RE: Label for Top Node of "triangle diagram"

David,

I don't think anyone disagrees with these points. But a role does NOT 
imply a central place
as many on this list have said, repeatedly. No one to my knowledge is 
advocating an exclusive, centralized
model for discovery. I don't think the architecture implies this at all 
(especially since we haven't written 
it down yet because we're wasting all our time on this inane debate!!!) I 
think that some assume
that "registry" is some secret code word for UDDI , but this is not the 
case. If the architecture
were written such that it said in prose:

        "Thou shalt use the UDDI node at foobar.com and none other; also 
sprach
        Zarathustra"

with accompanying thunderbolts and lightening, and music playing in the 
background
then that would certainly merit a spirited debate such as this. However, 
no one has proposed
such language, we are discussing the label for the top node in the triange 
diagram that
defines an abstraction of the architecture.

If that label were "UDDI", then it might warrant such a debate, but that 
has not been
proposed (limiting/restricting/constraining discovery to UDDI) either to 
my knowledge.

A colleague of mine recounted a story when the CPP/A team was stuck on 
naming
something. It involved a similar situation on a project where someone 
stepped forward
and said; "What's in a name? Let's call the darn thing 'Herman' and move 
on".

DaveO has proposed that we label the top node "DRegistries" marking it as 
draft and has similarly
asked that we table this debate and move on with defining the 
architecture. I couldn't
agree more.

Cheers,

Christopher Ferris
Architect, Emerging e-business Industry Architecture
email: chrisfer@us.ibm.com
phone: +1 508 234 3624

www-ws-arch-request@w3.org wrote on 10/02/2002 05:37:48 PM:

> 
> 
> >[Ricky] It has to be "exposed" in some way. But it doesn't need to be a 

> >"central place". Think about the Gnutella model.
> >
> ><daveh> Nope, not a central place, but where and how? There is no 
implicit 
> >hypertext network to crawl, so how are service providers to assure that 

> >their service is found? No matter what we say, the providers and 
discovery 
> >service operators will find a way. Sometimes they will be satisfied 
with 
> >an more organic like Gnutella, other times they will insist on a more 
> >regimented model like MP3.com.
> 
> Yes!  I agree completely.  There is no need to have a "central place" 
for 
> advertising/discovery of Web Service descriptions, and our architecture 
> should not imply that any such thing is required.
> 
> Then where and how can a Client find a Service?  Well, the "where" 
depends 
> on the "how".  In the general case, the "where" is the Web of all 
> electronically accessible data.  If the "how" is "by a UDDI request", 
then 
> the "where" is easy to identify: It's a particular UDDI registry.  But 
if 
> the "how" is "by a Gnutella search", then "where" is not so easy to 
> identify, nor is it so relevant.  The important thing is just that 
Gnutella 
> finds the data you want -- somewhere on the Web -- and gives it to you.
> 
> In short, there are many ways that a Client might find a Service, and 
the 
> Services that the Client finds may depend on the mechanisms that it uses 
to 
> perform the discovery.  This is a feature, not a bug.  It means freedom 
of 
> choice.  It's one of the nice things about the Web.
> 
> Of course, the downside is that you will never be sure that you have 
> discovered ALL of the available Web Services.  Unfortunately, that's 
life, 
> and it's okay.  It is inherent in the Web's open-world model.  New 
Services 
> can come and go as "dark matter" unbeknownst to you.  And if you don't 
like 
> dealing with the uncertainty that that implies, then you can just sign 
up 
> for a single, big, centralized discovery service and pretend that 
nothing 
> else exists.   Depending on what you want to do, that may be a very 
> reasonable approach.  It is analogous to shopping only at a single, big, 

> centralized department store instead of dealing with a plethora of 
> merchants scattered throughout an entire city.  Sometimes it makes 
> sense.  But our architecture shouldn't imply that we only support the 
> centralized model.
> 
> 
> -- 
> David Booth
> W3C Fellow / Hewlett-Packard
> Telephone: +1.617.253.1273
>