RE: i0001: EPRs as identifiers - alternative proposal

Hi Savas!

i'm puzzled by the 'scaling' argument. OK the Bank might not 
hand out specific telnos to customers, but only because the CLI
scheme is of limited length. However the URI scheme does scale 
pretty well and could well be used by the bank to give each 
customer their own address:

http://mybank.example.com/customer/paul_downey
mailto://accountno-12223213432@mybank.example.com

as opposed to addressing messages:

mailto://soap@mybank.example.com

 <someHeaderWhichCouldBeAParamOrAPropertyHelpImConfused>
   <accountNo>12223213432</accountNo>
 </someHeaderWhichCouldBeAParamOrAPropertyHelpImConfused>

Paul


-----Original Message-----
From: public-ws-addressing-request@w3.org
[mailto:public-ws-addressing-request@w3.org]On Behalf Of Savas
Parastatidis
Sent: 03 December 2004 10:31
To: David Orchard
Cc: public-ws-addressing@w3.org
Subject: RE: i0001: EPRs as identifiers - alternative proposal



> -----Original Message-----
> From: David Orchard [mailto:dorchard@bea.com]
> Sent: 02 December 2004 21:07
> To: Savas Parastatidis; Francisco Curbera; public-ws-addressing@w3.org
> Subject: RE: i0001: EPRs as identifiers - alternative proposal
> 
> I don't know what the difference between a bank account and a banking
> service is.  What is the difference between the 2 from a software
> perspective?
> 

You are not given a telephone number specific to your bank account to
address it whenever you want because that doesn't scale for the bank.
Only very few customers (the very-very good ones) may be given a
direct/personalised phone number to access their account.
No-money-in-bank-account-Savas gets the banking service's general phone
number and have to _explicitly_ identify the bank account for which a
balance check is requested (always getting the same "no money" answer
though :-) This scales.

However, this is now turning into an architecture/design discussion but
that's the only way I can think of I can explain the difference I see.

> I'll note that WSDL 2.0 service elements can describe most, if not all
> of the Web resources.
> 
> We do Web services so that we can have interactions that are typically
> characterized by sending SOAP messages and has WSDL descriptions, but
> there is no clear technical distinctions between Web services and non
> Web services (such as Web or Semantic Web) technologies.  There are
> marketing differences definitely.
> 

I think we continue to disagree. While I see the role of semantic web
technologies in a service-oriented architecture (semantic descriptions
as metadata for services), I personally see a difference between the
resource-focused Web and a service-oriented world. But we are going to
architecture matters again.

> I think that if you look at the heart of your differentiation, you'll
> find that they are circular or illogical.  Resources are things that
are
> identified by "universal resource identifiers".  Services are things
> that are interacted with in service oriented manner.  Etc.
> 
> So, I suggest not trying to differentiate between "services" and
> "resources" for the purpose of EPRs as identifiers issue.
> 

Apologies for starting a "EPRs are addresses of what" thread under this
subject but it seemed relevant.

Best regards,
.savas.

Received on Friday, 3 December 2004 15:38:21 UTC