🚀 go-pugleaf

In article <Bw2E9L.9Kx@mtholyoke.edu> jbotz@mtholyoke.edu (Jurgen Botz) writes:
>In article <1992Oct13.113431.7685@infodev.cam.ac.uk> ag129@cus.cam.ac.uk (Alasdair Grant) writes:
>>like POP2/POP3/IMAP2/IMAP3/..., and these protocols are not stable enough
>
>I don't buy this argument.  What do you mean by "not stable enough"..?  Do
>you think that these vendors never change their protocols?  A lot of them
>won't even bother maintaining backward compatibility when they do!  (They
>have less need to do so since customers can usually upgrader their whole
>site if they choose to upgrade).  Internet protocols are far more concerned
>with interoperability and are generally "stable" for much longer stretches
>of time than proprietary vendor protocols.

But MS, Lotus or whoever will sell the client and server as a package.
The _whole point_ of POP and IMAP is that they provide interworking 
between different vendors' clients and servers; in particular the server
is usually public-domain (no fancy user-interface stuff). So if a vendor
wants to have control over the protocol, they would have to sell both
client and server.

Yes, Internet protocols are generally stable, but remote-mail-management
is an exception! Just look at the IMAP2 vs. IMAP3 controversy for example.
It does look as if IMAP2bis is settling down, but there are a number of
unanswered problems (authentication in the absence of Kerberos, quota 
management, manual and automatic forwarding, interface to the Directory)
and I get the feeling some people are waiting for CMU's Andrew II messaging
system to take a more coherent approach. (IMAP2's specialisation on the
'Retrieval' function is in line with Internet no-redundancy policy, but
has pitfalls in practice. To me, redundancy means having five competing
mail-retrieval protocols!)