It seems as if everything I’ve tried to write in the last couple of weeks is an exercise in contrariness. I apologize in advance.
Eugene Barsky and Allan Cho have an article in the current issue of the Journal of the Canadian Health Libraries Association, Introducing Web 2.0: social search for health librarians.
It’s great that Eugene and Allan gently introduce resources like YouTube and Flickr to Canadian health libraryfolk, but I find myself uncomfortable with their calling these resources “social search tools.”
I’ll attempt below to explain in simple terms what I understand a social search tool to be and offer some examples you can try.
So what’s “social search”?
Wikipedia has as good a definition as any:
A social search engine is a type of search engine that determines the relevance of search results by considering the interactions or contributions of users.
Get the idea? When Google returns your search results, these results are ordered by the PageRank algorithmCheck out this excellent recent article for a long, complex, and good explanation of PageRank, which notes (among other things) how many other Web pages link to each search result. The more Web pages link to a search result, the higher up the list of results it’ll appear. (Note to geeks: Yes, I know this is a dramatic oversimplification)
A social search engine puts this aside and asks itself: What have users liked when they searched for this topic? The more other users have indicted they liked the result, the higher up the list of results it’ll appear.
If YouTube and Flickr aren’t social search tools, what are they?
YouTube and Flickr are searchable collections of user-created content. In the case of Flickr, users upload and tag images that others can search. In the case of YouTube, users upload and tag video that others can search. Searching social content is not, in my thinking, the same thing as “social search.”
Is del.icio.us a social search engine?
You could call it that and not be wrong- but I don’t call it that. I haven’t been able to find any documentation on how del.icio.us sorts its search results, but Iit appears that the number of times a URL has been saved DOES figure into it…so it is accurate to say that searching del.icio.us is a kind of “social search,” but I wouldn’t call del.icio.us a “social search engine” because it is so much more than that. (I looooooooove del.icio.us.)
What about Google Custom Search Engines? Are they social search tools?
Some of ‘em are, but most of ‘em aren’t. Google Custom Search Engines are really pretty simple: Each uses Google’s engine, but its creator get to decide from which sites the CSE will return results. When I made the Consumer Health and Patient Education Information Search Engine I restricted it just to the sites that were recommended for consumer health information by the NLM or CAPHIS. Why don’t we call it a social search engine? Because there’s no social component. A single individual (me) plugged in which sites to include and set it out to be used. It sorts search results by Google’s PageRank, just like regular Google. Heck, even if I tweaked the way it sorts results, it still wouldn’t be social because I’d be the only one controlling it.
But a Google CSE could be social. When you create a CSE, you can invite others to add sites from which the CSE will return results or even place a form on its front page by which users can request access to add their input. In that sense, a Google CSE could be called social….
…but would I want to open up my CSE to be tweaked by anyone?
That depends on your goals and asks one of the most important questions about “social software.” Like a lot of social software, the value of social search is significantly impacted by who is participating. In the same way that AskDrWiki and Ganfyd have more value because they limit participation to licensed clinical professionals, some very specialized kinds of social search might similarly benefit from restricting participation to only expert contributors. If we take a look at the Google CSE that Alan and Eugene give as an example, we find out that (with good reason) it isn’t social at all and is maintained by one person:
There’s nothing wrong with a good CSE that is only created by one person- but it is mistaken to call it an example of social search.
So what are some good examples of social search engines?
Here are three:
URL.com retrieves the top ten results for your search terms in Google, Yahoo and MSN, then lets the users rate how appropriately ranked they are.
Click thumbnail for larger screen capture
Sproose is sort of a cross between Google and Digg. You run your search, then vote for the results which you think are best.
Phil Bradley on Sproose
A Swicki is sort of like a Google CSE, but a lot more social in that it learns how better to sort search results based on the activity of users, even if the users aren’t specifically seeking to tweak it.
Is Social Search for Medical Libraries?
Sure. The library staff might make their own collaboratively-created Google CSE or Swicki of favorite, subject-specific sites (or have a CSE generated from a del.icio.us account’s links). Librarians should seek to be familiar with technologies for finding and organizing online information and social search is not likely going to go away as an idea any time soon.
My bottom line is that when it comes to health information for healthcare professionals, social models are only worthwhile if participation is restricted to those whose input is qualified. The Google Health Co-op, after all, is just a large-scale Google CSE where the invited 26 participating organizations are expert in healthcare information.
What am I missing?
Are there other ways social search tools can be used in medical libraries? Do you have any favorites? Please leave a comment and let me know.
Melissa L. Rethlefsen looks at social search engines, where search is heading, and what it means for librarians
Melissa is a Web-savvy medical librarian and I read everything she publishes. This article gives a very nice overview of a few social search tools.