[ietf-dkim] Re: Issue 1382

[Scott Kitterman]

On Tuesday 17 October 2006 17:05, david repking wrote:
> > I'd be satisfied if the requirements draft were to say:
> >
> > The protocol MUST NOT require use of a new DNS RR type.  The protocol MAY
> > allow for optional use of a new RR type.
>
> And then nobody will use the new (optional) RR type, if there's an
> alternative.
>
> IMHO, new types of data MUST require new types of RR
>
> If we would re-use RR types, why don't you wan't arbitrary numeric
> values on A records (like telephone numers, extensions, system IDs,
> etc) ???

Reuse of TXT in an underscored subdomain is benign. 

It all depends on if you actually want SSP to be deployed or not.  If we have 
to wait for a new RR type to be broadly useable, then SSP goes nowhere out of 
the starting gate.

An optional new RR type would be there to provide a path away from TXT if one 
should be required.  Get started now and deprecate reuse of TXT in the far 
future if there proves to be operational value in doing so.

I guess we'll get to argue about this again later.

Scott K