> I think it would be useful to start using this idea more widely to > improve the quality of DNS listing. So roll the idea around and see if > we can build on it.
It's somewhat similar to the "trust levels" we use in dnswl.org (where, incidentially, we partly import data from different sources that also use some kind of "scoring", eg from junkemailfilter.com). I believe this kind of aggregated scoring will prove pretty effective; whether the scoring is better done at each individual site (using SpamAssassin, for example ;) ) or centrally (at some DNSxL provider) remains to be seen. Personally, I would lean towards decentral solutions, but time will tell. -- Matthias
