Minority Report and Usability
9/14/2005
In spite of having a great story and cast, the most striking thing about the movie is the way they portray the future. Oddly, the society and culture that's painted attracts me.
Even though there are a lot of high fly gizmos that might seem to clutter a person's mind, the world looks simple. While it looks high-techy and geeky, it's not! Here are some points that grabbed my immediate attention:
- Personalization rests upon physical identity. We see this when Tom Cruise gets on the subway or walks around the mall. We are starting to see countries and digital economies pick up the concept of federated identity now. Privacy skeptics will always resist this.
- Voice activated commands can be used to operate tools. Think of Tom Cruise's house or office environment. How great would it be to have an object respond to you at a whim?
- Driving a car is automated, not just automatic. Why bother using the wheel or dealing with traffic when you could sit back, read the news, or stare at the scenery of a thriving world.
Google and Boston?
9/07/2005
I came across this post where Philip Greenspun asks: "Should Google have a Boston office?" I think that they should. He asked bloggers to give some thoughts across specialization and location.
Specialization: mobile applications. Thousands of college kids armed with phones. It's literally a walking distributed network. I know they just consumed Dodgeball - they could use Boston as a great test bed.Boston works better than a city like Los Angeles because it's smaller in size - you can go from downtown to Chinatown in a few minutes. That would take a lot longer to do here in LA. Not to mention, there are a lot of students that have the financial resources to own at least one mobile device. Come to think of it, what has Google done with Dodgeball recently? It looks like a potentially rich network - anyone with a mobile device is an active user. Also, is Google doing anything AI related for natural language processing? If so, being situated near MIT might be a good thing for obvious reasons.
I'm actually quite surprised that they don't have a place in Boston. It seems that they should. Hey - is there a Google office near you? I know there's one down here in Santa Monica.
Where Search is Headed
9/07/2005
Note, this is a rush post - I'll touch up on this again. Searching on the internet truly is a two dimensional experience - it's flat and non-engaging. I don't any of the existing providers suck. I do however feel that there is plenty of space for it to grow. I might even go as far as saying that giants like Google and Yahoo! could be taken down by another garage developer.
Most engines are optimized for servicing the client. They load within a browser instantly and return results relatively quickly. I think it is at this point where we have sacrificed innovation. What if I didn't care about the time it took to load? What if I wanted to have a better, dare I say, experience? The underlying message is a better interface that will let me do more things than typing queries with additional syntax.
Rather than going around in circles, I'll come and simply say that I think the future of search is clustering. You might even consider it aggregation - rounding up all the pertinent data, chunking through it, making connections, applying weighted averages of credibility, and giving a nice clean list back to the user. If all these insanely dense knowledge repositories (including blogs and wikis) keep popping up, then we'll think of search not as data retrieval but rather as data mining.
Relationships should not be judged based on hyperlinks. Who cares about how many pages your engine crawls. It's not about the size of your collection - it's about the quality. There was a time when Google returned higher quality results than the other engines. Now everyone is playing catch-up.
There are two distinct benefits of a clustering engine. First, as a customer, I only need to go to one place. I hate opening up tons of tabs and running different sessions that are related to the same topic. Second, all the data returned would be pertinent. However, unlike typical user interfaces, a clustering tool should give me tools that make me feel like I'm doing heavy lifting and give me a scratch pad for annotations. I want to do live bookmarking, dig on any search topic, and drill into related categories. An engine like this would kill del.icio.us or any other search tool for me. Researching with this kind of tool would be incredibly efficient.
