Re: [Pan-users] Scoring on Message-ID & arbitrary headers

[Duncan]

On Fri 21 Feb 2003 17:31, Charles Kerr posted as excerpted below:
> I will eventually get to scoring/filtering on arbitrary headers,
> it's not a high priority.  This is a lot more expensive in newsreaders
> than in mailers, because instead of having the entire message like
> mailers do, Pan only has what XOVER gives us -- number, message-id,
> references, author, subject, lines, byte count, and usually xrefs.
> Anything else is going to cost you a round trip to the server, or
> a painfully slow search through cached articles.
>
> That's not to say that it can't be done, or that it wouldn't be
> useful.  But, just so that you know, there *is* a rationale for
> only having limited criteria for now.

I understand that.  It's basically the difference between filters/scoring that 
can be applied while the body is still on the server, and ones where the 
entire message must be d/led in ordered to apply the filter/score.

However, it would still be a cool feature, as even if a message has to be 
downloaded to apply it, that's still far better than me having to apply a 
mental filter manually and act accordingly.  IOW, if the first thing I do 
when I see a message is delete it, it's already downloaded anyway, so having 
to do that to apply the filter/score wouldn't alter the net traffic situation 
at all.  However, it'd still be a huge bonus to never have to see that post 
in the first place, with the deletion happening automatically according to 
the filtering/scoring.

In the best case scenario, those filtering criteria that would require that 
the message be d/led would be marked.  Perhaps even a deliberate two-pass 
system could be used, and set up in the GUI to make it plain.  In the top 
part of the GUI would be all the pre-d/l scoring, perhaps in some 
prioritizing order.  In the bottom would be all the post-d/l scoring, with 
expanded criteria available, and a note to the effect that these would 
require downloading the entire message to apply.  Pre-d/l, only the top 
section would be applied, and if the message hit the delete threshhold there, 
it wouldn't even be d/led.  Post d/l would apply to what the user had to 
process, but obviously wouldn't apply to d/l traffic, only end user display.

-- 
Duncan
"They that can give up essential liberty to obtain a little
temporary safety, deserve neither liberty nor safety." --
Benjamin Franklin