Friday, August 31, 2012

Re: [NGC4LIB] Searching for a man without a name

Posting to NGC4-LIB

On 30/08/2012 09:13, Owen Stephens wrote:
<snip>
Some disconnected comments/ideas on this:

Earlier this week Hulu added facial recognition to deal exactly with the "What has she/he been in?" question http://www.geek.com/articles/news/hulu-adds-face-recognition-2011128/

Another example of searching by image to find the solution to a web layout issue: http://herr-heaven.blogspot.co.uk/2012/08/solving-problem-on-website-using.html?spref=tw

The links being added between Wikipedia and VIAF (http://en.wikipedia.org/wiki/Wikipedia:Authority_control_integration_proposal/RFC) might facilitate some kind of image search related to person authority records. My guess is that book cover images could be very helpful in this way. My knowledge of the services offering cover images to libraries is a bit out of date, but my experience is that these are currently implemented with a single image per book (which doesn't capture the variation in covers across editions, territories, etc.) and retrieve the cover image on-the-fly at time of display which is a bit limiting (I thought LibraryThing offered bulk download of covers but can only see the API details at the moment - although since they encourage local caching would be more possible to think about using some image analysis s/w across a locally cached set of covers relating to a specific collection)
</snip>
The very concept of search is changing from what it used to be. One of the most enlightening blogs about search I have seen is SearchReSearch http://searchresearch1.blogspot.co.uk/ by Daniel Russell, who has the remarkable job title: "Über Tech Lead for Search Quality and User Happiness" at Google. The questions he asks and answers are of a completely new type. The latest one deals with a photo of a sculpture with no other information. Who is the sculptor? What is its name? Where is the sculpture housed? He says that this is a simple one that took only two minutes!

Another one is a photo Mr. Russell took of a street (somewhere) and the question struck me: what was I doing when I took this picture? (What a question!)

While the questions themselves may not be very realistic of genuine questions people ask, it is important to understand the sorts of questions that can be answered today. Earlier, such questions were more or less hopeless so no one asked them. Perhaps as people understand that these kinds of questions can be answered, they may find themselves asking them more often.

The way I found out about all of this is after watching a public lecture he gave, which I suggest to everyone, "What Does It Mean To Be Literate in the Age of Google?" http://hulk03.princeton.edu:8080/WebMedia/flash/lectures/20120228_publect_russell.shtml

No comments:

Post a Comment