Re: Email archiving [was Re: Stop the "John in Alaska" thread title!]

From: Jacques Du Pasquier (
Date: Fri Jan 25 2002 - 09:50:49 MST

Jacques Du Pasquier wrote (25.1.2002/12:18) :
> Robert J. Bradbury wrote (24.1.2002/16:53) :
> > On Thu, 24 Jan 2002, Simon McClenahan wrote:
> >
> > > I have saved about 10 years worth of email (various formats) in the dream
> > > that one day someone, maybe me, will create software to match whole email
> > > messages together and identify quotes as links to other messages. [snip]
> > >
> > > Has anyone done any work on this?
> >
> > (...)
> > (see:
> > and
> Note that forward (replies to this message) and back links (this
> message replies to) are constructed based on "Reference"/"In-Reply-To"
> header fields, not by looking at quotes.

I have to correct that a bit. From the hypermail FAQ :

How does hypermail decide whether messages are in the same thread?

It uses the In-Reply-To: header if that is available. If not, it uses the References: header if that is available. If these are not available, it looks for previous messages with the same subject header. Matches based on the subject are listed as "maybe" replies.

If the linkquotes option is used, it will also search the message bodies of the previous searchbackmsgnum (default = 500) messages for text that looks like it is being quoted in the current message. If it finds one or more prior messages with such text matches, it will treat the one with the longest match as the message being replied to. The exact algorithms for deciding what is quoted text are a bit complex. These matches override the In-Reply-To: and References: info (this rule may deserve further thought).

So some kind of attempt is made in that sense. You should test and see
how/if it actually works.


This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 13:37:36 MST