For a long time we have been wondering if Internet browsing will continue obviating visualisation, delivering endless lists of results that we have to individually consider. We have already spoken about the attempts to make browsing and searching a friendlier activity, easing the navigation through Internet and through the mountains of information that fall daily on our shoulders (see in the archive numbers 24, 47, 48, 51, 52, 55, 70).

InfoVis.net seen by Grokker. Notice the lef window showing in circular form the links related to InfoVis.net.  Source: Screenshot of Grokker by the author. Click on the image to enlarge it  (177 Kb).	InfoVis.net seen by TouchGraph. In this case the links afound through Google are shown in the form of a graph. Source: Screenshot of the Java applet by the author. Click on the image to enlarge it (60 Kb)

Recently the efforts have evolved to include more sophisticated elements. One of the examples, following the thread of the last articles about graphs, is TouchGraph LLC a Java plug-in for Google that builds, from a primary link, the network of the relevant links to it. An interesting tool, somewhat in “á la KarTOO” (which we spoke about in number 97) style, with less colour and possibly a more “academic” look.

The current KarTOO Interface. The word used to initiate the search was "Barcelona". Source: Screenshot by the author. Click on the image to enlarge it  (226 Kb)

With a different philosophy and a much friendlier look comes Grokker, a meta searcher and browser produced by Groxis, that allows you to dive into the information in a visual form with a different metaphor. It can be downloaded for a free trial period of 30 days.

Unlike other similar systems, Grokker isn’t a plug-in or a website that produces graphical output. It’s a program that is installed in your computer and acts as an alternative to the usual browser. Grokker has a series of options that allow us (at this moment) to classify and visualise information from

• Our hard disk or whatever one connected to the local area network, 
  
  
• Internet, as a meta searchers using the output from AltaVista, MSN, WiseNut, Fast, Yahoo, and Teoma, 
  
  
• the products available at Amazon.com. 

In the near future they are working to complement them with the Library of Congress, Google and a series of plug-ins.

As in any search engine, you can enter a line of text expressing what we are looking for. Grokker searches for it, organises the output into categories and shows it visually in one out of two main spaces in which it divides its window. In the other one it shows, like a normal browser, the specific pages that you reach by diving in the visual metaphor of the other one.

Several zoom levels of Grokker. The word used to initiate the search was "Barcelona".Top Left: The usual Grokker window with a web on the righ half-window. Top right: We have eliminated the right half of the window in order to see only the visualisation window. Down left: zooming by clicking on the "Hotel"circle. Down right: Clicking on "Discount Hotel" to obtain a closer view. Source: Screenshot of Grokker by the author. Click on the image to enlarge it  (180 Kb, 108 Kb, 195 Kb y 157 Kb)

The visual metaphor that Grokker uses is circular and recursive. The search space is presented as a circle that occupies the maximum area that its window has. The search results are grouped into categories that Grokker derives automatically from the results. Each main category appears then as a circle itself within the larger one. 

Again within these circles new ones appear that pertain to subcategories. This schema continues until individual pages or items are found. The latter are presented as rectangles instead of circles and you can access them by clicking on them. The size of the circles is proportional to the amount of information they contain. There’s also the possibility to show nested squares instead of nested circles.

The possibility of changing the visualisation from circles to squares in Grokker. Source: Screenshot of Grokker by the author. Click on the image to enlarge it  (180 Kb y 208 Kb)

One of the keys of Grokker is its automatic classification algorithm that extracts the categories and subcategories. In this sense it’s not so different from Vivísimo, a tool that instead of representing the categories in visual form, produces a hierarchy and shows it like the Windows Explorer, as an unfolding tree. 

Searching "Barcelona" through Vivisimo. Notice the category tree to the left od the image. Source: Screenshot of Vivisimo  by the author. Click on the image to enlarge it  (86 Kb)

The point that differentiates them is the visualisation, classical in one case, more risky in the latter since the cluster of nested circles is at least infrequent. The visual proposal of Grokker is elegant and consistent. Each circle appears in a colour that can be codified by the user, for example in function of the age of the information or the price or ranking of a book if we are diving into amazon.com.

It has a series of filters that reduce the amount of information shown. There are basically four types:

• Textual, they allow you to hide contents related to specific topics, for example the information about hotels, if what we are looking for has nothing to do with them.
  
  
• Range, using sliders you can limit the ranges where the information lies in-between, for example books between 6 and 12 euros or hard-disk files bigger than 2 Mbytes.
  
  
• Selection, asking, for example, that we only want to see Japanese sites or non profit making ones.
  
  
• Colour, These type of filters are associated to ranges and are used for their visual codification.

Deciding whether Grokker TouchGraph, KarTOO, Vivísimo or whichever other existing alternatives to searching and browsing will turn themselves into standards, sending the current browsers to the museum, is certainly risky and uncertain. Nevertheless, there’s a fact that appears very clear to me: increasingly more solid and useful alternatives are beginning to appear, that are based on consistent visual metaphors that employ diverse complementary techniques that gather into one powerful and versatile tool. 

Few of them resist the temptation to call themselves “the Revolution”. But, although they are, in reality, the natural evolution of a series of existing developments, they are probably settling the foundations of what, no doubt, will be an easier and more intuitive search, a visual reference like Google is now a textual one.

Links of this issue: