W3C home > Mailing lists > Public > www-ws-arch@w3.org > October 2002

Re: Words for the Triangles

From: David Booth <dbooth@w3.org>
Date: Wed, 02 Oct 2002 18:17:52 -0400
Message-Id: <5.1.0.14.2.20021002165218.03f3d4d8@localhost>
To: www-ws-arch@w3.org
Cc: "Christopher B Ferris" <chrisfer@us.ibm.com>

"Christopher B Ferris" <chrisfer@us.ibm.com> on Tue, 1 Oct 2002 12:45:51 
-0400 wrote:
>There is more to making a Web resource discoverable via the Web than
>simply making the resource available via the Web. . . . there needs to be some
>action taken that makes a given resource discoverable via the Web.
>
>The action can be implicit or explicit, but it MUST be recognized in the
>architecture.

Yes, I totally agree.  But the ACTION is the act of publishing.  It is the 
verb, or arc, on the right side of the triangle diagram -- not the cloud at 
the top.  And the act of discovering is the verb or arc on the left side of 
the triangle -- also not the cloud at the top.

The cloud at the top is a noun.  It represents "where that description 
resides in between the time it is published and the time when it is 
discovered".  And it could be anywhere in the Web of electronically 
addressable space.

The acts of publishing and discovery do NOT need to be using the same 
previously agreed repository or "meeting place".  A Service might publish 
its description in one place (a "registry", perhaps), and the Client may 
find that description in a totally different place -- completely unrelated 
to the original registry where the Service had placed it.  This happens all 
the time on the Web.  Someone publishes some info in one place, and soon it 
pops up someplace else, in a location that is totally unrelated to its 
original location.

Yes, an act is needed to publish the Service description.  And yes, an act 
is needed to discover that description.  Those are the right and left arcs 
in the triangle diagram.  But they do not have to share a common "meeting 
place" or "registry" (short of the entire Web itself).

>I believe this action to be the act of registering with, or
>publishing to, a "registry", even if that "registry" is merely a 
>collection of links off
>the default page of an origin server. . . .

It CAN be, if you choose to use a centralized, previously agreed "meeting 
place" model.  My point is that it doesn't HAVE to be.  And our 
architectural diagram should not imply that it has to be.

>I am also happy with "discovery agencies" if "registry" is too overloaded 
>a term.

That term only describes the agencies that the left arc might use -- the 
discovery action.  The right arc might use "publishing agencies".  (And 
remember, my whole point is that the left and right arc do not have to be 
using the same mechanism, or agency or "meeting place"!)

I believe the best term for the top cloud is simply "The Web".  And the 
prose can explain that it represents the entire Web of electronically 
accessible data -- some of which could even be a local file system, a 
repository or whatever.

However, if others really like the word "agencies" then I would be okay 
with using the term "publishing/discovery agencies", since that at least 
acknowledges the fact that they are separable roles.


-- 
David Booth
W3C Fellow / Hewlett-Packard
Telephone: +1.617.253.1273
Received on Wednesday, 2 October 2002 18:16:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:09 GMT