RE: REST wrap-up (was Re: Web Services Architecture Document

 

> -----Original Message-----
> From: Mark Baker [mailto:distobj@acm.org] 
> Sent: Thursday, January 29, 2004 2:29 PM
> To: www-ws-arch@w3.org
> Subject: REST wrap-up (was Re: Web Services Architecture Document
> 
> 
> 
> I just wanted to say that I quite liked the REST/Web material 
> that was there, especially the comparison text.  It's not 
> perfect, but good enough to, I think, make many readers 
> wonder why REST wasn't adopted as a base style for Web 
> services.  It would have been nice to have seen an answer for 
> *that* question; it used to be because REST was felt to 
> require humans in the loop, and while I know some still feel 
> that way, I wonder what the motivation is for others?  That 
> it doesn't matter; that regardless of technical merit, all 
> one needs is widespread support?
> i.e.  that REST might have worked, but we're doing SOA now, 
> so nah nah neeny nah? 8-)  I'm seriously curious.

It's really very simple: the purported technical merit of REST for the
problems that people are addressing with Web services remains speculative.
Of the perhaps 100 people who participated in the working group at one point
or another, I recall three who put forth the argument that REST is suitable
as the "base style for Web services."  Given W3C's consensus-driven decision
making procedures and the distribution of opinion on this subject, what
amazes most observers is how seriously REST is taken, not that is is not
enshrined as the True Path to Web services.

Considering how few people in the Web services world even considered REST of
any interest at all two years ago, I guess my closing thoughts on the
subject as Chair of the WG would be: Why the perpetual complaints?  Why not
celebrate the support for REST in SOAP 1.2, WSDL 2.0 [in progress], and the
WSA Note?   Why not encourage people to use this support where appropriate
rather than argue against the utility of the familiar Web services
technologies in the domains where they are appropriate? W3C has given vastly
more support to REST than a count of the member votes and dollars would
warrant. REST it is treated as at least as suitable an architectural style
as RPC in documents from the SOAP, WSDL, and WSA WGs even though there are
*vastly* more member companies, development tools, and well-known
applications supporting RPC-oriented Web services than there are for RESTful
Web services. 

Y'all need to stop shouting and start showing. There are dozens of vendors
with hundreds of tools, and thousands of whitepapers, demos, webinars, etc.
that give tangible reality to SOAP/WSDL/WS-splat approaches to real world
business integration problems. They may or may not be optimal or even
suitable for all that many real world problems, but they are out there for
the world to use and make its judgment about. RESTifarians can point to a
couple articles on XML.com, the Amazon raw XML over HTTP interface, a
bazillion email and weblog rants, and ..... ???? Go out and build the tools
that ordinary developers can use, publicize the case studies where REST was
successful in machine-machine interoperation settings, and trust the
evolutionary/market forces to reward those who bet on it.  When concrete
reality speaks, then people will start to listen.

Received on Friday, 30 January 2004 00:37:19 UTC