Inspired by this article at Read/Write Web, I decided to make a search engine out of one of my bookmark collections. Discover Semantic searches about 70 sites dealing with the semantic web (aka "web 3.0"). It covers W3 recommendations, specifications, tools, tutorials, and dozens of blogs about semweb.

If nothing else, it was interesting to see what can be done with the CSEs. I also used Microsoft Expression Web 2 Beta for the coding. Unlike the first version of Expression Web, this one's got PHP intellisense (language library and code completion) and the built-in web server can run PHP. It's a lot faster than Eclipse, which really doesn't like the size of my Symfony projects these days.

I've been thinking about Twine a lot recently. I see a lot of potential in the semantic web and I think a lot of people will be thinking that way once Twine is here. Smarter systems that start learning from our siloed information all over the web and on our desktop, then start making intelligent connections for us based on that learning, are going to change how we work. They're going to start doing the connecting we've always had to do ourselves, between our own information and relationships, and between what we know and what 's related that we have yet to discover.

The concept of intelligent systems, software that learns, applications that become more intelligent and more useful the more they're used and scale, is really appealing to me. Thinking about it today inspired a product idea I'd really like to build. I don't know why I didn't think of it sooner, since the basis is similar to something I programmed way back in 2003 between the end of high school and my freshman year of college. It was a learning algorithm that determined who a sales lead in an SFA system should be routed to based on sales representatives' past experience with similar types of leads. It worked pretty well, despite my lack of experience leading to a very inefficient implementation; the company that had contracted me to develop the product applied for a patent on that part of the system.

It's been a long time since I've had an idea for something that's not been done a thousand times already. I think I'll work on this one after I finish the work on W3Counter.