[
Lists Home |
Date Index |
Thread Index
]
Chiusano Joseph wrote:
>I agree that it is quite easy and tempting to consider Web Services as
>being represented by a single "block" of standards, but I believe it is
>more useful to think of it in various "waves" (or phases), and to talk
>about the success or failure of those waves. I assert that there are
>current 4 "Web Services Waves" in sight - one of which is already behind
>us (but still ongoing for future enhancements), and one of which we're
>in the midst of now. These are:
>
>(1) Core Specifications
>
>- Your basic SOAP and WSDL, whose early versions are mature and
>ubiquitous, and whose specification is still ongoing;
>
>- We have completed the initial "phase" of this Wave;
>
Perhaps so, but it's been more like a series of intersecting and
competing ripples. The "mature and ubiquitous" SOAP and WSDL are still
in an extraordinarily confused state after several years of development.
After starting off down the path of rpc/enc in SOAP 1.1, which basically
provided an XML representation for simple data object graphs, this
approach has now effectively been dropped in favor of doc/lit, which
uses a schema description of data and leaves the interpretation of that
data (as objects or whatever) to the applications. However, WS-I Basic
Profile apparently couldn't swing enough support to *only* support
doc/lit, so we're left with the alternative of rpc/lit also a part of
the profile (neither one required, both allowed). This is not currently
used by much of anybody, but at least some implementations (JAX-RPC
included) plan to add this just because it's allowed by the profile.
Meanwhile, even such a basic (and necessary, for real-world services)
feature as attachment handling is left without any single specification.
We've got the Microsoft-proposed (but then dropped) SOAP with
Attachments (SwA) standard as the basis of WS-I Attachments - but
Microsoft doesn't support it, and AFAIK has no plans to do so. Then
we've got the Microsoft-proposed (but then dropped) Direct Internet
Message Encapsulation (DIME) proposal. This is supported by one
Microsoft add-on, and by several other frameworks just so they can
exchange attachments with Microsoft code, but AFAIK Microsoft has no
plans to support it in the future. Instead we've got the spiffy new SOAP
Message Transmission Optimization Mechanism (MTOM,
http://www.w3.org/TR/2004/WD-soap12-mtom-20040209/) apparently based on
yet another Microsoft proposal
(http://www.gotdotnet.com/team/jeffsch/paswa/paswa61.html) that throws
away all concepts of layering in favor of having everything as part of
the Infoset and leaving it up to the transport to sort out how to send
each component. Gee, that sounds like fun - layered protocols are such a
20th century concept, after all.
Given the ongoing confusion and incompatibilities at the level of
basics, it looks like Web services have a very long ways to go before
the current "wave" of WS-* proposals really gain relevance. The one
exception may be WS-Security, if it does get approved by OASIS and does
get supported by all the players. Meanwhile, we've got yet another layer
of hype being added to the top of the Web services heap with the latest
"SOA" buzz acronym... Ah well, at least it looks like the market for Web
services training isn't likely to dry up anytime soon.
- Dennis
--
Dennis M. Sosnoski
Enterprise Java, XML, and Web Services
Training and Consulting
http://www.sosnoski.com
Redmond, WA 425.885.7197
|
