I’m always cautious in posting anything about Google Earth, given my involvement. But I get enough questions from friends and I see enough interesting developments, I decided to make the rare comment.
First, a word of congrats to the GE team for being nominated for a webby. It’s a bit unusual, since it’s not a website that’s up for the award. But since the Google acquisition of Keyhole, the thing has taken off in popularity and all of the community enhancements are just amazing.
I can say I feel some sense of achievement for having helped build the thing, but I realize the people who stuck with the company through the lean times deserve full credit for the current level of success. Suffice it to say, I still get some pleasure from watching CNN use it in place of expensive graphics and seeing how it still kicks the pants off its competitors six years later.
Which brings me to item #2. It’s not well-publicized, but a maker of another terrain viewer is still attempting to sue Google over claims of patent infringement. The following news story comes as a welcome bit of relief that there may be some residual sanity in IP law, though the fight is clearly not over. The trial is apparently set for this fall.
I’m staying out of it for obvious reasons, though I do have strong opinions on the validity of Skyline’s patent (and many other over-reaching software patents that pop up long after companies work hard and innovate — the opposite of what the patent system was intended to do). I won’t say much more than that. But if their patent is so clever and Keyhole had actually copied any techniques, wouldn’t you expect Google Earth to perform as poorly (IMHO) as Skyline’s own software?
Interestingly, I remain on cordial (but clearly out-of-the-loop) terms with Keyhole senior staff and, at the same time, the investors in Skyline also have a stake in another company I’ve worked for (the work itself was graphical, but unrelated). So you can understand why I’ve kept so quiet and steered clear of any requests for information by Skyline. But it’s hard not to call shenanigans on a lawsuit that is, in my opinion, a mistake at best.
Anyway, the most common question I’ve been asked is just what it was I did for Keyhole. And since I already have a portfolio on-line which includes this work, I’ll answer the question here in more accessible terms.
In 1999, I joined Intrinsic Graphics, an advanced game technology company, as the 2nd employee after the founders. As a result of work I did there on an internet-enabled Earth Viewer demo showcasing the "universal texture" rendering technology my co-workers had developed, I was offered a co-founding role in the spin-off company, soon to be named Keyhole. Intrinsic also hired a CEO (John Hanke), and a former SGI employee (Mark Aubin) and the three of us (well, five, counting Michael Jones and Brian McClendon from Intrinsic on our board and giving expert guidance) co-founded Keyhole, with John doing the business work, Mark handling the massive data-acquisition task, and I was officially responsible for building the software that runs on your PC.
We all did an enormous amount of work, so I don’t want to claim extra credit at anyone’s expense. But some of the work I did remains highly visible, like the way you zoom across the planet from one spot to another, the labeling of roads and cities, and hashing out much of the original user-interface design. A good chunk of my time was spent in "directing," though I think I did better on the technology side. On the back end, Chikai, Phil, and I worked together to define novel ways to stream very-multi-resolution content over the internet, taking advantage of some fun properties of image compression, implicit geometry, clever caching, and so on. Although there are some competitors to GE that look pretty polished, and certainly some new and useful features around, in terms of sheer performance, nothing still comes close. As for overall usefulness, that’s the on-going race.
People don’t often realize, but a lot of the ideas about geo-targeted advertising and dynamic content were planned out from the beginning. Though KML was developed later, EarthViewer was scriptable and adaptable from the first prototypes. We very deliberately built a full 3D search engine inside the app to sift through just about any kind of spatialized data with minimal overhead. In truth, GE is only superficially an "earth browser." It’s actually very similar to Google’s massive search engine servers, but using spatial queries instead of keywords, with as much of the code and data residing on your computer as necessary to ensure the best interactive response. I imagine that’s the main reason Google bought the company, apart from the cool visuals.
It’s just now that Web 2.0 is catching up with the ideas from 2000 or before. It gives me some pleasure, and at the same time, some small regret imagining what more I could have done had I stayed in that line of work. I’m not complaining, of course, since I did other useful things with my time. But I’m looking forward to whatever the GE team comes up with next.
UPDATE: Google Earth did, in fact, win the webby, two of them. Congrats GE Team!
* IMHO = "in my honest opinion," which may or may not be humble