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

RE: Clarification on REST, GET and CGI

From: David Orchard <dorchard@bea.com>
Date: Wed, 24 Jul 2002 15:45:56 -0700
To: "'Paul Prescod'" <paul@prescod.net>
Cc: "'Mark Baker'" <distobj@acm.org>, <www-ws-arch@w3.org>, <rest-discuss@yahoogroups.com>, <fielding@apache.org>
Message-ID: <04ac01c23363$e0c040f0$0100007f@beasys.com>

Paul,

You didn't refute my point at all, which is that REST is not used
appropriately in many ways on the web but we don't document all the
inappropriate usages in the web architecture document and Roy's thesis.  We
still can talk about the architectural principles that the web embodies, as
described by REST.  Mark is proposing that the w3c web services group should
only document those architectures that have running code, and it should
document all the usages.  And he's specifically pointing out areas where
SOAP is used in different ways that he finds "sucky".  If the criteria is
running code, I would argue that a web architecture (aka REST) description
would include all the inappropriate uses.  And the "running code criteria"
precludes things like SOAP 1.2 which don't have much running code yet.  So
the running code criteria doesn't seem useful to specify architectures to
help people build better web/web services systems and standards.

So it's my opinion that the w3c web services architecture group should spend
some of it's time working on web services architectural principles, not
simply documenting what current systems do.  And the same goes for the TAG
wrt web principles.

Just because I mention REST usage (the generic method usage) and CGI usage
does not mean I want a REST vs whatever discussion.  In fact, I'm utterly
and completely exhausted by that particular topic area.  Roy's quote is the
reason why I mentioned CGI and I think his quote supports the notion that
some running code "sucks".  And I again assert that w3c groups shouldn't
document the sucky usages in our architecture documents, except perhaps as
being sucky.  But again again again, this is not a REST discussion per se.

I don't intend to follow through much more with this discussion about what
the group is going to do.  I'm more interested in writing up text to
describe the architecture and discussing the text of such.  "Do stuff rather
than talk about doing stuff".

Cheers,
Dave

> -----Original Message-----
> From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On
> Behalf Of Paul Prescod
> Sent: Wednesday, July 24, 2002 10:31 AM
> To: David Orchard
> Cc: 'Mark Baker'; www-ws-arch@w3.org; rest-discuss@yahoogroups.com;
> fielding@apache.org
> Subject: Clarification on REST, GET and CGI
>
>
>
> David Orchard wrote:
> >
> > ....  To compare, if I were to document REST as it is
> > done today, I'd probably ignore HTTP PUT/DELETE and I'd say
> that GET/POST
> > can be used interchangeably except for some bookmarking
> applications where
> > GET is a little better.
>
> That's a severe misrepresentation of the way the Web is used
> today. GET
> is used whenever safe references are important. Bookmarking is a tiny
> example. I can send you this link because of GET:
>
> http://search.bea.com/query.html?qt=REST
>
> I didn't bookmark it and probably you won't either. That fact
> that I can
> email it to you -- communicate it to you -- is what is important. You
> can interoperably, using any Web client, anywhere in the world,
> dereference that link. Bookmarking is a side effect of that general
> improvement in interoperability. I am confident that BEA's
> web designers
> understand this issue because they work with these safe information
> references ("hyperlinks") all day.
>
> > ... And I'd say that CGIs are great.
>
> CGIs *can be* great. I may be wrong, but I think that you're
> interpreting something from this mail by Roy Fielding. I've seen this
> post misinterpreted (IMO) in other places so I'm going to address it
> once and for all.
>
>  * http://lists.w3.org/Archives/Public/www-tag/2002Apr/0181.html
>
> Note the last sentence: 'The fact of the matter is that
> most CGI scripts are not HTTP compliant.  Most CGI scripts, in fact,
> provide interfaces to applications that suck. The "G" stands for
> Gateway.
> What people should realize is not that "CGI scripts should be banned",
> but rather that if the CGI script is written such that it behaves
> like a proper Web interface, then it won't suck.'
>
> CGI is a *gateway* interface. You can either use it to just
> publish some
> functionality through a URL (tunnelling) or you can use it to build a
> proper gateway between your software and the Web: URIs for all
> resources, hyperlinks as the engine of application state, etc. Some of
> the most REST-ful applications in the world are CGI and many of the
> least are not.
> --
> Come discuss XML and REST web services at:
>   Open Source Conference: July 22-26, 2002, conferences.oreillynet.com
>   Extreme Markup: Aug 4-9, 2002,  www.extrememarkup.com/extreme/
>
>
Received on Wednesday, 24 July 2002 18:47:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:02 GMT