a modern CPU architecture.
Alexander.