Re: [Lurker-users] Threading question
Brought to you by:
terpstra
From: Wesley W. T. <we...@te...> - 2003-07-14 02:33:30
|
On Sun, Jul 13, 2003 at 07:09:54PM -0400, Kevin Brosius wrote: > I notice that threading usually happens, even when a message only > matches in subject (meaning someone copied the subject line, rather than > replying in a mail reader.) I see one occurrence that confuses me > though. Take a look at http://kevb.net:3000/lurker/list/plug.html and > the pair of entries for "Re: PLUG Picnic! (Really Re: [PLUG] The Next > Step? -- was Community Service Project)" >=20 > What's interesting, is that it looks like the person copying the subject > left out a space, ie "Re: PLUG Picnic! (Really Re: [PLUG] The Next Step? > -- was CommunityService Project)" causing threading to break. That > makes sense to me. Then I noticed the "Old Topics" and "New Topics" > lines added to the head of those messages when viewing the text. They > link the messages together anyway. How does that work? I will probably be flamed over this but, ... lurker decides if two messages are in the same topic by roughly comparing subject. Empirically I found that this is more reliable than looking at in-reply-to/references. So, first lurker munges subjects in a way which is usually pretty smart. It ignores stuff like Re: Fwd: [some-list] stuff on the front, and 'was:' stuff on the end. Then it normalizes case and whitespace and does an md5 hash and some more magic to get those 8 digit thread identifiers from the subject. Unforunately, this person used 'was ...' instead of 'was: ' so lurker was not able to discard the stuff after it. Then the missing space was enough to make the subject just different enough that lurker considered it a new thread. However, lurker is not totally niave. If it has in-reply-to/references data which links a different subject to another subject, it considers this a new topic / old topic link. So, if you wrote a message 'about cats' and I replied with a subject of 'about dogs (was: about cats)' then lurker will put these in seperate threads, but link my message to yours with old/new topic links. I would wager that in this case a person replied in a mail reader (establishing the in-reply-to) and then accidentaly editted the subject line. --=20 Wesley W. Terpstra <we...@te...> |