Re: Commercial/Real-world Semantic Web Services?

Dr. Martin,

Dr. McDermott told me that, he was thinking "we were speaking of
'services on the semantic web'"; but semantic Web services usually means
"semantic treatment of web services."  

Do you want to clear that you were NOT speaking of "services on the
semantic web" when you designed and developed OWL-S? 

That introductory paragraph in OWL-S definitely reveals the purpose and
trick - OWL-S are about (semantic) Web-based services, or services on
the semantic Web. 

How about the processModel in OWL-S? You still ***ignore*** this
question and problem - the processModel is designed for Web-based
services, not Web services.

When you develop "a" hotel reservation service, why and how you need
that processModel?

When you develop "a" airline ticketing service, why and how you need
that processModel?

When you develop "a" rental car reservation service, why and how you
need that processModel?

When you develop "a" restaraunt reservation service, why and how you
need that processModel?

When you develop "a" address geocoding service, why and how you need
that processModel?

When you develop "a" IP address geocoding service, why and how you need
that processModel?

Your processModel is designed for a Web-site, not Web service. And that
processModel is full about modeling, not Web-site, or Web-service. 

Here, the ***nature*** of the question and problem is not a verbal
combat on that introductory paragraph in OWL-S, it's your processModel.

When you consider to make corrections to OWL-S, don't just change some
words in document, but focus on the processModel and tell people whether
it is designed for a Web-site, or a Web-service.

As for that VTA tutorial, it's well-known that WSMO is almost the same
as OWL-S with some differences. Both OWL-S and WSMO are for (semantic)
Web-based services, not Web-services. Almost their whole interest is in
service aggregation/integration/mediation through modeling. 

When we develop a Web-service, we don't need your processModel. When we
develop a Web-site, we MAY need it if we deal with multiple
Web-services. But we MAY NOT need it when we just ues a single
Web-service.

However, the framework for semantic Web services should be something
that can be implemented universally by Web-services, not Web-sites, or a
part of Web-sites. 

Don't mix up Web-sites with Web-services as they are a big different.
The new version of OWL-S should focus on the "semantic treatment of web
services". Otherwise, we have to ask you again whether your
service-provider is a Web-based service provider, or a Web-service
provider. If you do not clarify such difference, you probably misuse the
terminologies again.

Regards,

Xuan




>>> David Martin <martin@AI.SRI.COM> 10/16/2006 2:12 AM >>>

Hello Xuan -

I have no desire to engage in verbal combat, but there are a few of
your 
comments here that call for a response and some clarification.

Xuan Shi wrote:
> Luke,
> 
> If you agree that "True web services and websites are a different",
> it's so easy to understand the definition of a Web-service specified
by
> W3C  @ http://www.w3.org/TR/2004/NOTE-ws-gloss-20040211/ 
> 
> In this glossary, when W3C define the concept of "Web service", it
> said:
> 
> "There are many things that might be called "Web services" in the
world
> at large. However, for the purpose of this Working Group and this
> architecture, and without prejudice toward other definitions, we
will
> use the following definition:
> 
> A Web service is a software system designed to support interoperable
> machine-to-machine interaction over a network. It has an interface
> described in a machine-processable format (specifically WSDL). Other
> systems interact with the Web service in a manner prescribed by its
> description using SOAP-messages, typically conveyed using HTTP with
an
> XML serialization in conjunction with other Web-related standards."
> 
> Forget that section "without prejudice toward other definitions", at
> least (semantic) Web-sites are in those many things, not Web-services
-
> Web-service should at least have a WSDL interface, not Web interface,
a
> place for Web service integration/aggregation/mediation. 
> 
> Furthermore, what is a service-provider, and what is a
> service-requester? A service provider provides services through WSDL
> interfaces. A service requester consumes services through WSDL
> interfaces. It's a P2P relationship, not a server/end-user
> relationship.
> 
> Then we can see, OWL-S, or semantic-Web-based services, deliberately
> keep misusing terminology to confuse people for years. They mixed up
> their Web-sites with Web services to sell their modelings. 

Xuan, it's very unkind, baseless, inappropriate, and unreasonable for 
your to mis-characterize the motives of this group of people (the OWL-S

Coalition) in this way.  I believe you will find that when you use this

kind of unjustified, incendiary rhetoric, the readers of this list will

quickly recognize that your comments are without merit.

> 
> OWL-S claimed clearly from 2001 to 2006 that - "By 'service' we mean
> Web sites" -  it's wrong, right? as you know "True web services and
> websites are a different". 

It is inappropriate to take a partial quotation out of context in this

way, and try to make it seem as if the authors meant something else. 
We 
certainly did not mean to say that all Web sites are Web services.  In

fact, the passage in question was not really attempting to give a 
complete definition of "service" or "Web service", but rather was meant

to indicate that we were primarily concerned with services that have 
side effects (as opposed to purely information services).  Here is the

full paragraph where that appears:

"Among the most important Web resources are those that provide
services. 
By ``service'' we mean Web sites that do not merely provide static 
information but allow one to effect some action or change in the world,

such as the sale of a product or the control of a physical device. The

Semantic Web should enable users to locate, select, employ, compose,
and 
monitor Web-based services automatically."  [1]

Now it may be that this passage isn't written clearly enough.  It may 
also be that it is in need of some update, as it was written some time

ago.  We will consider whether to refine it.  The reason we used the 
phrase "Web sites" was because we were reluctant to use "Web service"
in 
this particular context.  This passage is in an introductory section, 
and is discussing the general context and motivation for the work on 
OWL-S.  Because we recognized that "Web service" was understood 
differently by different people, we tried to use a more neutral phrase

in this early part of the document.  When this passage was written, I 
believe there was probably less clarity on W3C pages regarding the 
meaning of "Web service" than there is now.  Also, we felt that this 
discussion was very general and abstract, and should not necessarily be

limited to services that are implemented using WSDL.

> 
> What's the goal of OWL-S? It's said again from 2001 to 2006 that
"The
> Semantic Web should enable users to locate, select, employ, compose,
and
> monitor Web-based services automatically." -  it's wrong, right? A
Web
> service is not a Web-based service.

Here again, I believe we were trying to avoid the phrase "Web service"

simply for the reasons given above.  Again, there is probably room for

improvement in the writing.  But certainly there was no intention to
use 
the phrase "Web-based service" in a confusing manner, and I doubt if it

has caused any confusion.  I think most readers have understood that it

was just a phrase that meant "services accessible via the Web", without

being too precise about the technology by which they are accessible. 
Again, please remember that this passage is in an introductory section,

and is discussing the general context and motivation for the work on 
OWL-S.

> 
> Besides what they said in their theory, we can see what they did in
> their practice. From this well-known tutorial
>
http://www.wsmo.org/TR/d17/resources/200507-ICWS/SWStutorial-iswc05.ppt

> 
> as well as other papers, demos, we can see they are really talking
> about and dealing with Web-site. What is VTA on slide #63? By their
> terminology, VTA is a service-provider. Unfortunately by W3C
> specification on Web-service, VTA is only a service-requester,
because
> VTA consumes the true Web-services (FlightBooking and HotelBooking)
> through WSDL interface. VTA is only a Web-based service provider
target
> the two end-users (Consumers ) on the Web. That's why I remind
everyone
> when Bijian boasts his "Web services", he may actually talking about
> Web-sites. And that's why I suggested that OWL-S remove processModel
> becasue it's designed for a Web-site (VTA), not a Web-service
(either
> FlightBooking or HotelBooking).

While I'd love to claim some credit for it, the VTA example you mention

above was created by researchers working on WSMO, not OWL-S.


> 
> Who has been keeping misusing terminology to confuse people? OWL-S -
> their service provider is not the same terminology used in
Web-service
> definition. Are they really interested in Web-site? No, they have
> nothing about how to develop a Web-site. Then are they really
interested
> in Web-services? No, their tutorial has little about those two true
> Web-servcies, FlightBooking and HotelBooking. What are they really
> interested in? If we know what are their vested interests, the answer
is
> clear.
> 
> OWL-S, if not completely wrong, is problematic. Do they want to hear
> any truth? They fear to see and acknowledge their problems. They
have
> little or no courage to face the questions and problems. Some of
them
> are eager at gossipping in the darkness while betrayed their
partners
> due to carelessness, rather than make correction to generate a more
> appropriate theory to guide their practice. You already see that
> gossiper yelled - ignore Xuan, he is a notorious  troll. Another one
of
> them said in this IG before that Xuan is an outsider (of us).
However,
> unfortunately they cannot ignore those questions and problems.
> 
> Until now, they have no courage to face the real world Web-services,
> rather than any problems in real practice. In the past six years
from
> 2001 to 2006, what we got from OWL-S are only those
> Kindergarten-toy-like service models, buybook, buyticket,
purchaseorder,
> stokequote, etc.

While it's true that the examples in the release are simple examples
for 
tutorial purposes, it is also true that there are a number of 
publications describing projects that have used OWL-S in more 
interesting ways -- some by members of the OWL-S Coalition and many by

people in the larger community of users.

Regards,
David Martin

[1] The quoted passage appears in the Introduction of the OWL-S 
Technical Overview available here:
   http://www.daml.org/services/owl-s/1.1/overview/ 
The same passage also appears in earlier releases.

> In 2007, if they want to play with some real world
> Web-services, they may wish to try my old examples that are still
living
> in practice -
> http://www.arcwebservices.com/services/v2006/AddressFinder.wsdl and
> http://arcweb.esri.com/services/v2/AddressFinder.wsdl 
> 
> Anyone would like to identify which string data type actually
contains
> a hex code? Adding semantic annotations into these two WSDL
documents?
> OK, it may tell us this variable is a string and is a hex code. Is
this
> the goal of SWS? Do you think that stupid "agent" understand when
this
> variable = "A" or "C"  (a hex code), the agent knows what and how to
do?
> We need something more than those Kindergarten-toys.
> 
> At last, if you check the archives, you can find what I suggest is
just
> that - we need more standards/agreements/protocols to develop
semantic
> Web services, not avoid the need for agreement. However, the leading
> roles of this IG don't believe that we cannot reach any agreement
> because they believe any service provider has the absolute right to
do
> what s/he wants to do. They try to avoid the need for agreement - in
> that case, they can find the value for their modelings.
> 
> Regards,
> 
> Xuan
> 
> 
> 
> 
> 
> 
>>>>"Luke Steller" <Luke.Steller@infotech.monash.edu.au> 10/15/2006
> 
> 11:08 AM >>>
> Xuan,
> 
> I dont understand your preoccupation with the definition of a
service.
> True
> web services and websites are a different. If thats a problem, just
> use
> SWS for web services. Use semantic web for web sites.
> 
> Sure, agreement is an issue but at least SW and SWS gives us a
common
> structure for agreement, which we did not have before.
> 
> Do you have work/suggestions for avoiding the need for agreement?
> 
> Luke
> 
> 
> 
> 
> On 10/15/06, Xuan Shi <Xuan.Shi@mail.wvu.edu> wrote:
> 
>>
>>Michael Uschold concluded a general law in his paper "Where are the
>>Semantics in the Semantic Web?" published by AI Magazine, 24 (3),
> 
> 25--36,
> 
>>that
>>
>>"The more agreement there is, the less it is necessary to have
> 
> machine
> 
>>processable semantics." What are "machine processable semantics"?
> 
> Maybe
> 
>>Bijan's logic modelings?
>>
>>This means, if we have more agreements in this community, people
> 
> like
> 
>>Bijan will have nothing to do. And that why he ignored such issues
> 
> and keeps
> 
>>misleading, if not cheating, the world.
>>
>>In the history, most people believed that the earth was the center
of
> 
> the
> 
>>universe. Eventually it's proved such an idea was wrong. But we all
> 
> know
> 
>>what happened in the history when someone told the truth. It's the
> 
> same in
> 
>>SWS community.
>>
>>I just hope that all people just don't believe such "authority" like
> 
> Bijan
> 
>>or something else, but have to have more critical and independent
> 
> spirit in
> 
>>"scientific" reesarch - pursue the truth, not the authority, as
> 
> those
> 
>>well-respected scientists may be wrong, and definitely can be wrong.
>>
>>If anyone would like to read this paper:
>>
> 
>
http://www3.interscience.wiley.com/cgi-bin/fulltext/109560959/PDFSTART

> ,
> 
>>you can see how a Nobel Prize winner in 1940s, prevented those
> 
> different but
> 
>>correct approaches from discussion and publication in the history -
> 
> his
> 
>>students could only make corrections and published the result after
> 
> he died.
> 
>>In the history, even some well-respected "scientiests' wanted to
> 
> send
> 
>>those who against them to Mars. Today, someone(s) in this SWS-IG
> 
> just
> 
>>blocked my discussion and emails two times - they just repeated the
>>historical events in nowadays.
>>
>>I actually don't care whether Bijan ignore me or not - my future
life
> 
> and
> 
>>career will not depend on anyone in this community. But Bijan cares
-
> 
> he
> 
>>fears about that when people know the truth and generate more
>>agreements/standards/protocols, he might have to find somethign else
> 
> to do.
> 
>>If we know he has such obvious vested interest on his specialties
> 
> rather
> 
>>than agreement/standard/protocol, we can ignore him and his
products.
> 
> And
> 
>>that why he told us that he "personally don't know of any
> 
> (successfully)
> 
>>commercial or production uses of OWL-S, WSMO, or the  like,..."
>>
>>When Bijian boasts his "Web services", ask him first whether he is
> 
> talking
> 
>>about a Web-site or not. When he boasts his modelings, just ask him
> 
> why and
> 
>>how can we use his modeling when we develop ''a" Web service, such
as
> 
> the
> 
>>favorite "AirlineTicketing" or "HotelBooking" kind of services, then
> 
> we can
> 
>>understand he is actually talking about how to modeling a Web-site,
> 
> not that
> 
>>two real services. If Bijan would like to remove the process
modeling
> 
> from
> 
>>OWL-S, I would welcome it, though there's not even a tiny bit of
> 
> evidence of
> 
>>that, alas. Sigh.
>>
>>As a "scientist", you can igore me or anyone who is againts you, but
> 
> you
> 
>>cannot ignore the problems and questions. As the chair of this IG,
> 
> you
> 
>>cannot fear that people who are against you will ask you questions.
> 
> At least
> 
>>in this country, the bi-partisan politics can teach you even the
> 
> President
> 
>>of US  cannot ignore the problems and questions from the other-side.
> 
> As a
> 
>>team leader, you have to learn a lot.
>>
>>Regards,
>>
>>Xuan
>>
>>
>>
>>>>>Bijan Parsia <bparsia@isr.umd.edu> 10/13/2006 3:14 PM >>>
>>
>>And sigh, I meant that to be private, obviously. I fully expect a
>>Xuanslought. Which I shall ignore. As I recommend to everyone. If
>>Xuan Shi would like to change this dynamic, I would welcome it,
>>though there's not even a tiny bit of evidence of that, alas. Sigh.
>>
>>Well, I might as well take this faux pas as an opportunity to point
>>out that, contrary to certain fantasies:
>>       <http://www.w3.org/TR/ws-arch/>
>>is a *note*, not a recommendation. Just one more distortion
>>exaggerated into a crusade.
>>
>>Which is *such* a good tag line!
>>
>>Cheers,
>>Bijan.
>>
>>
>>
>>
>>
>>
> 
> 
> 

Received on Monday, 16 October 2006 15:28:21 UTC