W3C home > Mailing lists > Public > www-ws-arch@w3.org > January 2004

Re: Web Services Architecture Document

From: Francis McCabe <frankmccabe@mac.com>
Date: Sat, 31 Jan 2004 18:20:03 -0800
Message-Id: <24B6B372-545D-11D8-81A1-000A95DC494A@mac.com>
Cc: "Katia Sycara" <katia@cs.cmu.edu>, "Stephane Fellah" <fellah@pcigeomatics.com>, <www-ws-arch@w3.org>
To: "Ugo Corda" <UCorda@SeeBeyond.com>

Ugo:
  Services do have URIs; that is clear. However, the URI of a service 
may not be dereferenceable. The URI of a service is entirely different 
to any URI that might be used to access its functionality, description, 
etc.
The primary purpose of a service URI has to be so that you can make 
inferences and declarations of the form: XXX used service xxxx at time 
TTTT.
Frank

On Jan 30, 2004, at 5:27 PM, Ugo Corda wrote:

>
> Katia,
> This is a very simple example, but I already have a problem with its
> practical usefulness. How would I verify that the service has a URI?
> Since we have not defined exactly what a service's URI is (is it the
> port address? Is it some document describing the service? etc.), then
> the compliance criterion derived from that reasoning does not buy us
> much.
>
> I suspect the same problem would surface with most other logical
> conclusions we could derive from the WSA ontology. (That's why the
> concept of WSA-compliance has a much more fuzzy meaning than the usual
> concept of standard compliance - this, by the way, is not the fault of
> WSA, it just comes from its special nature of being an architecture
> document instead of, for instance, the definition of a specific
> vocabulary).
>
> What I am challenging is the practical benefit of using the semantic
> machinery in the context of the WSA ontology, not the abstract concept
> itself.
>
> Ugo
>
>> -----Original Message-----
>> From: Katia Sycara [mailto:katia@cs.cmu.edu]
>> Sent: Friday, January 30, 2004 1:55 PM
>> To: Ugo Corda; 'Stephane Fellah'; www-ws-arch@w3.org
>> Subject: RE: Web Services Architecture Document
>>
>>
>> Ugo,
>>  If you define an X as a Web service then, since a Web
>> service is a service and since a service is a resource, then
>> this X has to have a URI (this is a very simple example, but
>> if this X does not have a URI, then it is not compliant with
>> what the wsa document calls a Web service).
>>   --Katia
>>
>>
>> -----Original Message-----
>> From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
>> Sent: Friday, January 30, 2004 4:44 PM
>> To: Stephane Fellah; Katia Sycara; www-ws-arch@w3.org
>> Subject: RE: Web Services Architecture Document
>>
>> Stephane,
>> I understand the idea of "automating the search of services
>> based on the agent criteria and perform semantic translation
>> of parameters between the services" (the UDDI TC has being
>> discussing exactly this kind of issues recently). I just
>> don't see how all that would relate to the WSA ontology and
>> leverage it.
>>
>> Ugo
>>
>>> -----Original Message-----
>>> From: Stephane Fellah [mailto:fellah@pcigeomatics.com]
>>> Sent: Friday, January 30, 2004 1:35 PM
>>> To: Ugo Corda; Katia Sycara; www-ws-arch@w3.org
>>> Subject: RE: Web Services Architecture Document
>>>
>>>
>>> Ugo,
>>>
>>> I have been actively involved in different OpenGIS web
>>> services testbeds
>>> (http://www.opengis.org/initiatives/?iid=79). The goal is to
>>> enable the integration of different information communities
>>> using geospatial information and services (map, feature,
>>> coverage, processing services...).
>>>
>>> One of the most obvious need for a web service ontology is to
>>> enable web-agent to perform automatic (read intelligent) web
>>> service discovery and choregraphy of services.
>>> Let's suppose you want to perform a complex task such as
>>> create a 3D map on a specific area. Your 3D Map agent will
>>> need to find the map from some Web Map Server (WMS) and the
>>> DEM from a Web Coverage Service (WCS) and use a Web Terrain
>>> Service (WTS) to create a 3D view from the data retrieved
>>> from the WMS and WCS.
>>> To be able to automate this task, there are two approaches.
>>> The first one is a syntaxic one. You define XML schema to
>>> describe each service and data information. The problem with
>>> this approach ? It does not scale. You have to write code to
>>> parse each schema and make semantic mapping  between the
>>> terms of different XML schema. With the floraison of XML
>>> schema standards that exist out there, you can be sure that
>>> integration of different systems is impossible.
>>> The second approach is a semantic approach , which deal with
>>> heterogeity. You describe the services and data with metadata
>>> using a common metamodel (read RDF/OWL). Using inferencing
>>> and rules and common upper ontologies, you can automate the
>>> search of services based on the agent criteria and perform
>>> semantic translation of parameters between the services. WSDL
>>> is far to be sufficient to be able to automate service
>>> chaining. In my scenario, you need to find geospatial
>>> information within a specific location. The information
>>> generated by the service will need to be provided in
>>> compatible formats for the WTS.
>>>
>>> To enable the semantic web, all the backend services and
>>> information needs to be viewed by agent as RDF graphs. Using
>>> semantic protocol, the web becomes a huge semantic bus and
>>> expert system. Instead of relying on specific protocols and
>>> syntax, the agents are communicating using semantic information.
>>>
>>> Best regards
>>>
>>> Stephane Fellah
>>> Senior Software Engineer
>>>
>>> PCI Geomatics
>>> 490, Boulevard St Joseph
>>> Hull, Quebec
>>> Canada J8Y 3Y7
>>> Tel: 1 819 770 0022 Ext. 223
>>> Fax 1 819 770 0098
>>> Visit our web site:  www.pcigeomatics.com
>>>
>>>
>>>
>>> -----Original Message-----
>>> From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
>>> Sent: Friday, January 30, 2004 2:47 PM
>>> To: Katia Sycara; Stephane Fellah; www-ws-arch@w3.org
>>> Subject: RE: Web Services Architecture Document
>>>
>>>
>>> Katia,
>>> I am trying to think of examples of how your idea of spec
>>> compliance verification could be applied.
>>>
>>> Are you saying, for example, that if the WSDL 2.0 spec were
>>> to be rewritten using OWL, then I could run a compliance
>>> verifier against the WSA ontology and find out that WSDL 2.0
>>> lacks intermediaries support? This seems rather far fetched to me.
>>>
>>> Ugo
>>>
>>>> -----Original Message-----
>>>> From: Katia Sycara [mailto:katia@cs.cmu.edu]
>>>> Sent: Friday, January 30, 2004 11:34 AM
>>>> To: Ugo Corda; 'Stephane Fellah'; www-ws-arch@w3.org
>>>> Cc: katia@cs.cmu.edu
>>>> Subject: RE: Web Services Architecture Document
>>>>
>>>>
>>>> Ugo, for one, as Stephen suggests the OWL formalization can
>>> be used as
>>>> an upper ontology for the work of groups such as the OWL-S
>>> coalition
>>>> or the Semantic Web Services Language committee (SWSL)
>> and Semantic
>>>> Web Services Architecture committee (SWSA). The upper OWL
>> ontology
>>>> could be further specialized by these groups, constraints
>> could be
>>>> added etc. In a long term view, one could imagine that if a
>>> new spec
>>>> for example were to be expressed in such an ontology, then
>>>> inferences about compliance of the new spec with the architecture
>>>> could be inferred.  Cheers, Katia
>>>>
>>>> -----Original Message-----
>>>> From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
>>>> Sent: Friday, January 30, 2004 11:34 AM
>>>> To: Katia Sycara; Stephane Fellah; www-ws-arch@w3.org
>>>> Subject: RE: Web Services Architecture Document
>>>>
>>>> What I have not been able to figure out so far is the "then what?"
>>>> part. In other words, what is the goal for the OWL
>> formalization of
>>>> WSA (besides being a showcase of semantic technologies)?
>> Is there a
>>>> plan to do anything with that formalization? What kind of results
>>>> would you like to achieve once you apply a reasoning
>> engine to that
>>>> information?
>>>>
>>>> Ugo
>>>>
>>>>> -----Original Message-----
>>>>> From: www-ws-arch-request@w3.org
>>> [mailto:www-ws-arch-request@w3.org]
>>>
>>>>> On Behalf Of Katia Sycara
>>>>> Sent: Friday, January 30, 2004 7:52 AM
>>>>> To: 'Stephane Fellah'; www-ws-arch@w3.org
>>>>> Cc: katia@cs.cmu.edu
>>>>> Subject: RE: Web Services Architecture Document
>>>>>
>>>>>
>>>>>
>>>>> Staphane,
>>>>>  We are working on an OWL formalization of the concepts and
>>>>> relationships in the Web Services Architecture. It will
>>> be published
>>>
>>>>> along with the final Working Group product by end of next
>>> week.  As
>>>>> for OWL-S it is not a Working Group of the W3C, though
>> some of us
>>>>> would like it to become one.
>>>>>   Cheers, Katia
>>>>>
>>>>> -----Original Message-----
>>>>> From: www-ws-arch-request@w3.org
>>> [mailto:www-ws-arch-request@w3.org]
>>>
>>>>> On Behalf Of Stephane Fellah
>>>>> Sent: Thursday, January 29, 2004 5:27 PM
>>>>> To: www-ws-arch@w3.org
>>>>> Subject: Re: Web Services Architecture Document
>>>>>
>>>>>
>>>>> Hi,
>>>>>
>>>>> I have a couple of questions related to the scope of
>> your working
>>>>> group. Is there any chance to see an OWL formalization of the
>>>>> different concepts and relationships exposed in the WS
>>> Architecture
>>>>> Document ? What would be the next step for W3C : define
>> again new
>>>>> XML schemas (syntaxic
>>>>> approach) or using semantic web technologies (OWL). I clearly
>>>>> favor the last option because the syntaxic approach is too
>>>>> brittle to scale on the web. The OWL-S effort seems to
>>>>> address the same problem, but uses different terms. Is there
>>>>> any harmonization effort between the working groups ?
>>>>>
>>>>> Thanks in advance.
>>>>>
>>>>> Best regards
>>>>>
>>>>> Stephane Fellah
>>>>> Senior Software Engineer
>>>>>
>>>>> PCI Geomatics
>>>>> 490, Boulevard St Joseph
>>>>> Hull, Quebec
>>>>> Canada J8Y 3Y7
>>>>> Tel: 1 819 770 0022 Ext. 223
>>>>> Fax 1 819 770 0098
>>>>> Visit our web site:  www.pcigeomatics.com
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>>
>
Received on Saturday, 31 January 2004 21:20:49 GMT

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