RouseWorld | Main -- Google, the Internet's scrapbook

Google, the Internet's scrapbook

Although Tim Bray doesn't like Google Image's fixation with file names, that fixation gives way to an interesting process of discovery.

As Bray mentions, many people put up images from their digital cameras without giving them meaningful file names. The camera itself gives a (mostly) numeric file name and a lot of times that is the file name used when the image is linked into a web page.

What that means is you can enter in almost any number and get all the images Google has found that have that number in the name. To me the right balance is struck with a four digit number. Too many numbers and you either get catalog images or nothing, two few images and the mix isn't focused and the results too many.

Of course, there are web pages that automate this random search.

But what interests me is what could be done. Camera images now contain a lot of metadata — information like what date the picture was taken, what model of camera, what the exposure was, if a flash was used or not.

What if Google indexed that information?

Imagine looking for pictures taken during daylight hours, inside the last month, without flash, by a Nikon D100. That would be good way to search for recent outdoor shots by "prosumers." Of course, it gets more interesting. There are fields for GPS data and GPS units are getting smaller and more ubiquitous. Indeed for cataloging one's own pictures it would be helpful to have that information with the picture.

So what happens when you can search for all pictures within 20 miles of your house — or inside 30 meters of Jenna Jameson's house?

The future is going to be interesting.

Comments