Microsoft TechFest 2008: A glimpse into the future

TechFest 2008 | The greater Seattle area, USA -- home of Starbucks, Valve Software, and of course, Microsoft. While the latter two aren't in Seattle itself (Valve is in Bellevue and Microsoft in Redmond), everything seems a stone's throw away and you get the feeling that this place, just like Silicon Valley, is a tech centre of the world.

TechFest is Microsoft's annual event for its researchers to demonstrate projects they're working on. The event is huge in scope, pulling in people from Microsoft Research (MSR) centres around the world -- from Redmond, Cambridge (both in Massachusetts and London), Silicon Valley, Beijing and Bangalore.

Despite this, only around 40 of the projects are open to journalist scrutiny, another 110 are limited to the 6,000 odd Microsoft staffers who file through. We can only assume the remaining projects are so hot that they can't risk the information getting out, or that they're not far enough through development for journalists without mathematical, physics or computer science degrees to be able to grasp the concepts. That and it'd take a good week or two to see them all!

Media coverage at TechFest is a fairly new thing, the primary role of the event being for Microsoft employees to get an idea of what is being worked on. If something sparks their interest, the work could end up being used in a product, licensed to another body or may cross-pollinate with other research projects -- a process Microsoft calls "technology transfer", and has a dedicated team purely to manage it.

Free, as in beer
MSR centres are a rarity in the modern commercial world -- the research work has no requirement to be profitable, or end up in a product.

As an isolated case this means that MSR runs purely at a loss -- something a lot of companies would struggle to justify. However this freedom from commercial realities gives MSR the scope and vision to explore the crazy ideas that under ordinary circumstances would not be supported -- and those ideas often lead to amazing progress.

Microsoft Research started with just Rick Rashid on the books, and has now expanded to over 800 PhD employees.

Hit the Start button
The event began, as these things tend to do, with a keynote speech. Revealing his inner geek by wearing a Star Trek tie, Rick Rashid, senior vice president of Microsoft Research, gave a fairly interesting speech about MSR's beginnings, noting that it basically started from just him, and expanded at a rate that was likened to "creating a [University of California] Berkeley research department every year for 16 years".

"The fundamental raw material of Microsoft is smart people," said Rashid, going on to mention over 800 PhD qualified people are currently working at Microsoft Research, and last year 300 PhD interns were added to that number to help things along. He stressed the MSR works very closely with university and educational institutions, both sharing information for mutual progress.

Theme music!
The message was convergence this year (or as Microsoft puts it "collecting and amalgamating"), particularly in the imaging field -- bringing collaborative data together into one system, to give a more complete picture than previously seen. Rashid gave the example of medical imaging, in which the different types of scans used -- X-Rays, CAT scans, MRIs and so forth -- are still kept very separate, rather than compiling them into one scalable, explorable image of the patient in question, being able to switch between views and information as needed, giving a more complete and natural understanding of someone's medical history. To get a better idea of the sort of thing Rashid is talking about, check out the TED talk on Worldwide Telescope -- a similar project Microsoft will be releasing for free later in the year, which collates imagery from many different sources to give an inspiring view of the universe.

Part of this information convergence relies on the collection of data, but not in the form of deploying recording devices purely under a company's control, or user generated content as we know it today. Rather, the solution lies somewhere in between.

"The population at large is becoming the sensor network," Rashid said, and the demonstrations at TechFest certainly supported that. Almost every device seemed to report back to a central or distributed database of some sort, whether recording bumps in the road so other users could determine the safest roads to take, analysing travel routes of previous holiday makers so new holiday makers can see where is popular, or suggesting you watch Kevin Costner's Field of Dreams because you visited a field in Iowa four years ago that happened to be in the film.

Alan Alda was wheeled out to chat with Craig Mundie, for no discernible reason other than "Hey! We've got a celebrity!".

Stop: indulgence time
Craig Mundie, chief research and strategy officer for MSR then came out with Alan Alda of M*A*S*H fame, and proceeded to chat with him for the next hour. Sadly it seemed a simple case of "wheel out the celebrity" as the session was devoid of any useful or interesting information, save Alda's concerns that every time media updates (VHS to DVD, cassette tape to CD) certain works of art are lost as the current decision makers don't see them as financially viable to transfer to the new format. Mundie assured him that we're at the point where storage is almost a nil concern, and nothing need be lost from here on in -- what we really need to have now are natural, intuitive search systems to be able to find what we want amongst all that data.

And with that, TechFest began...

Disclosure: Microsoft Australia covered CNET's travel and accommodation costs while at TechFest.