Monday, April 5, 1999 by Dave Winer.
With the Frontier 6.0 shipment done, I'm going to do something unusual, I'm taking a vacation! I leave tomorrow and will be back on April 20.
It's *conceivable* that there will be updates to the Scripting News home page while I'm traveling, I can update it thru a web browser, but don't count on it. This is my chance to get away. I never know what that means, every time it's different.
My main objectives for this period are to do a lot of body surfing in the Atlantic Ocean, and to read some non-serious stuff and visit with friends. I will have a couple of business meetings on the trip, but they're long-term things.
I'm at a resting point. Most of the technical goals we set out to achieve in the spring of 1996 are now achieved. Frontier is a now workgroup web server, a category that's going to be important, I think, in the next few years. Everyone else seems to have missed this, that there's a position between personal webservers and the honkers that can run Yahoo, Excite or Lycos.
Jakob Nielsen wrote about this subject today! In a very eloquent piece that I couldn't agree with more, he says: ""An intranet should have a single home page that integrates a directory hierarchy, search, and news. Most intranets are chaotic, under-funded, and lack design standards, causing huge losses in employee productivity."
I think Jakob has been reading my mind!
All the necessary components are there. Robust HTTP client/server, a full content management system, Edit this Page functionality, a solid network computing model, editing and development tools, and a high level programming interface. Remember, this is scripting, not low-level stuff.
My friend Marc Canter, a very wise man, shared some of his thoughts on this at his Passover seder on Wednesday night. "Now you're writing an operating system in a scripting language," he said. I had lost sight of that. It's true. We made a bet ten years ago that personal computers would get faster and bigger and stay cheap. The bet was that script code would perform like C code. It's a good time to reflect on that. We got there!
Everyone who's working in Frontier now, whether they know it or not, is using a new kind of operating system. It's reasonably platform-agnostic, it runs on Mac OS and Windows, but in a real sense, the services those systems provide to Frontier developers are commodities. A file system, interface to the screen and keyboard and mouse, a TCP stack. Most of the juice is supplied by scripts. Thousands of lines of script code are running every time Frontier does something; code that is unique, that does stuff that the lower-level OSes don't do.
Moore's Law will push us forward even further into uncharted territory. There will 8-CPU machines, then 16-CPU machines, and so on. RAM will continue to get cheaper. Megahertz and disk space will get cheaper too.
As that's going on, more web users will become web authors. Full-time net connections will become more commonplace. Our software will distribute, now running primarily on a content management workstation or server, and soon on the desktops of people who write for the web. Imagine a portal that not only aggregates the brain power of thousands of people, but also gives them powerful tools to connect with other people whose brains and experience are available over the web. That's our roadmap.
Today's users are web browser users. They usually are not using the same machine that the code is running on. When they are, this is the unusual Fractional Horsepower HTTP Server concept that's going to be very big, I think. Apps that can be moved off your desktop to a server and back again, whenever you want.
OK, that's the pre-vacation vision. There's no doubt that my perspective is clouded by eight months of intense development work. When I come back I will have the next steps outlined, and be ready to build some more Internet applications. That's where I want to work, for the forseeable future. In applications.
So, I hope while I'm gone that the lions of the Frontier community will dig into the sample apps we've created and explore some of the directions they suggest. Think in terms of guest databases that have code and templates, and other guest databases storing data and user profile information. Assume you know every person coming into your website. What can you do to make the experience more useful to them?
The best way for you to participate is to figure out ways for your users to participate. That's how we all get onto the growth curve. Think about the wishlists of your users. How many of them can you address now that the base software has grown so much more powerful?
Our vision is unlike that of any other company, rather than have a small group of invisible people behind an HTML interface, we have a large group of autonomous developers, some of whom use our tools, and others who don't (thanks to XML-RPC we work together). Many of them run their own servers, many are capable of producing excellent content and helping thousands of less technical people share their thoughts thru the web.
People often ask the question "Does Frontier scale?" My answer is "Unlike anything you've ever seen before." The old vision of a web server being monolithic and mysterious and hard to access will melt over the next few years. Now web servers are simple. People are complex. Having software meet the people is the real game in web space, that's the tricky problem.
When Netscape announced RSS I knew almost immediately that it was something we were waiting for. The idea of a distributed writing system is one we've been building around since the early days of AutoWeb and Clay Basket, it actually predates that, going all the way back to LBBS in the early 80s.
The importance of RSS is that it marked a moment when the market was ready to go that way, towards a distributed writing environment. We solve so many more problems than RSS, but many if not all of them fit neatly behind RSS. We see the web as a two-way medium that evolves. News, search and directory, for today, who knows what else in the future? You can see the beginnings of this on my.userland.com and newssearch.userland.com. Where ever it takes us, we're going there, and RSS is definitely on the path. If you do content, or content management software, I strongly encourage you to get up the curve on RSS and see if you can be compatible with my.userland.com.
When software is centralized, as we've seen in the Melissa problem, it encourages people to be irresponsible. They have an illusion of separation. In fact, we are all connected now. And our responsibility is connected too.
This is where the biggest juice is. More powerful distributed ways for people to connect at points of responsibility. It will allow more informed opinions. That's why I'm interested in the Internet. That's why we're developing the software we're developing.
I know that only one tenth of one percent of the people on the net today will understand what I'm talking about. But that's cool. Those are the people I want to talk with.
See you in a couple of weeks!