W3C home > Mailing lists > Public > www-ws-arch@w3.org > May 2003

Re: Separate concepts for "service" and "targetResource?" (was RE : /service/@targetResource ?)

From: Walden Mathews <waldenm@optonline.net>
Date: Thu, 22 May 2003 01:41:35 -0400
To: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, www-ws-arch@w3.org
Message-id: <00fc01c32024$cf8ad6c0$1702a8c0@WorkGroup>

> As far as I know, a "resource" is anything with identity that can
> meaningfully be accessed via the Web.  What the WSDL people call a
> targetResource has identity (it's a specific piece of code that performs a
> well-defined operation) and it is accessed via the Web (or at least by Web
> gateways).  For example, a specific printer attached to a mainframe that
> be accessed by a HTTP/SOAP gateway is a "resource" with identity, or at
> least the specific software agent that "drives" the printer from the Web
> service point of view is a Web resource.

It seems like "the printer", conceptually, is the resource, while
the driver software and even the hardware are kind of ephemeral.
The "software agent" view of that strikes me as a sort of 'techie'
way of trying to be conceptual, and not quite making it.  What if
there wasn't any software?

> If this is too confusing from the Webarch perspective we can drop the term
> "resource" and just refer to the software drivers as "agents."  That's OK
> with me, but the WSD group specifically chose the term "resource" because
> seemed to fit the Webarch concept of a "resource" ... and has a URI that
> WSDL can use to determine equality of identity.

A agree that if people want to refer to running snippets of code,
they're better off saying "agent" than "resource".  You can resource-ify
anything, but it can sure get confusing.  I think the error here is in
deciding that the URI identifies some bits of code, instead of something
meaningful in the application domain.

> > The URI binds to the meaning, not to the code.
> Hmm, that's an interesting point of view.  We did talk about this a bit
> clearly the targetResource is not a specific bag of bits, because it's
> identity is defined by the operations it performs on a specific "thing"
> is hard to define.  In the printer example, the "thing" is the physical
> printer, but printers aren't really Web resources in the Webarch are they?

If someone's car can be one, why not a printer, so long as it's on
wheels? :=)

> It gets messier if the "thing" is some piece of software: if we rewrote
> Web service interface so that it is in Java rather than C#, has its
> identity/URI changed?  I don't think so, so in that sense I agree with
> Walden ... but on the other hand its manifestation *is* a piece of code
> one runs.  Unless we want to get into the idea that a specific piece of
> is the "representation" of the abstract resource that defines its
> functionality, I think it's easier to say that the agent sitting on the
> "turtle" is the targetResource.

Keeping with the car <slash> printer-on-wheels metaphor, that
"representation" idea is definitely in the break down lane.  Let's not
go there.

 >  But if WSD is happy calling it targetAgent
> rather than "resource", I'm happy ... so long as something has a URI that
> does what WSDL needs to do with it.

The good news would be that the term "resource" would not
get so badly abused.  The bad news is that a technology ostensibly
built on some aspect of the "web" would lose perhaps the most
important word in the lexicon.  Flip a coin. :-)

Received on Thursday, 22 May 2003 01:40:03 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:05:52 UTC