RE: Resource definition

James

I think I agree, but here are a few more questions ... 

When you say a "The *protocol* that defines how to flip the switch *is* a
resource", what exactly are you referring to, for example? Is it:
1. The specification that describes the behavior of the protocol and the
structure of the messages it contains, or
2. The execution of an instance of the protocol, or
3. The representation of the information used to flip the switch (i.e. the
message)

I think only the first one is a resource.

The reason I ask is that you can draw analogies with what happens in a B2B
communication where an order, for example, is sent from a buyer to a seller.
Sending this message changes the state at the seller in the same as
"flipping the switch".

Also is the order a resource? In which case where does the resource reside
at the buyer, at the seller or at both. If it's at both is it the same
resource. Alternatively is the order just the message and therefore not a
resource.

Thoughts?

David


-----Original Message-----
From: James M Snell [mailto:jasnell@us.ibm.com]
Sent: Tuesday, February 18, 2003 5:02 PM
To: Burdett, David
Cc: Assaf Arkin; David Orchard; Michael Mealling; 'Cutler, Roger
""(RogerCutler)'; www-ws-arch@w3.org; www-ws-arch-request@w3.org
Subject: RE: Resource definition


The *act* of flipping the switch is *not* a resource.  The *protocol* that 
defines how to flip the switch *is* a resource. 

- James Snell
     IBM Emerging Technologies
     jasnell@us.ibm.com
     (559) 587-1233 (office)
     (700) 544-9035 (t/l)
     Programming Web Services With SOAP
         O'Reilly & Associates, ISBN 0596000952

     Have I not commanded you? Be strong and courageous. 
     Do not be terrified, do not be discouraged, for the Lord your 
     God will be with you whereever you go.    - Joshua 1:9



"Burdett, David" <david.burdett@commerceone.com>
02/18/2003 04:38 PM

To
James M Snell/Fresno/IBM@IBMUS, Michael Mealling 
<michael@verisignlabs.com>
cc
Assaf Arkin <arkin@intalio.com>, David Orchard <dorchard@bea.com>, 
"'Cutler, Roger \"\"(RogerCutler)'" <RogerCutler@chevrontexaco.com>, 
www-ws-arch@w3.org, www-ws-arch-request@w3.org
bcc

Subject
RE: Resource definition



In an earlier email, Francis McCabe said ...

>>>... to use one of Mark's earlier examples, turning on a light switch. 
The
state of the light is a representable resource; no question. The action of
flipping the switch is not so representable. The arm which is used to flip
the switch is, however, a resource; although any representation of it in
terms of bits is merely a symbol and not the real thing.<<<

So is the act of moving the arm to turn on/off the switch a resource?

The reason I ask is that I think that moving the arm in this example is
analagous to the sending of a business document between partners, as it is 
a
message that is designed to change the state of the destination and not
actually resource.

Anyone agree/disagree?

David


-----Original Message-----
From: James M Snell [mailto:jasnell@us.ibm.com]
Sent: Tuesday, February 18, 2003 4:18 PM
To: Michael Mealling
Cc: Assaf Arkin; David Orchard; 'Cutler, Roger ""(RogerCutler)';
www-ws-arch@w3.org; www-ws-arch-request@w3.org
Subject: RE: Resource definition



> I think that's the right approach. I personally would like to see the
> _web_ architecture to be broad enough to be synonymous with the RFC 2396
> definition. Its things like RDF, REST, web services etc that should have
> the more narrow constraints....

Agreed. The Web Architecture needs to define the key resources classes,
and their relationships to one another.  For example, you have Application
Resources and Protocol Resources.  Protocol Resources constrain how you
access and manipulate Application Resources.  REST, Web Services, etc are
types of Protocol Resources.  Application Resources have such and such
properties.  Protocol Resources have such and such properties.  While the
naming may stink, the idea should be clear.

- James Snell
IBM Emerging Technologies
jasnell@us.ibm.com
(559) 587-1233 (office)
(700) 544-9035 (t/l)
Programming Web Services With SOAP
O'Reilly & Associates, ISBN 0596000952

Have I not commanded you? Be strong and courageous.
Do not be terrified, do not be discouraged, for the Lord your
God will be with you whereever you go.    - Joshua 1:9

Michael Mealling <michael@verisignlabs.com> wrote on 02/18/2003 03:34:37
PM:

> On Tue, 2003-02-18 at 18:21, James M Snell wrote:
> > Ok, that's fine, but where do you draw the line?  If everything
becomes a
> > resource then the idea of a resource begins to lose value.  Simply
> > asserting that "All Web services have an identity and are therefore
> > resources" doesn't seem to add any value.

> Its not intended to add very much value other than the fact that you now
> have an identifier for it. Its what  you build on top of that by making
> system specific assumptions and standards that ends up being useful. But
> its also a fact that you can't build something system specific until you
> have a common way of identifying nodes on the network.

> Take IP for example, without TCP, UDP and ICMP its rather a useless
> protocol. But it does give you an addressing scheme and a  basic packet
> format that you then build on to make useful things happen. Resources by
> themselves are pretty useless. But once you have something like RDF and
> a set of rules for what to expect in an RDF graph, well then you have
> something!

> > Personally, I believe that Dave
> > Orchard's question in the original note that I responded to should
have
> > been something like "1. The W3C TAG should darned well write up what
> > **class of resources are most relevant** from a web architecture
> > perspective"

> Now that's a much more interesting and tractable question!

>
> > If the answer to that question includes the class of
> > resources that describe mechanisms of accessing other resources then
> > you've got something valuable because you've narrowed down the scope.
If
> > the answer is "any resource with an identity" then y'all will just
> > continue to spin your wheels and not get anywhere.
> >
> > Ok, I'm done... shutting up and going back to my corner ;-)

> I think that's the right approach. I personally would like to see the
> _web_ architecture to be broad enough to be synonymous with the RFC 2396
> definition. Its things like RDF, REST, web services etc that should have
> the more narrow constraints....

> -MM