yelvington.com

At work this week we let another website escape from the zoo. Ruby Baboon is another take on the user-driven context engine concept: a metasite that pulls together lots of interesting stuff around a topic, all nominated and graded and ranked by users. It uses individual item nominations and RSS feed nominations from users, and pulls in related images and videos from Flickr and YouTube.

Ruby Baboon is all about celebrity gossip -- that addictive guilty-pleasure stuff that we all claim not to follow. Take a peek but don't let anybody see you do it. Use your home Internet connection. :-)

This doesn't reflect a strategic shift toward building national websites (or a fascination with celebrity scandal). It's really about what Scott Anthony talks about: Invest a little, learn a lot. Ruby Baboon and its sporty sibling FanaticZone Remix may become big worldwide hits but I am not holding my breath.

The real point is for us to launch a real-world test in a low-risk, low-cost setting, so that we can understand how these user-driven context engines work. You can't get there with a whiteboard and a bunch of meetings. Someday the concepts may find themselves in a local market, which is where our real focus continues to be.

Previously I wrote about six-week projects. This one took a little over a week, with much of the work done over the July 4 holiday by Nik Wilets and Ken Rickard. I think beer was somehow involved, along with the open-source platform Drupal.

Back in the last century when I was at the Star Tribune in Minneapolis, we built one of the first online traffic maps. It definitely wasn't the first even in the local market -- a local tech development firm had already built one as a demonstration. We considered using theirs, didn't like their terms, and assigned a programmer to create our own. It took some time, some socket-level Perl programming, some artwork and some creative uses of ImageMagick. MNDOT was happy to provide the data. A few weeks later we had it online.

That traffic map has been rebuilt several times over the years and is still around, but it's not any better and in fact is not as good as the original.

Today the climate is different. There's a whole genre of website, often developed quickly by an individual, using a technique called "mashup" in which Web services from a number of directions are combined into a single result. Adrian Holovaty's ChicagoCrime.org is an example.

I discovered that someone is building a mashup of traffic data, Google's mapping interface, and Google's Adsense revenue program at www.minneapolistrafficmap.com. I don't know who's doing it (the "about" page is blank, and the GoDaddy domain registration is cloaked). But it looks like the kind of thing that an entrepreneurial programmer could whack together in his/her spare time and, because of AdSense, create a small personal revenue stream.

The mashup actually has much more detailed information than the Star Tribune's map.

And I see that 10 years later, MNDOT still doesn't have any pavement sensors in the northeast quadrant, despite the notoriety of the I-694/I-35E intersection as a favorite place for trucks to tip or jackknife in winter weather. Fortunately I don't drive there any more.