On Mon, Jan 12, 2004 at 12:16:52PM -0800, Hallam-Baker, Phillip wrote:
| This is true, you can perform all the compensation factors yourself based on
| your interpretation of the other person's data. Or you can put the
| responsibility on the provider to make all the compensation factors based on
| their first hand knowledge of their data and the way that they collected it.
|
| This is all about complexity and where it gets managed. The design you
| propose looks very much like the arrangement in traditional X.509 PKI where
| the responsibility for managing complex trust relationships was pushed to
| the application client. The result was horrendous complexity and the clients
| kept having to upgrade to keep up with the latest version of X.509.
|
| In this case there is no real variation in the output variable that people
| are after. Push the complexity to the data provider.
|
Sure, I expect that will be the common case.
But even if you just give a "credit rating" style number, which is
really based on a confidence interval (95% probability this is spam), it
would still be nice to preserve the ability to pass the original data to
the querant. There are quite a few dimensions (age of domain,
spamtraps, human reported complaints, etc) and in practice they may not
reduce elegantly to a single number.
In any case this is all minor detail.
Hmm ... they might use XML or TXT or YAML to encode the data.
Let's put XML into DNS! Bring on the new RR types! :)
-------
Sender Permitted From: http://spf.pobox.com/
Archives at http://archives.listbox.com/spf-discuss/current/
Latest draft at http://spf.pobox.com/draft-mengwong-spf-02.9.4.txt
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to
http://v2.listbox.com/member/?listname(_at_)©#«Mo\¯HÝÜîU;±¤Ö¤Íµø?¡