Re: Late binding

On Tue, Jun 25, 2002 at 02:09:55PM -0700, Henrik Frystyk Nielsen wrote:
> Just to be precise, I think the benefit your are pointing to of HTTP's
> methods (GET in particular) is that they are universally supported and
> centrally described as compared to some WSDL document describing
> operations that may be defined in a decentralized and local manner.

Yes, exactly.

> I don't think either is in conflict with the Web - central as well as
> decentral control are both fundamental concepts of the Web architecture.

Very true, but I think you'd also agree that some parts are better
centralized, and other parts better decentralized.

> The problem in HTTP is that it is very difficult to go from local to
> global definition which makes introduction of new features very hard.
> One of the purposes of SOAP is to bridge the gab between centrally
> defined features and decentralized features through the use of URIs.

Right, mandatory extensions rule.

> This doesn't make the need for universally adopted features go away,
> however.

Definitely.  Especially for universally adopted semantics, since you
need to have something from which to be able to bootstrap communication,
before you can even discuss extensions.

If I open a TCP connection to your server, we're not going to be able to
accomplish anything.  Ditto for if I open a TCP connection over which we
agree that SOAP will flow (without any other agreement).  But with
agreed upon application semantics, we can coordinate whatever those
semantics let us coordinate.

(I know you know this 8-)

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, 25 June 2002 20:25:11 UTC