RE: Underscores

Given that the document hasn't yet been adopted, there's plenty of room for adjustments before last call.  😉



Personally, so long as there's a way to go from an HTTP origin to a DNS lookup, I'm amenable to just about any encoding that makes DNSop happy.



-----Original Message-----
From: Dave Crocker <dhc@dcrocker.net>
Sent: Friday, May 11, 2018 8:26 PM
To: ietf-http-wg@w3.org
Subject: Re: Underscores



On 5/11/2018 9:16 AM, Ben Schwartz wrote:

> Hi Dave,

>

> I'm working on a draft in HTTPBIS proposing a new RRTYPE that uses

> underscores:

> https://tools.ietf.org/html/draft-schwartz-httpbis-dns-alt-svc-02#sect


> ion-2

>

> Does this underscore pattern match your new recommendation?  If you

> could send your thoughts to the HTTP mailing list [1], that would help

> us make sure we're getting the structure right, since we're not DNS experts.

>

> Thanks,

> Ben Schwartz

>

> [1] https://lists.w3.org/Archives/Public/ietf-http-wg/






Folks,



G'day.  Ben asked me to respond to the wg mailing list.



First, thanks to Ben for noting the parallel effort on the global _underscore registry and for opening a dialogue about it and his draft.



Second, unfortunately I think that draft-schwartz-httpbis-dns-alt-svc

is, in fact, problematic with respect to its _underscore-related text, since it draws from whatever 'scheme' is desired, to be used as the global underscored name.



The SRV and URI specs introduce a imilar meta-problem for the _underscore registry, by drawing their global underscored DNS node name from other tables.  The underscore work in the dnsop wg is splitting the topic between a 'clean sheet' specification that defines the registry and creates initial entries, but does not try to fix any existing specifications.



A second dnsop draft:



    DNS Attrleaf Changes: Fixing Specifications with _Underscored Node Name Use

https://www.ietf.org/internet-drafts/draft-ietf-dnsop-attrleaf-fix-00.txt




was just created, to offer text that updates existing documents, which specify _underscored DNS node name use.



The dns-alt-svc spec adds yet one more such 'external' table as a source of underscored name, lacking coordination with the central registry we are defining.



My reading of the dns-alt-svc document status within this working group is that there is still time to adjust its text before wg last call and I eagerly hope the working group will consider making that modification.



The attrleaf-fix draft contains proposals for language that should be used for the meta-specification of SRV and URI underscored node name use. I believe the template I've offered for them can be easily adapted to your own draft.



The world won't end if you decide not to make the change at this time, but gosh it would be nice to get IETF specs to start using the new registry...



Thanks.



d/



--

Dave Crocker

Brandenburg InternetWorking

bbiw.net



--

Dave Crocker

Brandenburg InternetWorking

bbiw.net