RE: Late binding

Part of the trouble we're having here is that we are debating between what exists today in SOAP, and what Mark would propose, instead.

Mark has been arguing against SOAP for more than two years, as I recall, and yet the spec has gone ahead anyway.  I point this out because we need to take into account what exists today as much, if not more, than we need to think about what should exist.  

To quote my colleague Oisin Hurley: in theory, practice and theory are the same, but in practice they are not.  Let's not get caught up too much in theory.

Fundamentally I also think we are having difficulty arguing in the abstract about the best use of Web architecture.  I don't think anyone will dispute that current Web architecture is working well for the browser-based Web.  However, I think that what we may need to focus on is how Web services will use the Web differently, and thus determine the extent to which current Web architecture applies.

In other words, just because Web architecture works well for what it's used for today, doesn't mean it will work well for Web services.  Let's please focus on what exists today with repect to Web services (I mean can we all please at least agree to confine the debate to what's already in widespread use and how it might best evolve) and focus on the use cases for Web services (and debate how, if at all, principles of Web architecture as articulated by REST might apply to those use cases).

Eric

-----Original Message-----
From: Sam [mailto:bytecode@Phreaker.net]
Sent: Saturday, June 29, 2002 10:42 PM
To: Mark Baker
Cc: www-ws-arch@w3.org
Subject: Re: Late binding



> SOAP will only help you do that if you know what method to use.

Correct.

This is based on information derived from WSDL. In the current state
of the universe, business parters would have to negotiate business
contracts seperately (Eg like an SLA).. and I look at the WSDL as
an artifact of that contract.

So yes, there is "some prior information".

The whole scenario of dynamic discovery and usage of services 
working transparently together on the fly without any prior
knowledge or anything, is, in my opinion, quite improbabilistic.
The number of variables in that equation increases exponentially 
everyday :)

/s





Mark Baker wrote:
> 
> I'll take one last stab at this.  I've explained this several different
> ways, but I guess I haven't found a suitable way to convince anybody
> here.  That's quite unfortunate, because this is very important.
> 
> On Fri, Jun 28, 2002 at 11:23:00PM -0400, Sam wrote:
> > Based on the WSDL I know what the input and output for the portType
> > is.
> >
> > SOAP will help me send the input information across the wire and
> > get the output information for that service.
> 
> SOAP will only help you do that if you know what method to use.
> 
> I liken the way in which the typical use of Web services work to a
> guessing game.  They basically say, "pick a number, any number, and if
> it's the one I'm thinking of, I'll give you something in return".  In
> other words, unless you have prior information about what the number is,
> you won't get anything except "sorry, pick another number".
> 
> The Web, on the other hand, assigns the number 1 to the "give you
> something" operation, and everybody knows that, so there's no question
> what number you'd say when you want something.  This makes it very easy
> for two parties that have never met to exchange information, which is
> pretty important on the Internet.
> 
> MB
> --
> Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
> Ottawa, Ontario, CANADA.               distobj@acm.org
> http://www.markbaker.ca        http://www.idokorro.com

Received on Sunday, 30 June 2002 15:06:53 UTC