* Dallman Ross <dman(_at_)nomotek(_dot_)com> [2005-02-18 13:57]:
The conclusion in the final paragraph, I find not to be entirely
useful. Duplicate messages rarely are absolutely identical in terms
of perfectly matching headers. Even the body can differ, in that
list identifiers are often added to the bottom, and so on.
That's true, but Adrian mentioned partial headers; so we would only
look at headers that match when a message splits off to the list and
the direct recipient. Maybe also ensure that headers differ in ways
that we would expect them to - ie. only one of the dupes should have
the same List-ID header.
Checksums on the bodies could get tricky because as you say, some list
software modifies the body. The slightest modification to content
produces a totally different checksum. Ideally the tool would be
sophisticated enough to measure degree of similarity between two
bodies of text.
I understand the danger being described. But I think it is
just another example of a solution looking for a problem. We
can think of innumerable theoretical problems. They become
actual problems when someone actually takes advantage of them.
I don't think throwing in the towel is an answer either. Duplicate
messages are a nuissance, and it's definately a problem worthy of a
solution. Message IDs are open to malicious attack and failure due to
not being unique, and social pressure doesn't work in a world full of
Outlook users.
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail