On Fri, Apr 02, 2004 at 07:15:53PM +0100, Tim Jackson wrote:
> many people deeply involved with writing networked applications (not
> limited to MS, by the way, though they are a primary culprit) either
> really don't appreciate the implications (often serious in their impact on
> others) of what they are doing, or have their priorities upside down.
This has been reinforced for me by the Glastonbury Ticketing fiasco (that
some of you may have read about). I don't believe the culprit is exchange,
so I'm not going to slate a particular product. It was, of course, also set
up by clueless muppets. For an operation that was handling 115,000 tickets,
and sending out the confirmation emails:
Mistake 1: sending the ticket confirmation email from a BT Openworld DSL line
Mistake 2: EHLO nobody.com
Mistake 3: somehow managing to get my MX confused with someone else's,
causing me to get relay mail for someone else
Their helpdesk autoresponder came from a machine with broken rDNS (PTR to
something which was NXDOMAIN). The inbound mail system for the return-path
of the confirmation message doesn't believe in PIPELINING.
I'm really less than impressed. I worry about the relay mail, because that
smacks of some threading problem, or DNS cache issue.
MBM
--
Matthew Byng-Maddick <mbm@???> http://colondot.net/
