XML Databases
Sunday, February 22, 1998 by Dave Winer.
Good morning and welcome back to DaveNet!
It's been a long time, but I think it seems longer to me than it probably would to a reader.
An example. A friend returned from a three-week trip from Bali, called me on Friday and was buzzing about the pieces I wrote about Karla Faye and Linux, weeks ago.
She just read the pieces. She talks about them in the present tense. To me they're ancient history! I've been thru so much in the last few weeks, so little of it has to do with the old struggles. A sense of time emerges. Loops! Interesting.
Anyway, I've learned a lot about management in the last few weeks. Some of these are old lessons. Here they are.
Conflict 
You can't avoid conflict. When a group of people wants to work together, no matter how much, conflicts arise.
People make assumptions, invent theories, fall into patterns, make accusations, tell others to shut up, people stay silent and the patterns can turn into dysfunctional ruts. I'm not saying I don't do it too. I do.
After the shipment of Frontier 5, I chose to go thru this. To challenge the assumptions, to confront the problems, to renegotiate our deal with the people we work with. It was time to do it. A lot of energy gets released when you ship a new version of a software product. All the questions that had turned into ruts came up, and this time I didn't run away.
I asked direct questions. Are we a community, and if so, what does the community do? Do we have developers, and if so, what is their commitment to their users? What level of professionalism do we aspire to? What is UserLand's responsibility? Our developers want independence, so do we.
I think we did really well. We decided we are a community and we have developers. I think it's clear now, after lots of heated email, much of it public, that we can work with other people who are gathered around our software. We value professionalism, and keep a respectful distance between us so we can work together more effectively.
People are working much better, from my perspective; I hope from theirs too. It's an old lesson, if you want to plant a gorgeous garden, you have be willing to dig out the weeds.
Free software
In a different world, there's a lot of energy gathered around free software, with the Linux operating system at its center. I get a few messages from the energy. These people are mostly young. They've been left out of the world that Microsoft carved out thru competition with Apple, Novell, Sun, Oracle, Netscape, IBM, etc.
Perhaps they don't see a way to win by following the economic structures of previous generations. I think this is a failure on my generation's part. Can we afford to include the entrepreneurial independence of the generation that's now coming of age? If I were in my early 20s now, I probably would be part of this group of people.
But I'm not in my 20s. I have a lot invested in the idea of customers who pay money for software and programmers who get paid to develop it. I'm being pulled firmly further into commercial software. The number one feature request of my users is that we start charging money for our software so they can feel comfortable building careers on it.
I'm being pulled. I want to build a team of business development, sales, marketing and support people for my users. The users want it and they deserve it. Same with our developers. And our investors.
So, in my world it's time to go commercial.
Now back to the other world...
What if you're young and the only role the world offers you is to take a job working on someone else's software? Do you have to be a Microsoft developer to make software? What do you work on in your spare time? These are the questions they raise for me.
Creativity is something to nurture and encourage and channel, and there's not enough of that in the economics of the current commercial software industry. So they invent a world without the old rules. I think we forced them to do this.
Somewhere in this cloud of energy there is a young version of Dan Bricklin or Bill Gates or Steve Jobs, I'm sure of it. They have what it takes to rock our world. Maybe not in 1998, but in a few years.
It's better to look into it now.
Learning Systems
Switching gears...
I've got an idea of nirvana in learning systems, now we just have to flesh it out and start implementing.
The ideal is a web page that's a simple form, like the ones you see on Infoseek or Excite or Alta Vista. Type in a phrase, get a list of pages.
The key is a smart human-authored index on the other side. Like Yahoo, but restricted to a single subject. It knows how to search its own database, and provide the most relevant information first.
Some sites, like www.scripting.com, have so much information to offer that a list of all the pages is overwhelming. People can't easily find the information they need.
Instead of a full screen of links, the only approach that works is an almost blank screen with a simple question.
What do you want to know?
We're going to try to make that work.
XML
XML is a bandwagon, and everyone is piling on. There will be tons of messages about XML coming from all corners. It will be hard to sort out.
I think I can offer some advice as you try to parse all the apparently conflicting opinions.
XML is a file format, a way to move information from one machine to another, from one piece of software to another.
The power of XML is the universal support it's gaining. In this case, the noise is the substance. Why? Because XML is so low-tech, so easy to implement, a company can commit to XML and deliver compatible software within weeks.
XML Databases
Further, I believe XML opens the door for a new generation of databases.
Relational databases will not go away, but a new kind of database will take root. Object databases, ones that can manage heterogeneous, non-tabular information. In other words, XML is more general than the relational form. Relational databases will only be able to manage a subset of XML.
My company makes an object database. It's at the core of Frontier. There are others, but we think we have the best one. Every developer is going to need an XML workbench and runtime, just as every Java developer needed an IDE and a VM and a JavaScript programmer needs an editor and a web browser.
The world is specializing. I think we found our niche in something hot. We're on Windows so people watch. Nice place to be.
A bigger picture -- we're going to need to work with other object database vendors to define interfaces that allow users to choose different databases without locking into any single vendor's product. Even in a niche, not all vendors will be able to do it all.
To the XML community
We've been doing a lot of work in our world. We're just learning about the other parts. XML is about breaking down the barriers. Let's all work together!
http://www.scripting.com/frontier5/xml/plan.html
Rain
Finally, the weather here has been spectacular.
One rainstorm after another. Bridges wash out. Creeks look like the Colorado River. Trees fall, mud slides, but the power has stayed on and the servers are up.
Today it's bright and gorgeous and sunny, and wet!
Let's have fun!
Dave Winer
PS: A reminder, you can respond privately to DaveNet pieces by putting Not For Publication at the beginning of your response. I read all responses personally, and select interesting or informative replies for publication on the Mail site.