Lists Home |
Date Index |
That's a good answer, Dave, and I certainly agree that the
history of the Internet is a forked history. Many forked
tongues and all. :-) But I'm not sure this history helps
me if I were faced with two different offerings for web
services, one based on RPC-style transactions and one
based on the web architecture. As for UDDI and WSDL,
the toolkits are there so there may be takers and given
the WSIO, maybe more than a few since so many of us are
developers of applications over the frameworks the core
technologists provide and the WSIO folks are ostensibly
not arguing over the baseline and global specs but
hammering out the fine points of interoperability. The
outcome to be determined.
So maybe my choice is made for me and all I get to choose
(in theory, at least) is the vendor of the toolkit.
Many thanks for the reply.
From: Dave Winer [mailto:email@example.com]
OK, I'll try to answer this question, with an analogy.
This week there are hundreds of thousands of people in Utah.
How did they get there?
They came from a lot of different places, but they're all there in Utah now.
Some are watching the Olympics, some are participating, some are covering
it, some are serving hot dogs and beer, some people live there, and some
people have no idea the Olympics are going on at all.
So how did we all get to be on the Web? Similar story.
I can tell you how I got here. I was doing scripting software that turned
into publishing software.
We were already building networks, using Apple Events, which is totally RPC.
So when we ported to Windows in 1996-98, we still needed to network.
Couldn't use Windows networking because that didn't work on the Mac, and
Further, by then, we saw ourselves as Internet developers, largely because
it didn't work well being controlled by the platform vendors.
That led us to XML over HTTP.
We already had systems and practices developed, we were just looking for a
way to encode and transport what we were already doing, but intsead of using
Apple's networking or Microsoft's we wanted to use the Internet.
We had to learn HTTP and XML. Perhaps if we came from somewhere else, we
would have come up with a different approach.
A big thing happened in 1993, when what remained of the desktop software
world, which I was a part of, merged with the Unix world. And of course this
was just a loop, because many of us came from Unix in the first place before
going to desktop computers. I think people often overlook this when
recounting the history of the Internet, as if it were just one thread, when
it was really a joining of many threads, and forking, and re-joining.
About UDDI and WSDL, I am hardly a defender of those bits. I don't think we
need them, and if we did, they're far too complex to gain traction. SOAP
itself can be massaged into a profile that's understandable, so we like
SOAP, but I think the disconnect betw the REST view and the RPC view is that
they come from different places, different points of view, different
But we're all one Internet now, no view is more right than the other, imho.