Re: Harvesting REST

Eric,

On Tue, Jul 16, 2002 at 07:44:52AM -0400, Newcomer, Eric wrote:
> My point was really that we need to be able to evaluate the Web services architecture according to use cases for Web services, not for hypertext.  Do you agree?

Sort of.  I think the group has been doing that so far, but as I said,
there's nothing in the usage scenario document that can't also be done
well with REST (of which hypertext is one part).

> I also did not suggest that the problems might not have multiple solutions -- all problems do.  I was suggesting that Web services use cases are different from hypertext use cases.  Do you agree?

No, I don't agree with that.  Use cases are use cases, and don't by
themselves suggest a solution, only a problem.  As you've said,
OMA/CORBA, REST, etc.. can all be applied to that problem.  Our job is
determine which architectural style best meets our requirements while
doing so.

> I think it's an important point because I don't want us evaluating Web services architecture against use cases for hypertext (unless, of course, they turn out to be the same thing, which currently they are not). 

I guess I'm unclear on what your view of hypertext is, because IMO, they
are currently the same thing.  That is, our usage scenarios can be
adequately addressed using it as part of REST.

FWIW, "hypertext" just refers to how applications change state in REST.
i.e. rather than requiring getStockQuote() to be invoked to get you to
the "knows the stock quote" state - which either a) requires a priori
knowledge of the remote interface, or b) requires downloadable code
which violates the principle of least power - you just embed a link to
the quote as part of the previous chunk of transferred state, which
implies that GET can be invoked on it.

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, 16 July 2002 11:18:47 UTC