Words You Can Use: The BAVC BLOG

Louis Juska, Lead Project Developer

Yeah, what about Ruby? I did say I was going to talk more about Ruby (and Rails) a couple of weeks ago. Guess what, I got sidetracked again.But this wandering is still in the ‘multi-platform’ programming vein. (You’ll see soon enough, that, for me, *anything* falls under that category.) It’s tangentially about a plug-in by Microsoft’s Live Labs for, get this, Firefox! A plug-in to uninstall Firefox, you’re thinking, right? Oh Ye of little faith.

It’s to have Firefox users tool around with Photosynth (don’t click there yet–the next paragraph is important). Without sounding too hyperbolic, I’ll say one thing: Photosynth–if successful–will change forever what we expect of visual archiving and retrieval. In other words, Google images is so 20th century (Flikr and its ilk, on the other hand, may provide much of the content we use in the future–see below).

The best place to start experiencing this thing is the TED conference presentation. Once I saw what this technology could do, and began trying to grasp how much data it accessed, parsed, and presented to users, I found that current attempts at aggregating image media with metadata are likewise so 20th century. Briefly, we’re talking about the ability of non-technical people to upload their vacation photos to, say, Flickr, and then have Photosynth find other posted images of the same places, collect them and their metadata (e.g. name, location, dates, comments), and return with an ‘essay’ of historical and contemporary images, including trivia, that would have taken teams of librarians days to cull. The resultant essay could also be a 3D rendering of many of the objects in the photos.

Although impressive, this example isn’t even the most profound one. Imagine being able to fabricate a virtual world based on real images and data for an encyclopedia. Children could virtually explore the Roman remains in Bath, England, hover over artifacts there and have discrete pop-ups of their materials and/or historical uses. They could visit reconstructions of ancient Athens, juxtaposing it with modern day, 3D representations (again, with easily accessible metadata). They could travel the span of China’s great wall, seeing trans-historical representatives in photographic and illustrative images. As more visuals and texts about other societies find place online, students (of all ages) will learn about people whom they didn’t know existed. As Blaise Aguera y Arcas says in the TED demo above, this is a tool to tap the collective memory living online. From a technical standpoint, we can further imagine the potential for metaprogramming to build resources automatically.

Of course, it’s also a tool that the more paranoid among us–wait, did that lamp just take my picture?–might feel can be used to harvest information about us against our wills and without our knowledge. Maybe some stalker will find someone’s name, enter it in every photo sharing site online, and build a composite image of his obsession. Or maybe, strapped for cash, the NSA will use it to surveil the world. Probably the most realistic danger comes from commercial information harvesters marrying mined Photosynth data with financial and publicly available data, thus building more granular profiles of consumers to more accurately assess them.

Come to think of it, maybe we should just unplug our computers.

(Thanks to Angelo Sacerdote for introducing Photosynth to tech staff here with a few links.)