[Ietf-not43] XML-RPC for an information service?

[Patrick]

On Wed, Aug 07, 2002 at 02:34:23PM +0200, Stephane Bortzmeyer took time to write:
> > such that a client starting with some label (e.g.
> > a dns name) can find the authoritative server and make the
> > relevant query. 
> 
> This could be done on the client-side, too, like many whois clients do
> at the present time.
> 
> "\.(com|net|org)" -> whois.networksolutions.com
> "\.(info)"        -> whois.affilias.info
> "\.(.*)           -> whois.nic.$1

A system that would not need to update *all* client software to
change something is something interesting... Like DNS for example:
you do not need to change your resolver when ICANN adds a new TLD.

It could be done, but it should not to. That is my point of view.
The mechanism in place should handle this sort of things
transparently.
Also I am not sure a system like the one you propose could be used
for all purposes in CRISP. Ok, easy for domain names (at least now),
but how for IPv4 adresses ? You plan to have each client software
having a mapping between netmasks and IP Registries ?

Personnally I would imagine that something like DNS SRV records may
be useful.

> > But until I'm a little more convinced you understand what problem
> > this group is solving, 
> 
> I understand much better the issues, thanks to your explanations. But
> I agree with the Ted Hardie's point of view: Crisp has several
> problems to solve and my proposal more-or-less fits one (querying a
> given server). Unlike you, I do not think it is the less important
> question.

I would dare to say that it is the easier part. This is also a part where
many ideas already exists (SOAP, RPC-XML, Corba, whatever), and it
``just'' needs to be adapted to the current setup. It is not to say
that your work is less important, just that there are other parts.
(the charter have 6 points, querying is one among the 6)

Ok, I have done nothing, so who I am to criticize anyway...

Patrick.