SAML, which I had not heard about before today, was much-discussed at the Catalyst conference.
Borland: "Kylix Open Edition is the version of Kylix specifically designed for open-source development and for building applications that are distributed under the GPL."
Paul Boutin is a senior editor at Wired magazine with a Manila weblog.
Dan Gillmor spotted Mitch Kapor at the Open Source Convention.
Evhead: "A dozen more SirCam emails today. I think the only way this thing is going to stop is if we create a disinfectant virus and release it the same way."
Peter Gabriel: "Real is anything you see." (With apologies for mangling the quote. Gabriel doesn't actually say that. But that's what I heard.)
Good morning, rise and shine!
Today's probably going to be a pretty light day on Scripting News. I'm giving a 25 minute talk on the state of distributed computing at The Burton Group Catalyst conference. That's a pretty awesome task, but then I realize it isn't.
First, I'll explain why I wanted to have an open distributed computing protocol in 1998. Here's how that goes. We had just finished porting our scripting environment from Mac to Windows. Frontier and networking was always a big deal, but until we had a Windows version we could just use Apple's networking, and we did. We could have used Microsoft's networking, DCOM, but that would have had several disadvantages. First, we didn't know DCOM. Second that wouldn't help if you had a mixed network of Macs and Windows machines, and third, it was against our religion. By that time we had learned the lesson of the locked trunk very well. In 1998 we were Internet developers. The platform with no platform vendor.
Like many other developers, in early 1998, I was playing around with XML, trying different ideas out and seeing what would stick. I was also very happy with HTTP, such a simple way to move information from one machine to another, so it was only natural to combine the two to create a distributed computing protocol. To make a long story short, I created a protocol, published an essay, Microsoft asked if we wanted to work with them, I said yes (of course) and the result was XML-RPC. We continued to work because they wanted more than that, and then came SOAP. Both protocols are in wide use now.
This is a good time to review the state of distributed computing because it's not totally messed up yet. The Internet favors simple low-tech protocols and formats. The BigCos prefer (for whatever reason) very rich and constantly evolving protocols and formats. To me, distributed computing is all about choice and no lock-in. As the CEO of a SmallCo, I can afford to say this. I win if lots of other people win. To me, distributed computing is about apps that are implemented on Linux, Windows, Mac, Palm, and PodunkOS. Small and diverse is good. Use the best tool for each part of the job. We offer far more freedom to the user than any other way of doing it.
To me, the key philosophy is inclusion. Remarkably, XML-RPC can work in every environment we've tried to make it work in. That is not true of SOAP, which has constraints that cannot be met by every scripting environment. And it is most definitely not true of SOAP plus WSDL. It's not much of a surprise, perhaps, that WSDL steers you into Microsoft's and IBM's favored environments. Microsoft and IBM developed WSDL behind closed doors and didn't include independent developers in the design process. This is why, to me, the distributed computing stack stops at SOAP. That's enough.
For details see What is XML-RPC? in the foreword to O'Reilly's XML-RPC book.
© Copyright 1997-2005 Dave Winer. The picture at the top of the page may change from time to time. Previous graphics are archived.