- From: John C Klensin <john-ietf@jck.com>
- Date: Fri, 24 Feb 2012 11:43:31 -0500
- To: Patrik Fältström <paf@frobbit.se>, Andrew Sullivan <ajs@anvilwalrusden.com>
- cc: mnot@mnot.net, IETF Discussion <ietf@ietf.org>, iesg@ietf.org, ietf-http-wg@w3.org
--On Friday, February 24, 2012 16:58 +0100 Patrik Fältström <paf@frobbit.se> wrote: > On 24 feb 2012, at 16:38, Andrew Sullivan wrote: > >> Over in spfbis, people are arguing that the SPF RRTYPE should >> be deprecated and abandoned in SPF because nobody uses it >> because of practical difficulties in getting new RRTYPEs >> deployed. What makes us think that the arguments in favour >> of SRV are going to find more fertile ground? > > Because people disagree on whether it is actually hard to get > new RRTYPEs deployed. > > I for example do completely disagree on it being hard. Sure, > your user interface in the gui of your favorite $EDITOR might > not support the new RRTYPE, but should that constrain > deployment of good standards? Patrik, While I don't see it as hard as Andrew does, I don't see it as easy either. The problem isn't one's favorite $EDITOR. It is the number of folks who, for lots of reasons, haven't upgraded from operating systems, resolvers, etc., that don't support newer RRTYPES. Remember that, while Windows 7 is getting some of the market share that Vista never did, there are still a huge number of XP systems out there -- many of them in the hands of people and organizations who aren't going to make the hardware investment to upgrade in a difficult economy. The situation with the Mac platform actually isn't much better -- I know a lot of people who haven't (and can't) upgrade from OS 9, much less to the latest carnivorous feline. SRV has the advantage of many year's head start and more utility to more protocols over, e.g., SPF but the reality remains that, if some users have support for an important new RRTYPE and others don't, we have either an inconsistent user experience (more confusion, more support calls, etc.) or, as we have seen with DNAME, a need to carry around complex workarounds that can lead to bugs, vunerabilities, or, ahem, inconsistent user experience. And, fwiw, I'm actually much more concerned about use of the DNS for any part of the process if we don't have broad consensus in and between the IETF and the implementer community about whether IRIs or URIs are authoritative and about how domain names containing IDN labels are encoded in either. I'm equally concerned if some organization of our mutual acquaintance is determined to declare the existence of equivalent names in the DNS that will create user expectations that certs and comparison mechanisms cannot support. --On Thursday, February 23, 2012 14:38 -0800 Doug Barton <dougb@dougbarton.us> wrote: > 2782 was published 12 years ago this month. I suppose it can be > considered mature enough to deploy at this point? :) One might wish. But real maturity has to be based on implementation and deployment, not just publication dates. I note, for example, that RFC 736 (an old favorite example of mine for several reasons) was published over 34 years ago is, like 2782, still at Proposed Standard, and, sadly, hasn't gone anywhere (at least in the last decade or so). Certainly it is elderly as Internet protocols go, but I don't think the leap from "elderly" to "mature" is obviously justified. john
Received on Friday, 24 February 2012 16:44:10 UTC