Re: some replies aren't really replies

On 17 May 1999, Christina Williams Heikkila <cwh_at_circe.gsfc.nasa.gov> wrote:
>
> One problem with using the In-reply-to message ids to determine threading
> is that many people (including myself) keep a large mail inbox. When we
> want to send an email to a person, we just scroll up to one they sent to
> us, and just reply to it, changing the subject line. This creates threads
> where there shouldn't be.
>
> I don't know if this has come up in the past, or if there's any conceivable
> way to work with this - I'm just throwing it out here in case anyone's
> given it much thought.

Christina,

Using a screwdriver to drive a nail doesn't make the screwdriver a hammer, and similarly using a backlog of messages in your mail user agent's indexing function doesn't make it an addressbook. You might successfully drive a few nails using the screwdriver, but you've got to expect some skinned knuckles and a few nicks in the wood.

There is an old agage in computer programming that dates to the punch card era. As the saying goes...

   garbage in, garbage out

You can't legislate social intelligence, of course. People are always going to "break" computer tools by coming up with "creative" ways to to use the tools which programmers can not envision. But unlike human beings, computers aren't creative beings. You can't expect a computer to do much more than process the data which is presented to it.

I for one believe a routine that would try to check the "In-Reply-To:" or "References:" header against the "Subject:" and make intelligent inferences about what the human user intended is a) going to be doomed, to failure and b) is going to significantly increase processing time in any event.

Personally, I don't want see it happen. I'm already dealing with large archives...

-- 

   Craig A. Summerhill, Systems Coordinator and Program Officer
   Coalition for Networked Information
   21 Dupont Circle, N.W., Washington, D.C.   20036
   Internet: craig_at_cni.org   AT&Tnet (202) 296-5098