Archive for June, 2007

US Route 90 Updated

In comments, Matthew Mims informs us that The Bay St. Louis Bridge has reopened just a couple weeks ago! This seriously throws up a lot of issues. Would it have been better if I hadn’t brought this to anyone’s attention? No, I don’t think so .. the wider issue is completely relevant, and fixing things by not doing anything isn’t really a solution. Shouldn’t I have made more checks prior to Where 2.0 (the bridge was not open when I first posted about it, but had partially opened days before Where)? Definitely. But if the model I’m advocating was followed, we’d already have updated the data to it’s new condition. The world is not in the least bit static. And now we’ll see how long it takes for the data providers to update again.

A couple months ago, I posted about a bridge destroyed by Hurricane Katrina still routed in all the major mapping providers. I gave it a brief mention at Where 2.0 too.


I checked again today, and the routing for Google & Yahoo (and others I suppose) has been updated to avoid the bridge. The bridge is still present visually.


This and other notable examples, like the MacArthur Maze fire, show that the geodata companies and mapping providers are listening to new sources — errors with high media profiles. And I’d think this must naturally lead to the thought that they need to listen to people generally — that the task of keeping maps up to date can’t be done without local people, and that many people are highly motivated to help. I sense a switch to OSM inspired models is going to happen quickly, or else lawyers are going to start pouncing on the data problems.

Steve gives his view in this interview.

Q: What has surprised you the most about OpenStreetMap?

A: The lack of overt hostility from within the traditional industry. Wikipedia got slated all the time by various content makers, but we’ve not really seen that (yet?). I think that geodata is a more fast-moving business and that people ‘got’ the idea fairly early, and it’s more of a question of what to do about it.

Comments (1)

just because you’re panoid doesn’t mean they’re not mapping you

Tom got the inevitable friendly take-down notice from Google, so the Google Street View Scraper is decommissioned. Tom tried to reason that he was scraping the metadata not the images, but such subtleties were politely dispensed with. Tom is now pushing ahead with open source panorama capture and display .. OpenStreetView. Gigapan should be coming around soon, similar cool stuff.

The time between release of a new feature, cracking, take down, and open source alternatives has collapsed completely since Google Maps was released two years ago. At this rate, a proper API for Street View should be appearing by next week 😉

And more hacks are appearing. These time lapse movies are awesome! and these visual analyses of the panoid are compelling moves towards a decode.


Where2.0 Presentations

I had a great time giving presentations at Where 2.0. Here are the slides and my notes.

Weaver House is split into 3, it’s the largest presentation I’ve ever given at 65 MB, and slideshare limits each upload to 30.

Weaver House Part 1
Weaver House Part 2
Weaver House Part 3

Weaver House Notes

This is my half of the Learning Lessons from Neogeography talk .. maybe Ian will post the whole thing.


OS-GeoRSS Notes

Comments (1)

Hacking Google Street View

We did some completely useless hacking Google Street View at WhereCamp on Yahoo’s campus. This is nice .. Google will have good mornings tomorrow with loads of strange requests in their logs from yahoo (evil neatness).

Greg Sadetsky cracked Street View a couple hours after the announcement at Where 2.0. It was pretty easy and comprehensible .. almost as if the Google engineers wanted it hacked. I’m sure they did.

There were loads of awesome ideas .. Grand Theft Auto ala Street View, and tricks to insert any panorama imagery into the viewer. That didn’t happen yet, but keep watched. Lots of stuff did happen around 3am this morning.

Google Street View in Upcoming

I took GeoRSS feeds from Upcoming, grabbed lat/long and associated that with a panoid via the RESTful “api” and pulled down the images (all that in a Ruby script) and then uploaded the panos to flickr with machine tags, via a perl script from Rich Gibson. A machine tag like “upcoming:event=200485” does magic stuff behind the scenes, those photos automatically included in upcoming pages. The upshot is that photos of venues are automatically added to Upcoming events .. and I think the 2nd Annual WebKit Open Source Party is going to be awesome since they’ll be enjoying these great panos on their upcoming event page.

Greg from Poly9 really pushed the technology above and beyond. Took panos, ran through facial recognition software, and added mustache and hat to everyone. GHATS. This should resolve all privacy concerns. .. Okaaay he really didn’t do it, but the idea is free for Google to implement if they take too much heat over this. Greg also built a handy street view debugger.

Google Street View in worldKit

Tom Longson built a database of panoramas in San Francisco, for future ease of hacking. And demo’d with a visualization containing only street view, no streets, in worldKit .. liberation! .. posted that to The original demo was bulit on Google Maps, but that couldn’t handle more than a couple hundred markers .. but that was pretty sweet since he had completely improved Google Maps by hacking in support for Google Street View (um right).

Google Street View on the Phone!

And we have Street View on the phone from Markus of uLocate, the builders of the awesome Where platform.

Comments (4)