Reversal of E-mail addresses

Alan Bleasby ajb at gserv1.dl.ac.uk
Mon Aug 24 16:15:51 EST 1992


It seems I'll get no rest until some sort of technical explanation is given
so, for the benefit of the millions out there interested in mailing systems
here goes (apologies to the few biologists who read this group as it'll
bore you to tears :-)

History: The problem was (I believe) pointed out by Mike Cherry some time
         ago; shortly after DL took over the IRLEARN lists. Rob has just
         brought the issue to light again although it had not been
         forgotten.

Problem: Direct insertion into NEWS (obviously) causes no problem. Email
         sent to DL from the continent does cause a problem as some UK
         gateways reverse the address [note that it is NOT Daresbury
         reversing the address].

Mechanism: Email sent to (e.g.) BIONEWS at DARESBURY is automatically inserted
         into NEWS by a sendmail/script mechanism.
         New NEWS messages are automatically extracted (by a similar
         mechanism) and sent to the mailserver which does the explosion
         from the lists of email subscribers.

What needs to be done is to reverse the address before insertion into NEWS.
The problem is complicated though by a move to SMTP. Any email sent via
SMTP from mainland Europe (etc) will have the addresses the correct way
round so these must NOT be reversed.

The problem is not insurmountable (in fact there are two or three ways it
could be solved... some rather gruesome). The most sensible way is to
alter the insertion script. This involves recognition of major domains,
domain name servers etc and is the slightly less obnoxious kludge. The
matter has been delayed as the chap responsible for the NEWS insertion
aspect has been on a rather long holiday (whatever they are).

In short, the matter is known and is in hand. This `feature' does not
affect BIOSCI distribution; you've just got to be careful how you REPLY
directly to the original poster. If it doesn't work reverse the address.

The application of such a modification will result in addresses being the
`wrong way round' from the point of view of UK subscribers (incidentally,
the system was not designed with inbuilt favouritism for UK users... its the
international gateways causing the problem & not DL remember), however
gateways from the UK to the rest of the world are intelligent. If the
address is not known they reverse it and try again automatically e.g.
the addresses `genbank.bio.net' and `net.bio.genbank' both succeed from
the UK.

I hope this lays the matter to rest. I'll post to the group when the matter
is (shortly) resolved.

Alan Bleasby
BIOSCI
SERC Daresbury Laboratory
Warrington WA4 4AD
UK
email: bleasby at daresbury.ac.uk ... or is that bleasby at uk.ac.daresbury
       I'm sure that driving on the left-hand side of the road will
       eventually catch on.



More information about the Bioforum mailing list