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

RE: Web services and the Semantic Web

From: Eric Newcomer <eric.newcomer@iona.com>
Date: Tue, 28 May 2002 07:22:10 -0400
To: "Anne Thomas Manes" <anne@manes.net>, "Mark Baker" <distobj@acm.org>
Cc: <www-ws-arch@w3.org>
Message-ID: <FFEDIILGIAFELAMIJNMNEECMDPAA.eric.newcomer@iona.com>
Yes, Anne has captured it well, it's the higher order services we need to
focus on, not the transport level, except to the extent that the transfer
protocol has a bearing on the architecture of the higher level services.

I take it this is what Mark is trying to point out, although sometimes I
have to say it seems like he's returning to his old anti-SOAP position,
encouraged by the recent "slap on the wrist" he keeps referring to.

Can we please just stick to accepting what's done and widely adopted (SOAP,
WSDL) delta any improvements or corrections necessary to support our
architecture (whatever it becomes), and work on the defnintion of the
requirements for and relationships among the higher level services?

I agree with the requirement to be consistent with existing Web
architecture, I don't think anyone disagrees with that, but I think we also
need to be consistent with SOAP and WSDL also, and build from there rather
than drag the whole discussion back down to first principles again.

Eric

-----Original Message-----
From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On
Behalf Of Anne Thomas Manes
Sent: Monday, May 27, 2002 4:20 PM
To: Mark Baker
Cc: www-ws-arch@w3.org
Subject: RE: Web services and the Semantic Web


For the most part I agree that ACID isn't suitable for the Internet, but
there are times when acidity is required, which is one reason why SOAP
supports other protocols (e.g., JMS)that might be more suitable for ACID
behavior.

I agree with you that we're trying to reinvent CORBA on the Web. But this
time we want to do it in a loosely coupled fashion that can tolerate immense
scalability, latency, and even intermittent connectivity. (Something that
you simply can't do with CORBA.) Looking back at the CORBA process, creating
IIOP was the easy part, even though it took years to achieve consensus and
to then make it interoperable. The hard part was designing the higher-order
CORBAservices that address transactions, security, etc. These high order
services never reached the point of pervasiveness and interoperability.
(Security has been particularly disappointing.)

Fortunately we're not trying to reinvent IIOP. We want to build this new
distributed computing infrastructure based on an immensely scalable network
that automatically manages really hard things like partial failures, network
transparency, etc. But we still need to build those gnarly high-order
services -- and we want to do it in such a way that they're interoperable
from the beginning.

I really think that the Web and XML can allows us to be successful this
time.

Anne

> -----Original Message-----
> From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On
> Behalf Of Mark Baker
> Sent: Monday, May 27, 2002 12:18 PM
> To: Anne Thomas Manes
> Cc: www-ws-arch@w3.org
> Subject: Re: Web services and the Semantic Web
>
>
> Oh, is this what Eric was talking about?  I thought by "distributed
> computing concepts" he was talking about managing things like
> partial failure, desirable degree of network transparency, evolution
> and extension, etc..  All things that the Web already addresses.
>
> If he meant those things, then I agree (sort of - e.g. ACID isn't
> suitable for the Internet).
>
> MB
>
> On Mon, May 27, 2002 at 08:44:22AM -0400, Anne Thomas Manes wrote:
> > The Web architecture doesn't address a number of core requirements for
> > application-to-application ccommunication:
> > - application level security
> > - ACID transactions
> > - conversations and message correlation
> > - reliable message transfer
> > to name just a few.
> >
> > The Web architecture addresses a number of other core requirements:
> > - pervasive communications protocols
> > - universal naming
> > - platform- and language independence
> > - etc
> > which is why we are developing an application-to-application distributed
> > computing architecture on the Web.
> >
> > Anne
>
> MB
> --
> Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
> Ottawa, Ontario, CANADA.               distobj@acm.org
> http://www.markbaker.ca        http://www.idokorro.com
>
Received on Tuesday, 28 May 2002 07:26:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:00 GMT