The Dog Ate My Homework
Tuesday, March 17, 1998 by Dave Winer.
This is so weird.
Eighteen years ago I had root canal work. I remember it being excruciating. The work went on and on, but finally it was over. I forgot it. Yes, I had two silver teeth. But they barely got one tenth of one percent of my consciousness.
Until last night!
I was eating pizza with a friend, getting up to get another slice, when I felt this foreign substance in my mouth. Hmmm. At first I thought it might be something impure about the pizza. I took the impurity out of my mouth. It looks like a tooth! Oooops. A piece of my mouth was in my hand.
I wondered why it didn't hurt. But this morning it does. Quite a bit...
I had planned a trip to NY to speak at Seybold, to meet with friends and explore new business opporunities. The trip is cancelled. My apologies to the people who were coming to the DaveNet Live session tomorrow night. It won't happen.
XML XML XML. It doesn't quite have the ring of Marcia Marcia Marcia or Java Java Java, but like Marcia and Java, XML does have substance; but there's a lot of extra hype and confusion.
XML is substantial, even more so, in my opinion, than Java. But the airshows get in the way. The hype balloons. Isn't this easy? they say. OK. But why?
The substance of XML is about two things, rendering-independence of content and as a format for inter-network plumbing. I'll also talk more about the hype.
First, XML has a strong place as a rendering-independent way of specifying structured information. This is a good idea. It's scalable content. Stuff that's specified in XML can evolve to the next rendering method. If all the multimedia CD-ROMs of the early 90s had been written in XML (it didn't exist then) they could have been re-rendered for the web, and whatever comes after the web.
This is the SGML pitch. When people say that XML is like SGML, this is what they mean.
It's a real benefit, not hype. Scalable content is especially important for the Seybold crowd, people and companies who invest big bucks in content. If you want that investment to pay off, no matter what comes down the road, it pays to systematize the development and management of the content.
My good friend Marc Canter is the constant evangelist of scalable content. He made me aware of the concept, and continues to teach me about it. Marc is an idealist, but listen to him, he's learned the hard way, and he practices what he preaches.
If you're in the content business, at any level, in any way, and you're not thinking about scalable content, read Marc's piece, or you'll get a big I Told You So in a couple of years.
I'm a systems guy. I love scalable content, but I live for plumbing.
My vision is hundreds of thousands of servers running compatible software, delivering beautiful HTML and DHTML pages over HTTP, with great workflow systems running behind the server, allowing creative people of all flavors to work together effortlessly, using the tools they like and understand, and have it all assembled into a constantly-changing presence.
I want the authors to feel like they're in touch with their audience, but to have a clear line between authorship and readership. I want everyone who creates value to make money. I want to compete and cooperate with other companies who deliver pieces of this puzzle.
To make that vision real, we need good plumbing. We need to be able to connect the authoring systems with the content management systems and the web servers. There are already many plumbing systems, but we could use a new one that supersets and simplifies all the existing protocols.
The promise of XML as a plumbing level is that now, in 1998, we can sort out the differences, allow a new kind of server to be created, and get ready for the next level of growth in net-based content.
It's happening, but not on the podiums of the new XML circuit. What's happening on the podiums? Posturing and positioning. The same kind of stuff that brought us Push last year. The hype is superficial. If it gives you comfort, or makes you rich, go for it.
But in the end the hype will subside, as it did with Push, and the companies that win will be the ones with deep and efficient workflow systems and substantial content tools that creative people like to use.
Please pay attention to who's *not* on the XML hype circuit.
I find it fascinating and encouraging that Marimba, the leader of the Push push is busy creating systems and selling software, and not latching their presence to the XML hype parade. That's cool!
I lost a few friends when I went anti-Push in early 1997. But I think I have been vindicated. The web is alive and well. HTML is still the way of the world, and it will be in the future too.
But Push has a place. If we had been more realistic about it in 1997, we wouldn't feel burned by it now. Let's learn from the experience. Don't go ga-ga over the XML promises. Let's wait for delivery. It's more interesting.
I'd like to offer my thanks and extend an olive branch to Marimba. They will continue to play an important role in the new plumbing system of the net. Our software will be compatible. We'll work together, that's the beauty of standards, that's what's beautiful about XML.
Life sure is funny!
I have a message for three companies, Adobe, NetObjects and Macromedia, and other vendors of content tools.
WebDAV. It's another XML-based standard, being developed thru the IETF, largely by Microsoft, that defines how work flows thru content management systems.
It's got Big Mo, in my opinion. It solves the problem. Proprietary workgroup software will fade as WebDAV gains traction. Open systems are the way of the future. The big money is going that way.
Tools makers like Adobe, NetObjects and Macromedia can connect to servers in an open way. WebDAV is the way.
Now I get wistful.
In my breast beats the heart of two Jewish men, Rube Goldberg and Albert Einstein. I love cobbled-together systems that magically work and I'm seeking a unified theory that allows systems to connect effortlessly.
I live with reality and I reach for nirvana. That's me! The plumbing is interesting to me, just because it's plumbing. Connections. Client-server is the way of the past. Servers everywhere. Huge virtual web servers. Giant distributed servers! Workflow. That's the way of the future.