Re: Merging RRP and Whois

I would expect that some object types (such as ones dealing with
domain names) will be fully specified as the result of the current
work on the RRP.  Some thought should be given to future types -- but
just enough to allow the protocol to grow later.  Very fortunately,
there is no need for too much work on these types now -- otherwise
there would be no time to do it.

A little can go a long way, however, to ensure the future of the
protocol.


On Wed, Jan 24, 2001 at 05:35:10PM +0000, Martin Oldfield wrote:
> >>>>> "George" == George Belotsky <[email protected]> writes:
> 
> -> snip <-
> 
>     George> In order to do so, it is very helpful to view the
>     George> registration system as a number of object repositories
>     George> (which may contain things other than domain names), with
>     George> each repository being managed by a registry.  The object
>     George> repositories should all support a single interface, which
>     George> allows for a variety of users.  These include superusers
>     George> (i.e. the registry performing maintenance) and privileged
>     George> users (i.e. the registrars registering, modifying,
>     George> deleting and transferring objects) as well as users with
>     George> only moderate privileges (e.g. paid subscribers to an
>     George> advanced Whois service) and minimally privileged users
>     George> (i.e. users of the public Whois).
> 
>     George> If we start separating users into different privilege
>     George> categories by writing a different protocol for each
>     George> category, how many protocols will we end up with?
>     George> Clearly, a single protocol is the most logical solution.
> 
> Quite so!
> 
> Presumably besides the notion of a general object registry, we would
> also need to concoct standard definitions for e.g. a domain, a
> contact, &c. ? If we don't do this then we'll lose interoperability to 
> a proliferation of different domain definitions.
> 
> If we're going to embrace such diversity, then would it also be
> prudent to let each registry publish details of the classes of objects
> they store, and the access policies which apply to them ?
> 
> Cheers,
> -- 
> Martin Oldfield,
> AdamsNames Ltd.
> 

-- 
-----------------------------
George Belotsky
Senior Software Architect
Register.com, inc.
[email protected]
212-798-9127 (phone)
212-798-9876 (fax)