The success of Flickr is hindering the advancement of media discovery. Don't get me wrong; Flickr is a good product - no question. However, its success has focused digital media discovery or whatever you want to call it in the wrong direction. House all your content in a large single repository. Then use metadata, folksonomies, and taxonomies (or tagging) as "the" method for relevant media discovery.

Working in the eLearning world for the last 3.5 years has taught me that singular repositories and tagging alone does not really work. The eLearning community has an impressive standard for tagging digital assets called IEEE LOM, which has over 60 elements with 9 categories. And even assuming that every learning object had all 60 odd elements filled out, the resulting metadata in many ways will remain as useless for discovering a relevant learning object as genre and song title are for discovering a relevant song.

However, Flikr's well deserved runaway success has gotten everyone talking/thinking/believing how critical tagging is for the discovery of content. This "Mind Space" is critically flawed. Tagging, be it strict taxonomies or loose folksonomies or something in the middle (Call them Gardened Taxonomies), provide no information on the actual context and quality of the underlying object.

Example 1.

(Search Result from OurMedia.org.)
Does the combination of search term and title provide enough information to figure out what these media objects are about?

Example 2.

(Search Result from Learn ALberta.)
Providing thumbnails and summaries offer a bit more information. Brands like National Geographic can indicate quality, but relevance is still uncertain. Some summaries are extensive some are not.

Example 3.

(Search Result from LavaLife.)
But do summaries and extensive metadata really provide the needed information about the underlying object?

Example 4.

(Search Results from Merlot)
Generic 5 star type reviews are not the answer to determining quality; like taxonomies they are only a small part of the solution. On their own they are of marginal value. Am I rating the quality of the production? The quality of the content? Its relevance to my need?

So why does tagging work for Flickr?
First, I think Flickr offers better viewing and management than discovery, and these are the reasons for its success. But still tagging works for discovering images because I can quickly ascertain Quality and Relevance without wasting a great deal of time. In fact the managed taxonomies of Getty and iStockPhoto are preferable to the folksonomies of Flickr for discovery.

Example 5.

Search Results from Flickr
It is quite easy to find images of sufficient quality and relevance from these results.

As with any large repository search, the level of quality and relevance varies greatly. However, presented with a tiling of thumbnails, a searcher can quickly ascertain both quality and relevance by visual examination. This simply is impossible with any other medium that takes more than 10 seconds to consume. Tagging alone simply does not work for any media that take time to consume. I would even go so far as saying that full text search on a repository of Word documents will provide only marginal gains for assessing quality and relevance.

So how do we find learning objects, music, and other-assorted-non-image-media without having to fully consume each and every object first?
Well we need to stop describing what objects are, and start describing how we use them and sharing that information freely with everyone in our community.

I'm a businessman at heart, and you sell a product or service by addressing your customer's needs not by showing her a features list.