To:
"Hollenbeck, Scott" <shollenbeck@verisign.com>
cc:
"'Damaraju, Ayesha'" <ayesha.damaraju@neustar.com>, ietf-provreg@cafax.se, brunner@nic-naa.net
From:
Eric Brunner-Williams in Portland Maine <brunner@nic-naa.net>
Date:
Wed, 16 May 2001 17:13:14 -0400
In-Reply-To:
Your message of "Wed, 16 May 2001 15:21:10 EDT." <DF737E620579D411A8E400D0B77E671D750AB4@regdom-ex01.prod.netsol.com>
Sender:
owner-ietf-provreg@cafax.se
Subject:
Re: Epp - Contacts
Scott, As Dave pointed out, there is ambiguity in <contact:addr>, and the order (read from the bottom) of multiple PO Boxes, Private Bags, Rurual Routes, and Street Addresses (as many as fit into a 6x30), is last one wins. As is, in either version of draft-hollenbeck-epp-contact-* that I'm aware of, we don't guarentee uniqueness, nor ordering, though finitness appears to come dead cheap. With a little ingenuity at least two postal addrs could fit in a 6x30 (I've lived with a Rural Rte #, a Private Bag, and a PO Box, all at the same offset from US 101 in S. Humbolt Co., CA US). I'd like to give my telephone number (which doubles as a fax), but in case I'm not in, I'd like my answering service (paid professional) to pick up, and I don't want my local "Intelligent Network" to do call forwarding, as that is insufficiently deterministic (to put it mildly). So, there's the motivation. The ordering is (reverse of the surprising postal bottom-up reading) order of incidence (which places a no-reordering on second and subsequent element list members requirment). Repeat for fax, my answering service (lady down the street) got the same model I did (I bought it for her as partial payment for professional services). I've a lot of email addresses, the one I'm using today is subject to the whims of a DSL vendor (now that's chancy, neh?), and if the wire or the host goes south, my primary alternate is pretty damn stable (world.std.com), but I could have used a TimeWarnerAOL dhcp'd and mailbox hosted bit-bucket, and that too can vanish (like a DSL vendor) without even a .vacation forwarder. So, there's the motivation. The ordering is (reverse of the surprising postal bottom-up reading) order of incidence (which places a no-reordering ...). Unbounded seems kind of goofy, picking a bound is going to be an exercise in engineering arbitrariness, and 1 is just as arbitrary as 2 or 0, though the last has the same ordering properties as the first. Cheers, Greetings, Regards, ... Eric