Re: Exentisibility in the requirements (fwd)

[Sheer El-Showk <sheer@laudanum.saraf.com>]

On Wed, 24 Jan 2001, George Belotsky wrote:

> Actually, it is quite clear that provreg is quite broad in capability.
> 
> This is why merging the Whois functionality into the RRP would work so
> well.  Whois will eventually need more features, while the RRP has
> methods to accomplish Whois-type queries (one does need to just look
> at the registered objects).  Both the Whois and RRP must reference the
> same objects in any case: the RRP at a very high privilege, and the
> Whois at the lowest possible privilege.

I don't think privilage level is really the relevant consideration -- as
an extensible protocol, ProReg/RRP would benefit greatly from a privilage
scheme whether or not it is used for whois.  Nor do I think the argument
about different usage types and load are relevant in considering whether
to integrate whois into the protocol we are considering.  The protocol
just defines the way we access the data ... the backend system
implementation will just have to be different for the transaction based
ProReg and the anonymous/read-only whois, but that doesn't mean they can't
share the same protocol.

I think the real reason _not_ to incorporate whois into the protocol is
the fundemental difference in the way data is accessed and stored in the
two cases (someone mentioned this point before but I havn't seen more
disucssion of it).

Whois requires a referral (possibly) to another agency, while ProReg does
not.  While this clearly diffrentiates them in terms of implementation, I
would argue that it would not be a good idea to mix them, even at the
level of a standard or protocol.  The resultant protocol would have to
support terminology for transactions and non-transactions, local and
referred lookups, anonymous and access-based privileges.  While such
extensibility would be admirable, most of these fundemental design changes
would be due to the very particular nature of whois and not benefit the
protocol in general (though I would be very interested to hear of any
other possible uses for these extentions to the basic nature of the
ProReg protocol).

To summerize I think whois is a very particular problem differing greatly
in character from the issues involved in ProvReg and I see no benefit in
mixing the two.

Sheer

> 
> 
> 
> On Wed, Jan 24, 2001 at 12:55:09PM -0500, Edward Lewis wrote:
> > With this talk about merging provreg and whois, it seems to me that some
> > folks have gotten the impression that provreg is limited to domain names.
> > Provreg is looking at domain names in the near term with an eye towards
> > other data types in the future.  That is what the charter says.
> > 
> > Perhaps the current requirements document does not reflect this.  The
> > current requirements document is focused at domain names because of the
> > near-term effort.  Perhaps the requirements need to strenghten the notion
> > that the protocol must be extensible.
> > 
> > I would enourage anyone who feels that the requirements document is too
> > domain- name-centric submit comments and words to reflect that domain names
> > are just a current focus, more work is to follow.  I don't want to bog the
> > requirements document in discussion of other data types but rather strongly
> > recognize that other data types will be considered in the future.
> > 
> > -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
> > Edward Lewis                                                NAI Labs
> > Phone: +1 443-259-2352                      Email: lewis@tislabs.com
> > 
> > Dilbert is an optimist.
> > 
> > Opinions expressed are property of my evil twin, not my employer.
> > 
> > 
> 
> -- 
> -----------------------------
> George Belotsky
> Senior Software Architect
> Register.com, inc.
> george@register.com
> 212-798-9127 (phone)
> 212-798-9876 (fax)
>