Re: [Pan-users] Scoring based on arbitrary headers?

[Jim Henderson]

On Tue, 06 Jan 2015 23:57:06 +0000, Duncan wrote:

> Jim Henderson posted on Tue, 06 Jan 2015 06:30:12 +0000 as excerpted:
> 
>> On Tue, 06 Jan 2015 04:10:09 +0000, Duncan wrote:
>> 
>>> So if you test this, please post your results. =:^)
>> 
>> Thanks for the detailed answer, Duncan - some of that I found with some
>> experimentation, but ultimately, yeah, it didn't work the way I'd hoped
>> it would.  I might have to try using an intermediate server (something
>> like the caching NNTP server that I can't remember the name of at the
>> moment) where I can set up the overview to use it.
> 
> Leafnode?

Yes, that's the one. :)  Brain overload last night after a long day at 
work. :)

> Meanwhile, just to confirm, arbitrary header scoring did work, but only
> after downloading the messages and possibly manually triggering a
> rescore,
> correct?

Hmmm, I didn't try a manual rescore, but the scoring that applied to a 
post that should have been affected didn't show up when I went to look at 
the rules applied.

> Did you have to manually trigger the rescore, or did pan automatically
> rescore after it downloaded the messages?
> 
> And what size cache are you using?  I predicted that at least if you had
> to manually rescore, you'd need to have a cache large enough to keep the
> set of messages you wanted to rescore, so pan would have the information
> at hand when you told it to rescore.  I normally work with a multi-GiB
> cache and for binaries at least, after deleting the obvious no-interest
> posts I download the remaining messages to cache for further processing,
> so that wouldn't be a big deal here, but for people who keep the default
> 10 MiB cache and routinely download and save in one go instead of
> working from cache, it'd be a big change.

I've got a 10 MB cache set, generally large enough to hold the messages 
I'm currently interested in in about 75 groups on various servers.

> Meanwhile (2), your idea of using a caching news server triggered
> another question I don't have an answer to.  Running a local caching
> news server and having pan access it is a setup most people won't have,
> but it does have some interesting implications, the most obvious of
> which is that pan could operate with little to no cache of its own (I'd
> probably run a small pan cache and put it in tmpfs, so as not to hit
> permanent storage at all in that case), since the entire server pan's
> accessing would be local.

Yep, I know a few people who use leafnode so they can use Pan, knode, or 
some other newsreader as an "offline" reader.

> But your comments triggered this thought process and question:
> 
> With control of the server as well, and thus its overview policy, you
> could of course put whatever headers your interested in, in the
> overview.  Can pan's arbitrary header scoring use such non-default-
> overview-header content if it's available in the overview, or must it
> still cache the post in ordered to score on those headers, because it
> doesn't expect them in the overview and thus simply doesn't check for
> them at that point?
> 
> Interesting question.  The answer really doesn't matter to most, but it
> would to those with servers that have a relatively liberal overview
> policy, putting many headers in it, as well as to those who do run their
> own server, either remote or local, and thus control what's in the
> overview.

That's kinda what I'm thinking as a fallback.

Jim
-- 
 Jim Henderson
 Please keep on-topic replies on the list so everyone benefits