Re: RFC3731: domain:roid in <info> response

Andrew Sullivan wrote:
> It might be useful to clients over time, though: a domain name need
> not be unique over time (whereas the roid is defined in such a way
> that it ought to be).  When the record is current, of course, the
> sponsor could know what the roid is (because they had it originally,
> and the object hasn't been removed from the repository).  But it is
> undoubtedly extremely useful for retrospective co-ordination for
> client and server operators to be able to refer to different objects
> of the same name by unique ids.  

Yep, that was the only use case that came to my mind - addressing a specific
instance of a delegation instead of "the currently active one", eg. for
retrieval of historic data of a ceased delegation (because the "name" would
obviously address the current one). However, for that to work, also requests
would need to allow using a roid instead of the domain name - and that is
not the case (yet? ;).

Anyways, points (especially about backwards compatibility) taken. now
crafting a creative way to derive roid's from ENUM domain instances ;)

cheers

Alex