Re: [DNSOP] redefining SRV, was New Version Notification for draft-ietf-dnsop-attrleaf-03.txt

[Dave Crocker <[e-mail]>]

On 3/20/2018 10:16 AM, John R. Levine wrote:
>>  We need to move away from the complexity created by having special 
>> rules for
>>  some entries in the registry.
> 
> That would be fine except that the Port and Service registry has 
> thousands of entries, and the named ones (nearly all of them) are valid 
> SRV names. Importing a handful of names that we guess are commonly used 
> with SRV is the worst of both worlds.

Again: the plan is to change the SRV spec so that it doesn't try to 
inherit all those values automatically.


> Either point at the real registry for SRV, or say that this explicitly 
> redefines the namespace for SRV and see if dnsop will go for that.  I 
> don't see any way you can just ignore RFC 6335 and its predecessors.
> 
>>  ps.  I thought the URI RR had no current actual use (or at least very
>>  little.)
> 
> I belive that's correct, but again, either we deprecate URI or we deal 
> with its naming rules.

Or, here too, we change the URI spec.


d/

-- 
-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net