Re: When to break threads (was: alternative threading algorithms for sloppy mailing list)
On Sat, Sep 06, 2003 at 10:24:42PM +0200, Johan Almqvist
<johan-mutt@xxxxxxxxxxxx> wrote:
> * "Daniel E. Eisenbud" <eisenbud@xxxxxxxxxxxxxx> [030906 22:17]:
> > > right, and I appreciate it. I think Alex wants to go a bit further:
> > > break thread automatically if subject changed.
> > I could do this.
> > Hmm, give me an hour or two for the patch. :-)
>
> I wonder if that's always the right thing to---see the subject of this
> mail :-)
It's not alwawys the right thing to do, of course. But on some mailing
lists it would be a net improvement, and since it will be an option,
people can do it on a mailbox by mailbox basis. I'll point out where in
the patch people can try different heuristics for when to keep a thread
together despite a subject change, if anyone's inclined to experiment.
Maybe a decent heuristic, which might be pretty easy to implement, would
be to keep the thread together if one subject is a substring of the
other, or if they have at least a certain number of characters in
common? Of course, varying levels or "Re:" can be automatically taken
care of by real_subj as determined by reply_regexp, but a sensible
heuristic should also deal with whitespace damage within the subject --
it's fairly common.
-Daniel
--
Daniel E. Eisenbud
eisenbud@xxxxxxxxxxxxxx
Computational Biology Center
Memorial Sloan-Kettering Cancer Center