Andrew Otwell points us to an interesting half-hour video on Microsoft’s Channel9 site, which showcases their forthcoming mapping/geolocation product, Virtual Earth. As Andrew suggests, it looks to be way more advanced than Google Maps in several ways, perhaps not surprisingly given Microsoft’s (relatively) long history as the daddies of the digital mapping business. If you can bear the awkwardly non-staged, perhaps Dogme-influenced presentation – “Hey shove the ‘tech guys’ in front of the camera!”, including the presenter’s authentically childish “Take that, guys down at Mountain View!” and “Can you see sunbathing in the backyard? Hur hur“ etc. – it’s a pretty fascinating tour through the product, including some juicy facts along the lines that MS have had a SOAP API out for their mapping products for around 3 years now, doing 20 millions transactions a day …

But more on Virtual Earth. A single killer feature is the hybrid view, which superimposes graphical map information on top of the photographic satellite imagery. Whilst I’m not 100% convinced of the efficacy of being able to zoom in to direct-overhead satellite photos, the beauty of this particular city view is utterly compelling.

In terms of the user experience, it’s the already-ubiquitous AJAX-based approach (nice quote from one of the developers here: “The best way to share an application is over the web”), apparently cross-browser (which to MS means Firefox and Explorer). This means dragging maps around in the Google Maps style, but with additional keyboard control (arrows keys; + and – for zoom), as well as double-clicking to zoom in. This latter is interesting, as at some point, the double-clicking could stop zooming into the map and instead ‘zoom into’ (i.e. open up in the same window?) the website of the organisation concerned. This would be an extension of the Eamesian tunneling zoom, but instead opening up/continuing the zoom into the virtual representational space, rather than virtually rendered physical space. Interaction design problems abound – talk about a ‘threshold’ point – but possibilities perhaps.

The ‘game panning’ of the compass tool looks very smart i.e. replicating different pan speeds depending on your handling of the compass icon. Search is built in very cleverly, with queries remaining live and contextual i.e. it updates the query results as you drag the map around … So searching for “public libraries” will keep a continually updating list going in a pane on the left, as you pan the map around. A scratchpad to the right contains those particular results you want to check out – or indeed one-click-blog to MSN Spaces (with deep links, jolly good). You can layer these searches i.e. perform multiple queries, marked with different coloured and numbered icons on to the map.

The user experience seems fairly smooth, and making the map itself the interface has a pleasingly recursive edge. One aspect of the experience is fairly fatally flawed though: the erroneously-named ‘Virtual Earth’ looks set to start with only one nation. Can you guess which it is? (Clue: It’s not Wales). Again, geopolitical patterns of power implicitly replicated in code. A less hubristic name may be more appropriate while the Earth is in beta. But you have to start somewhere, and we can see why they have started there. And yet, from visiting Microsoft researchers and designers last October, I know that approx. 46% of Microsoft revenue comes from the USA, so the 146 countries they’re also based in would already seem contribute a greater proportion.

The real killer feature, to follow apparently, is the “eagle-eye view” which gives a more comprehensible view of the city, drawn from a plane’s low elevation flyover at 45 degrees. This begins to create a usefully recognisable city representation. This still isn’t how the average punter sees the city though (exception for those in high-rises), and I’d like to hear Microsoft’s thoughts as to something on the scale of Amazon’s ‘block-view’ drive-by photos at eye-level. Zooming down to that level gets useful, over and above the gorgeous wow factor of the SimCity elevation. To zoom and rotate from overhead projection to isometric view to the view from the street level – in both graphical map and photographic form – is surely the long game here.

Right now, however, the zoom from these ‘eagle-eye’ photos goes potentially a lot deeper but the developers seem aware of the latent privacy issues involved. So these are not real-time images, and they don’t zoom down to show license plate data or people’s face, for instance. Although tantalisingly, they could. The images are actually “cleansed” (their word, and how Orwellian?!), which means that the images of the people are removed. Bizarrely – and inadvertently, wonderfully poetically – they leave the shadows that the people had cast in. So no people are present in Microsoft’s cities, but their shadows are! This immediately conjures thoughts of other “cleansed” imagined cities, but is also reminiscent of Antonioni’s Blow Up. Will people attempt to reconstruct the events from their photographic remnants? From the shadows that Microsoft are mysteriously offering us? The lipstick traces of human endeavour scored throughout their imaginary versions of Detroit and Duluth?

Andrew Otwell:MSFT’s Virtual Earth movie
Virtual Earth: MSN’s answer to Google Maps