Thursday, April 26, 2007

flickr at the organizational level

We're attempting to build an organizational collection of images for teaching around here using MDID, which is great open source web-based software for teaching with digital images. It's got lots of features tuned to teaching the images in the classroom, many of which center around its presentation tool which allows for zooming and side by side display of high res images.

It's got two major flaws, as I see it:

  • It's a silo of data. Folks looking for images don't go to it first when they're looking for something.
  • The UI on it for organizing images isn't nearly as nice as Flickr. I realize this as I use Flickr more and more for my personal stuff. Though MDID has a personal collection feature, there's no way I'd recommend it over Flickr.
It would be great if Flickr offered some kind of organizational account that facilitated organizations building image collections within it. Or perhaps we need to make MDID work with Flickr. MDID could drop all the functionality that Flickr does better and just concentrate on the education specific features.

Wednesday, April 25, 2007

Freebase explored

Got a chance to try out Freebase thanks to an account from Paul Miller.

Once you've logged in, the site has some nice Flash based tutorials that explain the idea behind the service pretty well. You can also browse the categories 'types' that are setup in freebase. Basically, they are designed to span the world's knowledge, a pretty ambitious goal. Top level types include:

  • media
  • sports
  • science and technology
  • art
  • travel and dining
  • special interests
...among others

The amount of data in these areas is still pretty small. They've pre-loaded data about films and restaurants; many of the demo applications take advantage of this data, such as cinespinner.

As I browsed through the categories, I thought about how similar this kind of organization is to library cataloging, especially authority control. It would be great if the catalog and authority data in WorldCat was pre-loaded into Freebase, as well as other datasets with a certain amount of authority such as the Getty Vocabulary Program. Seems like it would also make sense to massively load in data about consumer products.

How to convince organizations to put their data into this open database? Perhaps they could get a fee based on queries using their data when a Freebase account went over a certain daily query limit. Sort of an Amazon web services model. I guess that would make it unfree, technically, but what really is?

Thinking more broadly, once the rules were established for a global database of knowledge, it could be replicated and hosted by more than one organization--with each hosting company providing basic querying functionality but also perhaps some value added services. Those are my anti-monopolistic tendencies showing through.