The Sky is Falling!

Tim Bray's post on the Loyal WS-Opposition seems to have hit a nerve. The RESTafarian extreemists have become emboldened by such unpatriotic speech! They cackle with glee over the fact that some have realized that there is little semantic difference between the Get in WS-MEX and that in WS-Transfer. Such treasonous thoughts and mixed messaging are only giving aide and comfort to the enemy! Where is WS-Zell when you need him most?!

While I too lament the over-complexity of Web services, and have sympathy for those who are overwhelmed, I would only point out that the IETF has published 3913 RFCs to date, each of which started life as a series of obscure Internet Drafts, many if not most of which fell by the wayside before ever reaching the exalted status of Standards Track RFC. Just think about how many thousands of ID's have been published over the years.

Run for your lives! The Internet is drowning in a sea of complexity!

Granted, many of the RFCs obsolete others, but that still leaves us with roughly an order of magnitude more specs than have been produced thus far for Web services. And talk about your complexity... puhleeeze!

RFC2616 (HTTP1.1) cites no fewer that 49 references, 41 of which are other IETF RFCs. There are 25 IETF RFCs that have "HTTP" in their title which means that they either define the protocol or add extensions or provide clarifications. And here you thought that HTTP was such a simple protocol. Apparently, someone felt that the additional specs were needed.

And yet, my mother can send me email, though she has never read RFC821 or RFC822. She can surf the web, oblivious to RFC2616 or RFC2396. For that matter, how many developers have actually bothered to read RFC2616 or RFC2396?

There are a grand total of 50 WS-* specs listed here, although the list needs to be updated to reflect the recently publish MSFT specs. As has been pointed out by some, there are competing initiatives vying for the brass ring of broad, interoperable adoption for the same functional domain. Eventually, some will fade away into obscurity. Some may only be relevant in a few edge use cases, but that does not make them any less valid.

Should the Web services community be striving towards simplicity? Damn straight.

Should Web services be better leveraging the Web as it exists? Darn tootin'!

Is it true that HTTP solves all of the world's problems? Hardly. ("When I accept POST", said Humpty Dumpty, "it means just what I choose it to mean, neither more nor less.").

Is the sky falling? No.

Should we throw in the towel? No.

Should average developers be concerned with every new WS-* spec as soon as it is first published? Not unless they are masochists.

Do average developers worry about every new ID published? Hardly. Most have probably never even heard of the IETF.

Should average developers and IT managers be concerned with the WS-I Profiles? You bet, because the Profiles define the set of broadly adopted and widely implemented specifications, and that is what matters to developers.

Why then is it that every time a new WS-* spec is published, the blogosphere blows another gasket?

Get a grip.