Ubuntuforums.org, Message-IDs, and the mail archives

[Matthew R. Dempsky]

I noticed that ubuntuforums.org's mailing list gateway doesn't generate 
unique Message-IDs for all emails it sends.  Their Message-IDs are based 
on the poster's username and the time of posting, but with a course 
enough granularity that if a poster sends multiple emails within a few 
minutes time frame, Message-IDs will be reused.

Further, Pipermail, the software running the Ubuntu lists archives, 
doesn't seem to handle duplicate Message-IDs well.  It checks for 
duplicates simply by comparing Message-IDs (allowing trivial archive 
corruption).  In the Threaded view, only the first occurence is visible, 
while in Subject, Author, or Date views, all occurences are visible, 
although they each link to the same message.

For example, check [1] and [2] for posts by ``linuxcity''.  He posted in 
the ``[Q] upgrade Flight-6 to Flight-7?'' and ``Packages ! get the 
current ones???'' threads, but only one is present in the archive 
because both contained <linuxcity.27lf9z at gs1.ubuntuforums.org> as their 
Message-ID.

Is this a known issue?  (I noticed this issue when trying to index all 
of Ubuntuforums.org's ricochet spam[3], but the problem affects several
legitimate posts as well.)

[1] https://lists.ubuntu.com/archives/ubuntu-users/2006-May/thread.html
[2] https://lists.ubuntu.com/archives/ubuntu-users/2006-May/author.html
[3] http://odin.alkemio.org/ubuntuforums.html