February 19th, 2004
google location
Google has a location service that you can use to search for things, like wifi hotspots, in your area. It’s cool, but fairly inaccurate when using data from Web pages that aren’t specifically listing wifi hotspots. This is a natural result of the malformedness of Google’s database (the Internet). You simply can’t get great results when your dataset is not consistent.
For example, I see hotspots in Berkeley that don’t exist. When examining the Web pages behind these supposed locations, I see Peninsula area codes (650 area code shares street names with Berkeley) and Buck’s restaurant (in Woodside, also a street in Berkeley).
Obviously, the accuracy, or “quality of service,” sucks. But, maybe that’s OK.
Google has built its reputation on good search results and innovation, and I wonder how much it can rely on the old 80:20 rule for software, in which the last 20% of development takes 80% of your time and resources. Buttoning up a project is very difficult. Google doesn’t worry too much about that. It puts out cool products in “Beta”, and the geek community, which influences the analyst community, eats it all up.
But I wonder how the public, my mom for example, would react to a product that works only this well. If your yellowpages was as accurate as Google location, you’d be pissed.
If the potential for the Internet is as a mass media, and Google is its current 800-pound gorilla, I have to wonder how well Google is going to be at creating products for the general public that work really well.
And that’s exactly when I’m reminded again of that old stalwart Yahoo. The game is far from over, my friend.