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

RE: Does RM make a qualitative difference?

From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
Date: Wed, 18 Dec 2002 10:49:24 -0500
Message-ID: <9A4FC925410C024792B85198DF1E97E4049BD03D@usmsg03.sagus.com>
To: www-ws-arch@w3.org

> -----Original Message-----
> From: Mark Baker [mailto:distobj@acm.org]
> Sent: Wednesday, December 18, 2002 10:30 AM
> To: Christopher B Ferris
> Cc: www-ws-arch@w3.org
> Subject: Re: Does RM make a qualitative difference?
> I respect the choice of the WG to defer this issue (as much as it
> disappoints me), but let's call a spade a spade; the WG currently
> rejects the uniform interface constraint.

I think it's fair to say that we don't accept it as a requirement, and we as
a group remain unpersuaded that it is a UNIVERSAL architectural principle.
We can see benefits as well as costs, and hope to clarify the scenarios
under which the benefits outweigh the costs (e.g., when providers and
consumers don't know anything about each other and they want to bootstrap a
relationship).  IMHO it is horrible overkill in an environment where the
parties do have a well-established relationship and a fast/reliable network
> What's emphasized there?  I can't see it.  Presumably it's one of the
> last two sentences, but I can't see how any of those would give me
> pause.  

For the Outlook-blissfully-oblivious, Chris emphasized 

"The trade-off, though, is that a uniform interface degrades efficiency,
since information is transferred in a standardized form rather than one
which is specific to an application's needs. The REST interface is designed
to be efficient for large-grain hypermedia data transfer, optimizing for the
common case of the Web, but resulting in an interface that is not optimal
for other forms of architectural interaction."

> It's not like Web services are a replacement for telnet or the
> Quake protocol, which would be examples of different forms of
> architectural interaction.

Hmm, at the risk of displaying my apostasy towards RESTifarianism, I *would*
think that programs communicating fine-grained structured data over an
intranet, optimizing for the common case of B2B/EAI or RPC, is a different
form of architectural interaction than coarse-grained hypermedia over the
Web. One *can* do this, of course, but we are not going to say that people
*must* do it.

I think the quote from Roy Fielding was very apt for our purposes:  it
emphasizes that this is all about optimzation and tradeoffs, not fundamental
Received on Wednesday, 18 December 2002 10:49:31 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:41:01 UTC