On 2/29/2016 2:32 PM, Ray Bellis wrote:
On 29/02/2016 22:27, John R Levine wrote:
The existing port and service registry already has all of the _service
names, and is updated as people invent new services. What's the benefit
of duplicating it rather than just pointing to it?
+1
So this commentr sequence is tending to convince me to have the document
stay away from saying anything about SRV's subordinate _underscore name
field, _Service.
However for _Proto, there remains a basic problem with the current SRV
arrangement, in the face of an _underscore registry: name collision.
If SRV continues to specify _Proto choices from a separate IANA
registry, there there are two, /independent/ places for doing
_underscore registrations. That model ensures name collisions.
While I doubt that all 7000 service names, or even a small fraction of
them, get used for SRV, whatever percentage does get used is likely to
be a large-ish number. So, yeah, having to duplicate the registration
is problematic.
In the case of _Proto names, the set is tiny, so the burden of
explicitly registering these names in an _underscore registry is not
onerous, given that it is the only way to ensure that name collisions
are avoided.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
