Having a chance to attend the IKS Workshop, I enjoyed many of the talks and presentations that had a focus on semantic technologies in action. The workshop was very informative, (in a great location) with a good mix of attendees and some very nice presentations.
One of the talks was the “The Enterprise Search Market – What should be on your Radar?” by Alan Pelz-Sharpe arguing the need to improve Search tools in the enterprises.
Alan’s speach was a call to the community to still consider “search” as a problem in big corporations, and while many commercial and open source solutions exists, there is still need for improvement. What Alan was transmitting, was that corporations are not satisfied by the way their information is organized and retrieved. The keyword of his speach was “Search”.
I believe there is a need to extend and fully understand the meaning of search today. Google was a search engine, and so used to be many others. This means that they had a process that allowed anyone to search for a keyword (or phrase). Searching for something on internet based on keywords, used to be about finding new “stuff” (or finding information, if we define as information something new).
The search we enjoyed to do on search-engines was based on a random factor, where the result would serve something completely new and unexpected in content, but expected only in the relation with our keyword.
Findability is a new word being used a lot today. I was first introduced to an overview of this concept in the book “Ambient Findability” by Peter Morville (great book by the way). In practice Findability is exactly what corporations need today (in the context of Alan’s speach). They don’t need a search, their need systems that allow them to find the exact document/asset they are looking for. This is findability and this is where a search box is no longer enough!
Search engines worked on brutal way of organizing information based on keywords. They would basically go through a junk of information and try to extract out of there the best matches and put them on indexes. If we could note the order, is Junk of Data, to Order by a third party who analyzes your content based on keywords, NLP and some other great metrics.
What Findability should be, and what the Semantic Web promises is a new approach. Order first and then the rest will be easy. By using Faceted Search or other Information Retrieval interfaces findability is achieved. Computer Search is based on indexing a junk of data, while Findability should be a process defined at the moment when the data are created.