Whilst social photo sites like Flickr have grown over the last few years, many developers have been asking “now we have the resource, what do we do with it”. The first interesting application was tag clouds - it allowed you to see from the metadata in the system what was there and get an idea of how ‘heavy’ some concepts were. But what to do with the imaging itself? All those photos, and no easy way of making use of all the data they contained.

I’ve been ploughing through TED talks (no surprise that my favourite section is the What’s Next in Tech area), and have been meaning to post up lots of the talks, but one being discussed on a mailing list I’m on this morning is the demo of Photo Synth. Here’s the official video:

There are a couple of interesting things about this. Firstly, whilst the first half of the demo - a demo of Sea Dragon, a resolution independent image library - is interesting, there’s nothing truly novel about it. The only limitation stopping that system from being produced in the past is processing power. Every Computer Science/Software Engineering undergrad I knew had that idea whilst in the labs at University.

The second half though - the demo of Photo Synth - is what really grabs people’s attention. By computing vanishing points and common overlaps in images, it becomes possible to build a 3D representation of the object being photographed (in this case, Notre Dame) that you can take a virtual tour through. The applications are fascinating, not least because it takes mapping to a whole new level, and starts answering questions about what we’re going to start doing with all this social media.

One thing I noticed about that application, is if you upload a photo into a set that this software is processing, the software has to ultimately work out where you were in relation to the object. If just one photo in the set is geo-tagged (and many camera phones coming onto the market have in-built PS), I can work out your precise location. Now, let’s suppose you go on a tour of Paris. You take lots of photos all over the city. I now know your location when you were taking each of them. What’s more, the image will have within its metadata the exact date and time. I can, from that, construct a complete trace of where you spent your day from morning until night, with GPS-accurate location data, even though you didn’t have a GPS unit on you. Intelligence agencies are going to love this stuff…

Another interesting thought, is how this is being called a “Microsoft technology”. It wasn’t developed at Microsoft - they bought it in, and have worked out how to bring it to market. Well, when I say bring it to market, I mean do what Microsoft always do: make it available for Windows machines, but pretend the rest of the World doesn’t exist - the tech preview doesn’t work on Linux and OS X at the moment. This is a typical “flat World mentatility” prevailing at Microsoft I hope they’re going to change soon.

Microsoft are buying in a lot of innovation at the moment. They know they have a shortfall in innovative thinking (that’s what happens when your revenue is made up of sales of operating system software you can’t radically change and Office software everybody hates), but they have a big pile of money in the bank. By buying up the ideas and then pushing it out there, Microsoft are getting a lot of credibility within the geek community, and hopefully the idea-hungry culture will start to infect the rest of the company. I suspect a lot of people at Microsoft got a slight kick in the stomach when they saw Surface all over the web last month, simply because it’s such a radical change in how Microsoft looks at itself and answers the question “what is it we do?”

I have absolutely no respect for Microsoft, its software, or its business practices - I genuinely hope that for the sake of humanity the OSS community gets their act together and puts them out of business - but I’m starting to warm to some of the ideas and their employees.