Link This |
Email this |
Blog This |
Comments (0)
StumbleUpon New Ideas for Libraries
March 5, 2008
Ack! I just noticed that this post never went live. The blogging software used here was having some issues. This should have come out a couple of weeks ago to followup on an earlier post.
The idea is simple. You tell
StumbleUpon what kinds of websites you are interested in, and it shows you random sites that match your interests. Only it isn't that simple. StumbleUpon isn't just picking a site at random, but rather it is using a powerful recommendation engine to determine "alikeness" of sites to find other sites that might work for you. Sound familiar? This strikes me as being rather akin to Readers Advisory.
So what can we learn from StumbleUpon if we have already been doing the same thing for years. I think there are a couple of ideas we can take away from the site. One is diversity - Readers Advisory is great for people looking for the next book to read, but doesn't really help others. StumbleUpon's toolbar tries to broaden the appeal by letting users look for sites, pictures, videos, news, or what their friends are looking at. A bit of a daunting task for librarians who have a hard enough time trying to stay current in all the genres of books!
The second key is that StumbleUpon makes use of powerful tools to facilitate the recommendation process. Instead of recommendations being generated by a single person (or a single, closed source) the site turns to the collective recommending potential of a very large user base. Libraries have this same potential as is clearly demonstrated by
LibraryThing. Give dedicated readers an opportunity and an easy to use tool and they will gladly help you find connections between books.
The system is pretty simple; you need to have user accounts, a rating system, and a tagging system. User accounts let you track alikeness of items as perceived by a single mind. For example, User2345 liked these three books, but did not like those two. If you also disliked the same two, and liked two of the others chances are good you will also like the third. The likebility can easily be rated using the now standard 5 star system. Survey developers would note, however, that a 4 star system might be more effective in the long run as it eliminates the opportunity to select the middle ground and avoid an opinion. Finally, you need tagging to complement subject headings. Books might have very similar subject headings and yet be incredibly dissimilar in their tone, style, and even content. While subject headings can provide a starting point for alikeness, letting users describe items is a better option.
The final function that StumbleUpon uses to help ensure accuracy in their recommendations is feedback. Users can quickly mark items that they feel have been incorrectly tagged or categorized. Users can also mark whether or not they liked a site to quickly provide feedback.
So, do librarians need to be more like StumbleUpon? Not really. We actually work rather a lot like StumbleUpon already. We have access to vast sources of potential information to help users find what they didn't know they were looking for. It's just that we have to do a better job of letting people know that this is a service we provide.
Posted by Chris Harris on March 5, 2008 | Comments (0)
