Topic mapping for context, searching for content: graphical approaches in the new generation are almost "meta"-topic/concept maps.

Online, May-June, 2003, by Isabel A. Stirling

COMBINE SOFTWARE DEVELOPERS with visionary entrepreneurs and what do you get? The development of new and exciting ways to retrieve, organize, visualize, and comprehend information graphically and intuitively. Companies are springing up to commercialize the technology, called topic mapping.

One new company creating topic mapping information retrieval is Groxis [www.groxis.com], which has created a program called Grokker. It retrieves and organizes searches in a graphical and contextual way. (The name grokker is from "grok," a word coined by Robert A. Heinlein in his 1961 science-fiction novel Stranger in a Strange Land. According to The American Heritage Dictionary of the English Language, 4th Edition, it means "to understand profoundly through intuition or empathy.") Since Grokker's debut in October 2002, it has received much attention from technical editors in the press (New York Times, CNN, Fortune, Business Week). Although still in the final stages of its Preview Release version, it will be widely available in the coming months. As with many software programs, there will be a basic and professional edition. An enterprise-wide version is also planned. At Demo 2003 in Scottsdale, Arizona, where it won best of show and the coveted Demogod award, Groxis unveiled its Preview Release 2.1, which is priced at $99.95.

EVOLUTION OF TOPIC MAPPING

Earlier generations of taxonomy software include those from Semio, now owned by Entrieva [www.entrieva.com], linguistic software that does content indexing and categorization through its SemioTagger and visualization with SemioMap, and Inxight's Star Tree Studio [www.inxight.com]. Both of these companies have gone through change and development, and both are active, particularly in the business environment, in managing/mapping information software for companies. Inxight recently released its Inxight Smart Discovery software that extracts metadata from documents for businesses.

Other "second-generation" software included Mohomine's MohoClassifier (now evolved into Mohomine's Mohoclassifier, partnered in a strategic relationship with WordMap), Cartia's ThemeScape, and Metacode's Metasaurus. Cartia was acquired by Aurigin in 2000. Aurigin was then acquired by Information Holding Inc. in 2002 and merged into IHI's MicroPatent product. Metacode was acquired by Interwoven in 2000, but much of its technology is evident in Grokker.

Ontopia [www.ontopia.com] and Antarctica's VisualNet [http://antarctica. net/] both have some good ideas about the topic mapping potential, but are still developing the right technology to make the software compelling and workable.

TOPIC MAPPING COMES OF AGE

Topic mapping became a buzzword at online conferences during the past year, along with the Semantic Web. Both of these are somewhat ambiguous concepts that involve a complex arrangement of metatagging and XML. An XML topic mapping organization now exists called Topicmaps.org [www.Topicmaps.org], which is "an independent consortium of parties interested in developing the applicability of the Topic Maps Paradigm (first fully described in the ISO/IEC 13250:2000 "Topic Maps" standard [ISO 13250], which is SGML and HyTime-based) to the World Wide Web." Ontopia's CEO, Steve Pepper, is a founding member of the consortium and has written a white paper, which can be found on Ontopia's Web site, about the technology.

The early stages of topic mapping were primarily textual, with tree structures or lines linking one label to another. Some software continues with this theme, while graphical approaches in the new generation are almost "meta"-topic/concept maps. These maps are multidimensional and akin to creating a "PDF" of visual information universes. This is where Groxis is positioning itself.

GRAPHICAL INFORMATION VISUALIZATION

Some veteran searchers insist that graphical information visualization is a "waste of time" and that words and lists are quite adequate--until they see a demonstration, that is. We have grown accustomed to the page after page of lists resulting from searching using a Web search engine (such as Google, AlitheWeb, AltaVista) or a database that often includes redundant and repetitive information, one small clump of results at a time (available through Dialog, EBSCO-host, Factiva, ProQuest). In the case of databases, results can be selected or marked, but most Web search engines provide no way to gather up the good results and get rid of the redundant and unneeded ones.

For those who want powerful, text-based searching, metasearch engines such as Copernic [www.copernic.com] offer ways to classify and organize results into lists. Copernic was just named the Best Meta Search Engine of 2002.

It is difficult sometimes to pick a good word or concept to benchmark a new search engine or software. In the case of Groxis' Grokker, it is easier to see the beauty of visual searching and exploring by making a broad search with just one or two words. As an example, I open the downloaded software, and connect to the Internet.

I enter biomimicry in Grokker's search box and am immediately treated to an expanding array of spheres and circles of glowing colors (colors embed relevant information and characteristics) that keep moving as more categories and documents are retrieved and clustered. I find myself reading an essay on "Succession" in ecological theory by David Korten--while the map of colored balls continues to grow and organize itself into a map with labels such as chemistry, ecology; ecosystems, materials science, pesticides, pollution (see Figure 1).

I can read full text, launch a Web site, look at the categories being developed, and browse--all while the map continues to grow. I can mark "nodes" of interest to return to later, and I can change the label of the node itself. One half of the screen is devoted to the spheres, and in the other half I can ifiter (by date range, domain type, or text), browse, launch a Web site, or read a full-text article. I can close the map while I read an article, and the map keeps building behind the scenes. For me, this is analogous to seeing one hemisphere at a time, with movable boundaries, but that's because I'rm kind of a map nut anyway (see Figure 2).

VIEWING, SAVING, EDITING MAPS

When I like what the map has gathered and want to return to it later, I can save it or send it to someone else as an e-mail attachment. In the example of biomimicry, Grokker searched through the Northern Light search engine (one of the plug-ins for the Grokker software) and got results of 124 categories and 1,114 items. [Editor's Note: A spokesperson for Groxis confirmed that Northern Light remains a plug-in despite the bankruptcy of divine, inc., its parent company. She stressed, however, that Groxis is aware of the possible closure of Northern Light and has been working diligently to offer new and more powerful sources. Groxis intends to gradually phase out Northern Light starting with Grokker Preview Release 3.]

Grokker is not a search engine, even though it resembles one at first glance. By contrast, a metasearch engine that uses some interesting graphical mapping techniques is Kartoo [www.kartoo.com]. In the case of this visualization software, a few search engines are selected and a handful of categories are created. Kartoo has many nodes that you can click on and go to the Web site that the node represents (see Figure 3). Sponsored links are included in the nodes. There are big differences between what Kartoo does and what Grokker does.

In searching biomimicry on Kartoo, the software searched through a few search engines and developed 16 keywords/sites. Kartoo's design has garnered awards for its playful presentation using Flash technology; but it can only manage small result sets. It has also developed ways of saving and sending results. In Kartoo, the results are saved and e-mail is sent to the address you specify directly from the Kartoo server. The e-mail then has a link back to the Kartoo Web server with the search results.

In Groxis, the maps you create and edit are saved to your hard drive. You can directly e-mail the results (for example, I use Eudora and e-mail saved maps as attachments) to yourself or someone else. Then you can continue to work on and edit them.

Currently, the Grokker search will gather up to 10,000 nodes/documents into one map. You can delete unneeded material and save the final version with your own customization preferences (for example you can re-label spheres and nodes, and you can delete unwanted ones).

MAPPING YOUR HARD DRIVE

Another feature of Groxis: It will create topic maps of your hard drive. Imagine. Because this direction of mapping is software that you actually download to your own computer, it maps various "plug-ins" which currently include Northern Light, Amazon, and the Open Directory Project, in addition to the possibilities of your own hard drive. By the time this article is published, my guess is there will be several additional plug-ins (including search engines) that you will be able to search and then create maps for.

In the past year, I have become mesmerized by the potential of using the beauty and graphical qualities of mapping to create more holistic searches of information that provide connections and context to search queries. I think that this type of software will continue to grow, and that graphical representation of information will help us visualize topics in a more multidisciplinary way.

I have barely scratched the surface in editing and revising maps, but I find that in interdisciplinary areas, such as environmental research, it really shines. Searches on complex topics such as "What do the 2417 dredgers clearing the channel in San Francisco Bay do with the silt, and how can polluters be identified from analyzing silt?" flower into a systems approach to San Francisco Bay ecosystems and environmental research. There is something synergistic about having spheres with nodes and leaves. Seeing document types and categories with identified relationships is very exciting.

Maps, globes, and cartographic art have long fascinated me. Many years ago at a conference I gave a paper called "Mental Mapping in Academic Libraries." My love of traditional maps and mental maps, and the expertise that goes into creating them, is enhanced by this new breed of online topic mapping that gets to the heart of mapping metaphors and how we can use them to successfully enhance our knowledge and understanding.

MAPPING SOFTWARE PARTNERSHIPS

I find the connections and partnerships within this evolving field of topic mapping and information visualization to be quite interesting, much as one finds with following the paths and lineages of great jazz artists or photographers. For example, Entrieva has licensed Groxis for use in SemioMap and SemioTagger. LexisNexis Portal is powered by Plumtree, which is also a technology partner of Entrieva, along with Alta Vista, Antartica, Documentum, and Interwoven. Inxight counts COMTEX, Factiva, LexisNexis, Reuters, West Group, and Highwire Press among its publishing industry customers.

Often these partnerships are behind the scenes. Key Groxis development emerged from Metacode, an earlier generation of taxonomy mapping software (Metacode's Metasaurus). Although Metacode was purchased by Interwoven, the mapping software went to Paul Hawken, Groxis' chairman and co-founder, and Jean-Michel Decombe, chief design and technology officer and another co-founder.

Vivisimo [www.vivisimo.com] is a clustering type of search software that has created interest in the past year. Some publishers, such as the Institute of Physics and Highwire Press, have partnered with Vivisimo to enhance the searching options of their full-text electronic journals. Highwire also uses Entrieva to categorize its collection. You can use Vivisimo as a search engine as well, but when I entered the example of biomimicry, it returned only 149 documents.

Organizations can run topic mapping software on their intranets to help provide relationships and contextual data for overlapping areas within their businesses. For example, the human resources aspect of creating databases of staff with overlapping skill sets that can optimize working relationships is possible with this type of searching.

Applications for mapping and organizing other forms of content outside of search engines and desktops are emerging. Being able to explore, map, organize, and retrieve large data sets from environmental, legal, science, and other types of databases, as well as huge global information complexes such as healthcare networks and financial markets, is exciting the imagination of developers, executives, and information specialists.

VISIONARY NEW GENERATION SEARCHING

New generations of online searchers, both professional searchers and end-user searchers, will include younger people who have grown up with graphical, zooming interfaces as part of their daily lives. These searchers are at home with digital games, and the way they learn and create will be affected by their intuitive grasp of the context of what they are doing online.

Creating new software for the coming generations of searchers is a stimulating challenge.

For some of this new software, looking at who is on their management team and their advisory boards is revealing of how their vision is being shaped. In the case of Groxis, a stellar list is included in its management team: Jean-Michel Decombe, R.J. Pittman, Paul Hawken. The list of their directors and advisory board includes John Seely Brown, Lisa Gansky, Paul Saffo, Peter Senge, Stewart Brand, and Kevin Kelly, among others. All of these individuals are farsighted visionaries in their own fields.

John Seely Brown, who recently joined Groxis as a member of the advisory board, calls this new software "GII" for Graphical Informational Interface. "Finally, a visualization interface that has elegance, coherence, and compositional simplicity, all at the same time," says Brown. "I expect this to be the cutting-edge technology that moves us from the ubiquitous GUI (graphical user interface) to a new paradigm-one we might call a Gil, or graphical informational interface-that will let us play with and find information in a way that honors the way our minds work."

The language of Grokker, with its "grok" implication referring to deeper levels of conceptual understanding, is finding its way as a verb into the global informal commons (including blogs and discussion lists) as people refer to "grokking" something--many levels above the use of "google" as a verb.

This graphical topic mapping is a field that is moving fast--by the time this article comes out, there will probably be a whole new set of ideas and software developments to marvel at. Flash and XML (and GXML) have brought us quite a way from the olden times of HTML. I look forward to watching new mapping paradigms emerge and mesh colorfully with our linear, textual world.

Isabel A. Stirling

[istirlin@library.berkeley.edu] is associate university librarian, University of California, Berkeley.

Comments? E-mail letters to the editor to marydee@xmission.com.

COPYRIGHT 2003 Online, Inc. in association with The Gale Group and LookSmart.
COPYRIGHT 2003 Gale Group