Thoughts on this week's news

It's been a busy week for news, with many announcements timed to coincide with this week's Where 2.0 conference and the Google developer day. And for the past couple of weeks I have been on vacation in the UK with limited Internet access, but got home to Denver last night so am trying to catch up a little.

Both these events were covered extensively online, for example at Google Earth Blog and AnyGeo, as well as the official Where 2.0 site. I thought I would just add a few comments on some of the announcements that I thought were interesting.

Google had a busy week, with announcements including the new Street View in Google Maps, Mapplets which they describe as an easy way of doing "mashups of mashups", and the acquisition of Panoramio. Microsoft announced availability of 3D building data for New York City in Virtual Earth (for the second time - and unlike the last time it's really there now!), as well as a lot of additional data in other places.

The Street View is very nicely done I think, providing a good balance between rich data and simplicity of use. Exactly where you can go and which directions you can look in are more constrained than in a full 3D environment like Google Earth, Virtual Earth or SketchUp, but it's a lot easier to navigate along a street and get an impression of how a neighborhood looks with this more constrained approach. The way that one view smoothly transitions to another maintains context very well. The fact that the data comes from Immersive Media is interesting - I have seen their stuff a few times before and been impressed with them. They capture continuous data from a spherical assembly of video cameras, so this suggests that there is a richer set of data behind street view than is actually exposed at the moment. For example, when looking around Denver I found myself wanting to look upwards at times when I was in front of a tall building, which I could only see the lower portion of. But of course this would add some complexity to the user interface, so there are trade-offs in adding this type of functionality.

This raises the interesting question of the relative value of a "true 3D" environment like Virtual Earth or Google Earth versus a "pseudo 3D" environment like that provided by Street View - and how the value relates to the cost of capturing the data. Obviously there are things you can only do with a true 3D environment, but for many applications something like Street View may both provide a simpler user experience, while the data is much cheaper to capture.

The Panoramio acquisition raises a few questions for me. It's presumably not a technology play, as there's nothing difficult about displaying geocoded photos in Google Earth or Maps. So I suppose it's primarily a content play - they have around a million geocoded photos, but this small compared to the 18 million or so geocoded photos which are currently in flickr, and which can easily be displayed in Google Earth or Maps in many different ways. Google has its own online photo sharing service integrated with the Picasa photo software it acquired. They really need to integrate these two offerings quickly - you don't want to have to do something different for photos that you want to assign a location to. I want to be able to upload all my photos to one site (I use flickr these days), and assign locations to some of those as appropriate (soon that should be automatic via the GPS in my BlackBerry!) - I don't want to have to load those which I want to display on a map to a different site. So anyway, it will be interesting to see what Google does in this area.

Microsoft announced its touch table for what it calls surface computing (short video here). This incorporates the idea of multi-touch interaction, which has been around for a bit and will be included in the iPhone - this is a powerful way of interacting with maps. But it also adds in the notion of interaction with multiple different devices by placing them on the table. This is an example of what has been called sentient computing or ubiquitous computing, which I was involved with at Ubisense. An important notion here is computers reacting based on what is happening in the physical world. In particular if you can precisely determine location of objects (probably within inches for this type of application), you can do some very innovative things in terms of user interaction with computers. The Microsoft technology is based on video recognition, using cameras located in the table. Unlike a lot of other research systems in this type of area which are pretty far away from being commercially available, the table will be commercially available this year, and not too outrageously priced for something like this - they say in the range of $5-10K, and expect the price to come down significantly over time.

Finally for now, FortiusOne announced availability of GeoCommons which is interesting. They are trying to bring more sophisticated geospatial analysis to a broader market, and have done a lot of work to assemble publicly available datasets into the GeoCommons database. I plan to look more at that in a future post.