Sunday, October 22, 2006

Filter

I'm very interested in the idea of using the Internet to let people help each other.

You have surely seen websites such as Amazon and Netflix that use software called a recommendation engine to try and steer you toward other products you might be interested in. It's a very effective marketing tool, however it's rare for me to discover any new really great music this way. At the same time, I regularly go out to public places where I do hear really great music I'm not aware of. So I know it's out there.

Systems like Amazon's use a technique called collaborative filtering to try and match the tastes of others to your tastes. This method is good for finding others with similar music collections, but doesn't take any of the qualities of the music itself into account when making a recommendation.

A new service called Pandora tries to remedy this by hiring a staff of music experts to painstakingly listen to lots of music and catalog the musical qualities of the music. By matching songs you like to others with similar musical qualities such as genre, tempos, chord progressions, type of lyrics, etc, Pandora can pick music that is similar to songs you choose.

Pandora then plays a radio station for you of music similar to your taste. I like to listen to the music on Pandora, but I still rarely get any really great new recommendations. I'm still more likely to hear something better by going down to a local hot spot. I regularly ask staff members "what's playing right now?" only to get an indifferent shrug most of the time.

In the mean time, I have made two really good recommendations for friends in the past month. I hear something new and immediately think of people who I think would like it. The surprising part is that I have had a pretty stellar record of success in that my friends tend to really enjoy what I recommend.

To me this is vague evidence that maybe humans are better recommenders than computer algorithms. What if humans could recommend music to one another instead of computers. What a concept.

Wouldn't it be nice if you could put out a list of your favorite songs that are similar. Maybe even some text describing what you like about them. Then other users having the same favorite songs on their lists could read yours and suggest new ones to you.

Why would they do this grand favor for you? Well in my last few ideas, I've used the give and take model. By giving good recommendations to other users, you put yourself in a better position to get recommendations by others. Give some, get some in return.

I'll discuss this idea with others to see if they would think getting better recommendations would make them likely to spend some of their own energy to do so. I'll also see if I can find anyone else doing anything similar.

Here's one interesting article I've found.

Monday, October 09, 2006

Furniture

I decided last week that it was time to explore a new hobby.

One of the things that has always bothered me about my software development career is that it is just too darn abstract. The work I do only exists inside the computer. If the power goes out, it's as if it never existed. If I want to show someone work I have done recently, I have to bring them to a computer. Something about this has always bugged me. It's as if what I do is imaginary. I realize that my work is generally quite useful to those who use it, and often saves people large amounts of money in more efficient operations. But still, I have an envy for those who create things more concrete. Things.

Somehow, furniture has always intrigued me as something both useful and, at time, graceful and stylish. Furniture is something everyone needs. Furniture is something everyone understands. Furniture is something that lasts. Hopefully. But most importantly, furniture is something you can touch, sit on, sleep on, stand on.... Its real.

Making useful things has always been important to me, and has led me in to the software industry. What's missing to me lately, has been this notion of a concrete solidness to my work.

Furniture seems such an intriguing area, that I feel I should explore it further. I don't have any real woodworking experience, nor much crafts or design experience for that matter. I am completely green, with only curiosity to fuel me.

To document my journey, I have started a separate blog, Concrete Utility. The name describes the two most important aspects of furniture to me. It's certainly far from prose, but I hope it helps me remember what I have learned along the way and to keep track of tips and resources. Maybe it will even help someone else along a similar journey if they somehow come across it.

A new hobby. Just like that.