To:
"'Eric Brunner-Williams in Portland Maine'" <brunner@nic-naa.net>
Cc:
"'ietf-provreg@cafax.se'" <ietf-provreg@cafax.se>
From:
"Hollenbeck, Scott" <shollenbeck@verisign.com>
Date:
Fri, 21 Sep 2001 14:21:10 -0400
Sender:
owner-ietf-provreg@cafax.se
Subject:
RE: Another To-Do Thought
> -----Original Message----- > From: Eric Brunner-Williams in Portland Maine > [mailto:brunner@nic-naa.net] > Sent: Friday, September 21, 2001 12:04 PM > To: Hollenbeck, Scott > Cc: 'ietf-provreg@cafax.se'; brunner@nic-naa.net > Subject: Re: Another To-Do Thought > > > Scott, > > Hmm. > > Assume we've a base or core or MUST spec. E.g., epp-04 et seq. > > Assume we've one or more extensions or MAY spec(s). E.g., beep, containers, > push, data protection, et seq. > > Assume someone has something for which no standards-track I-D exists which > normatively defines. E.g., trademark (.info registry private xml blob, any > errors are mine, and copyrighted). > > Are r-*-private thingees <unspec>, or are they <extension>? If the thingees are shared between client and server and not part of the base specs, but they exist as a some form of additional elements added to the protocol, they are extensions. I don't think the form of documentation (I-D, RFC, proprietary document) matters, as long as there's a schema somewhere that defines the thingee. If a schema exists, the thingee isn't unspecified by definition. As I said in my original note, I don't think "unspec" conveys the real purpose of the element. It's a place to add extensions, and as such using "extension" seems more appropriate. <Scott/>