I strongly disagree. The DNS protocol deserve cleanup. Deprecating RRTYPEs 
doesn’t mean the will stop working on the day the RFC is published, neither are 
people going to backport the removal of RRTYPEs to existing DNS software 
releases.

It just means - whatever ancient stuff you are using - you are on your own now. 
It’s same as with the stuff that never got the RFC. You can you whatever you 
want, it will be your responsibility (and costs), not the implementors. So, the 
people can keep their old system running, they just can’t expect the current 
DNS to interoperate with them.

Paul, sorry, but the argument “but I know of people running” ancient systems 
can’t be used at every attempt to cleanup the kitchensink protocol that DNS is 
right now.

Ondrej 
--
Ondřej Surý — ISC

> On 23 Mar 2018, at 17:55, Paul Vixie <p...@redbarn.org> wrote:
> 
> 
> 
> Ondřej Surý wrote:
>> Thanks, now I understand what you are asking for;), so what about:
>> 
>> “No existing Internet Standard uses these Resource Records and there no
>> know practical usage in the public Internet.”
> 
> i think this is overbroad. if we aren't also sure that it's not being used in 
> some private network somewhere, we should not tell implementers to remove 
> support. a lot of private networks use internet protocols and implementations 
> to support their local apps and users.
> 
> mf, mg, mb, and mail1 are i think still in use on some as/400 intranets.
> 
> when i removed UID and GID from BIND it was because there was no RFC, not 
> because i wasn't fully aware of some older athena implementations still in 
> use at that time which used these instead of TXT.
> 
> ideally we'd put out an extended call for comments about anything we'd like 
> to remove if it ever worked to anyone's knowledge. if it never worked, like 
> extended label types in EDNS, they can just be removed.
> 
> this is how we handled IQUERY deprecation and i think it went well.
> 
> -- 
> P Vixie
> 

_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to