W3C home > Mailing lists > Public > www-ws@w3.org > April 2002

RE: potential users of web services

From: Andrew Layman <andrewl@microsoft.com>
Date: Mon, 8 Apr 2002 16:40:48 -0700
Message-ID: <C3729BBB6099B344834634EC67DE4AE105359D28@red-msg-01.redmond.corp.microsoft.com>
To: <www-ws@w3.org>
Agreed.  We should think about the concept, and think about it in terms
of what is new and different about the kinds of services that we are
building. Rather than looking at the name and trying to _deduce_ what a
Web service might be, we should look at the actual Web services that we
are building and _induce_ the definition that distinguishes them from
other nearby technologies such as RPC/RMI and Web sites.


-----Original Message-----
From: Toufic Boubez [mailto:boubez@saffrontech.com] 
Sent: Monday, April 08, 2002 1:06 PM
To: www-ws@w3.org
Subject: RE: potential users of web services

I'm sorry I didn't catch this thread from the beginning, having been
away
from my office last week and just catching up on my email!
Having sort-of "been there at the creation", let me first say that, in
my
mind, the term "Web Services" is a very unfortunate misnomer, since a
definition of a "Web" service doesn't really require the Web. It was
just a
marketing buzzword that some people who shall remain nameless came up
with
at the time. In my group at IBM, when we were hammering out the details
of
UDDI and other Web Services components, it was called Service Oriented
Architecture, and I'm sure Andrew Layman can tell us that it was
probably
also called something else at Microsoft.

Having said that, my view of web services is that it's any platform- and
implementation-independent software (or even functionality, although you
can
always wrap functionality such as pizza baking in a software interface!)
that can be:

* described using an agreed upon or well known description language (for
example WSDL is nice but not required)
* published to an agreed upon or well known registry (for example UDDI
is
nice but not required)
* discovered through an agreed upon or well known mechanism
* invoked over the network through its declared API

I usually add "composed with other services" but that's somewhat
circular!
Now, you'll see that there's no mention of XML, SOAP, etc. These are
wonderful and extremely useful standards (the "agreed-upon or well
known"
part) but nothing prevents an internal (within the firewall)
implementation
consisting of C programs listening on sockets and sending binary data
whose
format is well-known and described in the organization to be labeled
"Web
Services".

I realise that this view might differ a bit from the "accepted" view,
considering that the "Web Services Architecture Stack" consists of XML
and
Schema at the bottom of every layered cake, so I might make an exception
for
XML. But nowhere is it suggested that the other components such as SOAP
or
UDDI are required.

Thoughts??

  --  Toufic

Toufic I. Boubez, Ph.D.
Chief Technology Officer
Saffron Technology
1600 Perimeter Park Drive, Suite 300
Morrisville, NC 27560
boubez@saffrontech.com 

919-468-8201 Voice (x109)
919-468-8202 Fax

"In theory, there is no difference between theory and practice. In
practice,
there is"


-----Original Message-----
From: Andrew Layman [mailto:andrewl@microsoft.com]
Sent: Monday, April 08, 2002 3:40 PM
To: www-ws@w3.org
Subject: RE: potential users of web services


Private mail to me suggests, and I agree, that the definition could be
slightly improved as:

A Web service is a computational service, accessible via messages of
definite, programming-language-neutral and platform-neutral format, and
which has no special presumption that the results of the computation are
used primarily for display by a user-agent.

-----Original Message-----
From: Andrew Layman [mailto:andrewl@microsoft.com] 
Sent: Monday, April 08, 2002 12:29 PM
To: Mark Baker
Cc: www-ws@w3.org
Subject: RE: potential users of web services

I believe that the services you cite fit my definition of Web service
quoted below.  I could perhaps be more concise: 

A Web service is a computational service, accessible via messages of
definite, language-neutral and platform-neutral format, and which has no
special presumption that the results of the computation are used
primarily for display on a user-agent. 

Hope this works for you.

-----Original Message-----
From: Mark Baker [mailto:distobj@acm.org] 
Sent: Friday, April 05, 2002 12:03 PM
To: Andrew Layman
Cc: www-ws@w3.org
Subject: Re: potential users of web services

On Fri, Apr 05, 2002 at 09:10:46AM -0800, Andrew Layman wrote:
> The term Web service was created to contrast with two earlier
> technologies.  On the one hand, it identifies a distinction from "Web
> site" in that a Web site serves pages, typically in HTML, for display
in
> a browser to a human, while a "Web service" offers a computation
> directly to anther computer, with no special expectation that the
> computation will be used in a browser or for display to a human. Web
> services are not computer-to-human but computer-to-computer.

Well, if it's the HTML that you're concerned about, why not return some
XML or RDF via HTTP GET?  That's machine processable.  And any piece of
software can invoke HTTP GET on a URI, no human required.

What about this?  http://www.xmlhack.com/rss10.php

It's an RSS feed for xmlhack.com.  No "getXmlhackRss()", just
"GET /rss10.php".  It's also not easily human parseable.

I don't know why that's any less a Web service than getStockQuote().

MB
-- 
Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA.      mbaker@planetfred.com
http://www.markbaker.ca   http://www.planetfred.com
Received on Monday, 8 April 2002 20:11:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:40 GMT