pan-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Pan-users] Re: scoring rules question


From: Duncan
Subject: [Pan-users] Re: scoring rules question
Date: Mon, 28 Apr 2008 11:56:57 +0000 (UTC)
User-agent: Pan/0.132 (Waxed in Black)

Tom Daly <address@hidden> posted
address@hidden, excerpted below, on  Sun, 27 Apr 2008
11:51:48 -0300:

> 0.132 on Hardy
> 
> right-click > Ignore Author > New Scoring Rule > "and the Article's ::
> subject
> author
> <<references>>
> line count
> etc
> 
> 
> what does <<references>> mean?
> can someone explain?

[I went to sleep while composing this, waking up to find other replies 
already posted, but since this is more detailed and I had it almost done 
anyway, I might as well post it.]

Short but likely opaque answer, it's referring to the references header.

More practical but much longer answer...

The references header, which like all headers you can see by hitting the 
view-headers toggle (the "h" key by default, IIRC), tracks post lineage, 
parent, grandparent... original thread-starter post.  This is how proper 
(IOW, NOT just by subject) threading is managed.  New replies will append 
their parent's Message-ID to the end of the existing list (if any) of 
Message-IDs in the references header, creating the header if it doesn't 
yet exist (generally meaning it's a reply directly to a thread-starter 
original post).

Now, unfortunately pan can't score on just any header, only the ones 
normally available in the overview, the short "overview" of information 
sent from the server when updating "headers" (which are thus more 
correctly called overviews as shown in the pre-message-download list, 
because it's only a limited subset of headers).  The headers normally 
available in the overview include those necessary to populate a normal 
overview/header listing, thus Subject, Author, Date/time, bytes and 
lines.  They also include the those necessary for the client to track the 
post, thus Message-ID (globally unique message identifier), Xref (for per-
group message sequence number and cross-posting), and references (thus 
allowing the client to properly thread the posts).  That's why pan can 
score on most of those headers, but not stuff like Organization, NNTP-
Posting-Host, etc, which don't normally appear in the overviews and are 
therefore not normally available until after the message is already at 
least partially downloaded.

So scoring on the references header is what watch/ignore thread would 
do.  Applying the score on the references header, using the message-id of 
the original post should then apply it to the entire thread, while using 
a reply's message-id will apply it to the subthread below that specific 
reply.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman





reply via email to

[Prev in Thread] Current Thread [Next in Thread]