On Mon, May 10, 2004 at 09:59:51PM -0700, Daniel Quinlan wrote:
> > [re: net-daf] bizarre...  all spam?
> 
> Duncan had something weird going on with blank messages or something
> like that.  Maybe he should disable nightly runs until it's fixed.

I've disabled them; I'm going to need to look into this (I'm at work
right now...)
 
> I suspect some of his failures are for the same reason as mine, see
> below.
> 
> > [re: net-jm] fails are good, passes suck.
> 
> As expected.
>  
> > [re: net-quinlan] not terrific.
> 
> Most of my failures are historical.  SPF records unfortunately specify
> "now", not "then".

I suspect mine aren't as likely to be historic -- I'm running net
tests on only 2 weeks of mail (IIRC).

> > since SPF isn't a spam indicator, just a forgery indicator, as long as
> > the hits are valid, I'm ok with the results?  can people double check
> > so we can either fix things or close the ticket?
> 
> I'm not sure what we should do about failures.  This might be a case
> where we have to avoid using the GA and go with human set scores,
> initially very low until SPF is more generally reliable.

Or just "hint" perceptron accordingly...

-- 
Duncan Findlay

Attachment: signature.asc
Description: Digital signature

Reply via email to