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

RE: FW: draft findings on Unsafe Methods (whenToUseGet-7)

From: Anne Thomas Manes <anne@manes.net>
Date: Wed, 24 Apr 2002 00:28:43 -0400
To: "Mark Baker" <distobj@acm.org>
Cc: <www-tag@w3.org>
Message-ID: <CJEIKEMEBAONGDDNLEKFEECIECAA.anne@manes.net>
If you look back at all of my postings, I've consistently recommended that
we explore new and better ways to do web services, but not at the expense of
the current set of technologies. My customer base (12,000+ registered users)
loves SOAP -- very possibly because the RPC/Messaging semantics are familiar
and comfortable. If the "Web architecture" is such a *very* difficult thing
to understand, perhaps that's why most users don't feel particularly
concerned that SOAP doesn't conform. Things that are *very* difficult to
understand generally don't achieve wide adoption. On the other hand, SOAP
has been experiencing extraordinary success (even though Mark refuses to
acknowledge it). Even so, SOAP 1.1 is flawed (incomplete, imprecise, etc. --
not "bad technology"), and it needs revision. We need SOAP 1.2.

The reason I spoke up at this time is that it appeared apparent to me that
this Unsafe Methods finding would provide a convenient excuse to allow the
TAG to challenge the release of SOAP 1.2 as not "web-friendly". My immediate
goal (speaking for my customers, not for my *small* company marketing
machine) is to improve the current Web Services architecture -- to improve
SOAP and WSDL, and to define standard SOAP extensions to support
interoperable security, routing, attachments, and conversations.

Not that it matters, but I've never been overly found of the name "Web
services", because Web services (based on the existing Web services
architecture) aren't constrained by Web technologies. A more accurate name
would be "network services" or "XML application services". A Web service can
just as effectively communicate over MQSeries, Jabber, or WAP. Although an
HTTP binding is required, SOAP uses HTTP strictly as a transport (not as an
application protocol), and SOAP is inherently transport-independent. (Hence
our dispute.)There are only two Web-oriented requirements in the current Web
services architecture: a Web service is identified by a URI, and it uses XML
messaging. But Microsoft named them "Web services", IBM endorsed the name,
and that was that. The name stuck, for better or worse.

I have a question for Roy:
If you believe that SOAP is such "bad technology", why do you condone not
one, but two, SOAP projects at Apache?

And I have a question for the TAG:
If so many members of the TAG viewed SOAP 1.1 as such "bad technology", why
did it launch a Working Group with such a restrictive charter that mandated
that XML Protocol be based on SOAP 1.1? Based on this new Web Architecture
now being defined, I would say that the XML Protocol charter is
self-contradictory.

I encourage Mark, et al, to develop technologies and products that implement
Web services in a Web-friendly way, and market them to the public. I advise
you to figure out a way to explain the Web architecture so that the average
Joe can grok it and appreciate why it is so much better than SOAP. Then,
perhaps, this new approach to Web services will gain comparable public
adoption as SOAP has to date.

But first let's release SOAP 1.2. Either that, or let's be upfront and let
the members know that W3C is not the right venue to advance this technology.
I'm sure we can take our work someplace else.

Best regards,
Anne

> -----Original Message-----
> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org]On Behalf Of
> Mark Baker
> Sent: Tuesday, April 23, 2002 9:36 PM
> To: David Orchard
> Cc: www-tag@w3.org
> Subject: Re: FW: draft findings on Unsafe Methods (whenToUseGet-7)
>
>
> I'm so glad that Roy had the guts to speak up as he did.  I've been
> saying basically the same thing in the XML Protocol WG since I joined
> when it started.  I've also done my part to ensure that SOAP can be
> used in a Web architecture friendly way, but given that I've only seen
> two companies use it this way (my company, and KnowNow), I don't have
> much hope that SOAP will see more success than CORBA did on the
> Internet (read; "little to none").
>
> On Tue, Apr 23, 2002 at 04:14:49PM -0700, David Orchard wrote:
> > The reason that SOAP remains and grows at the W3C is because it
> and related
> > specifications are heavy and responsible users of web machinery,
>
> This is demonstrably not true; they are not responsible users, because
> they fail to work within the constraints imposed by the architecture
> (as commonly used).
>
> Web architecture is a *very* difficult thing to understand.  For most
> people whose exposure to distributed systems has been limited to RPC and
> messaging, it is unlike both, and so it's no surprise at all that so
> many people have gravitated to Web services; they're familiar.
>
> IMO, the closest "popular" non-Web architectural style to the Web's
> would be a tuple space, such as is seen in Linda or JavaSpaces.  But
> it's more general than even tuple spaces.  And obviously more
> successful.
>
> > I find it interesting that you feel comfortable being part of
> the TAG at the
> > W3C, which was voted on by the W3C Members, but you are
> uncomfortable with
> > Web Services as-is at the W3C, also voted on by W3C Members.
> And if the Web
> > Services folks don't do what you think is right, they should just all go
> > somewhere else regardless of the process that got us where we are.  I'd
> > observe that in the TAG elections, 3 of the 5 elected members
> are members of
> > the SOAP WG.  I interpret this as yet another plank in the
> mandate that web
> > services at the W3C is important to the membership.
>
> Without a doubt, Web services are important to the membership.  Nobody
> would argue with that.  The question being asked is, are Web services
> important to the Web?  Do they help lead the Web to its full potential?
> Roy, myself, and everybody else I've talked to who understands Web
> architecture, agrees that they don't.
>
> But let me be clear about my position.  The *goal* of Web services is an
> admirable one, and I fully support what people are trying to accomplish
> with them.  The only issue is, *HOW* do we best use the Web to achieve
> these goals?  All goals *are* achievable, just in a different way than
> Web services proponents advocate.  To demonstrate that this isn't an
> academic exercise, as some have suggested, my company sells software
> that incorporates many of my ideas in this space.
>
> So, how do we proceed from here?  IMO, we need some Web services
> proponents to have an open mind and admit that they might be wrong
> (this is how I learned it).  Then we need those who do understand Web
> architecture to give them a crash course.  I would be happy to help out
> here.
>
> MB
> --
> Mark Baker, Chief Science Officer, Planetfred, Inc.
> Ottawa, Ontario, CANADA.      mbaker@planetfred.com
> http://www.markbaker.ca   http://www.planetfred.com
>
Received on Wednesday, 24 April 2002 00:27:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:06 GMT