19 May 2008

Visual Search

A few years ago, I read about a new kind of internet search engine, which held out the promise of a different, better way of presenting information from the web: as a series of visual worlds, with lines delineating the interconnections between each piece of information, and thus emphasizing relevance. In other words, showing search results from the web as map of the web itself. That search engine was called KartOO, and from time to time I have returned to it, to see how it has changed or improved.

In the years since, the basic KartOO presentation has stayed the same, while the features and functionality have evolved and expanded. Users can subdivide their search according to specific categories (e.g., images, videos), or zero in on better results by selecting from a “topic” on the left side navigation. Moving the cursor over the results reveals the connecting lines between different bits of information. Scrolling forward generates a new “map,” showing the next set of results and their web of connections. Searches can also be saved for later use.

The reason I keep coming back to KartOO is the hope that it might somehow prove useful – more useful than it so far has. “Visual” searches have an intellectual appeal, a la mind mapping software, but do I necessarily care that page X links to page Y if one, both, or neither have the information I need? Maybe a system like KartOO is just ahead of its time: maybe the reliability of information within and throughout the internet remains so inconsistent that revealing the relationships is currently more confusing than clarifying. Still, I encourage people to test KartOO: each of us processes information differently, and some may find such visual representations more effective than Google, Yahoo, or other’s straight lists.

That said, I have also wondered when a competitor to KartOO would appear. There are two other visually based search engines of note, though neither functions with KartOO’s level of detail. The first is Snap, launched about two years ago, which tried to expand upon the basic search approach by providing users with a “preview” of the web site related to each search result, with a list of links and text on the left, and the preview pages on the right. Conceptually, this is used the idea of visual identification to support search: if the preview picture showed a blog, but users wanted a newspaper, it would be easy to dismiss the result and move to the next one.

SearchMe (currently in beta testing, but accessible to the public) takes a similar approach, but with an even stronger emphasis on the pictures. Results show “pages” for the user to scroll through, with the relevant text highlighted at the bottom and literally circled on the page. Recognizing this is not a fully launched product, it’s hard to be too critical – but I found that the search tool itself generated inconsistent results, while the idea of trying to decipher the often-small print on a pictured web page made determining the right result a real challenge.

Neither Snap or SearchMe aims to do what KartOO does, however: they are “visual” search engines only in their reliance on pictures of web pages. For now, as far as I know, KartOO stands alone in trying to map the process of searching the internet.

Users who have comments on these search engines - or recommendations for others - are encouraged to submit them, and relevant comments will be published.


At 4:51 AM, Blogger Ron C. de Weijze said...

If you would be interested in modeling, searching and navigating all that you ever learned yourself or will learn, then check PMM Personal Memory Manager. www.pmm.nl

At 5:01 AM, Anonymous Anonymous said...

one for your list, one week old


Post a Comment

Links to this post:

Create a Link

<< Home