HyperTech's improvements make large map, database, photo and text files accessible to all computer platforms -- even when users are on location away from their offices.
ACCESSING LARGE AMOUNTS of different types of data -- map, database, aerial photo and text files -- can be difficult using a conventional database interface like the one on your home or office computer. Now imagine trying to access such data at a remote location in Asia, or via a laptop computer during a business flight to Europe. How do you access all the information you need?
HyperTech, a graphical hypertext interface being developed at the Georgia Institute of Technology with funding from the U.S. Air Force's Rome Laboratory, may put the information right at the user's fingertips. HyperTech accesses large amounts of diverse information -- but unlike many of its predecessors, it uses a standard, commercial relational database system called Sybase. Not only is Sybase the Air Force standard, it also is heavily relied upon in the business world, says senior research engineer Kirk Pennywitt of the Georgia Tech Research Institute (GTRI).
Susan Liebeskind (left), the technical director on the HyperTech project, leads programming design. She and Kirk Pennywitt, a principal investigator on the project, discuss their work.
"Although our system is based on a Sun UNIX platform, the data can be exported to HTML [Hypertext Markup Language], the language of the World Wide Web," says Pennywitt, who is HyperTech's project director and works in GTRI's Information Technology and Telecommunications Laboratory. "HyperTech can be read by people on other computers using World Wide Web browsers available for PCs, Macintoshes, Unix and every other major type of computer platform."
Hypertext is a way of displaying information differently from the usual word processing methods. Instead of writing and reading page by page, as one does with a book, hypertext allows a computer user to display and traverse information as logically related pieces. Dr. Jay David Bolter, a professor in Georgia Tech's School of Literature, Communication and Culture, provided most of the input on how the hypertext interface should work.
"Electronic links in a hypertext can deliver information to a user according to his or her immediate needs and interests," says Bolter, a principal investigator on the project. "Let's say the hypertext contains a political map of the United States and that demographic information is linked to the coordinates of each state or city on the map. By clicking on a particular city or state, the user receives exactly the information that he or she is seeking."
Working with Pennywitt and Bolter on the project are Susan Liebeskind, the technical director, who is leading programming design; Maria Hybinette and Phillip Hutto, automated linking coding; Janis Roberts and Eric Ayers, user interface coding; Stephen Arnold, database coding; and Kelly Johnson and Susan Hatcher, user documentation and usability testing.
THE RESEARCHERS DEVELOPED HyperTech to work with the Air Force's "Electronic Footlocker" concept of storing information digitally, ready for deployment to remote locations, Pennywitt says.
"So, for example, you have information on a foreign country and a situation arises there -- you extract all the pieces of relevant information from the Electronic Footlocker, put it on tape or CD-ROM, and ship it off with people when they head for the field," Pennywitt explains. "You might have information consisting of message traffic, maps, photographs, reference information and other items to store in that digital format."
HyperTech lets users create links between pieces of information where an association might not be obvious. A link, familiar to hypertext and World Wide Web users, is an underlined or highlighted word or symbol -- and by clicking on it using a mouse, the user is transported to a related document.
"You could have an article talking about the economics of certain products and that could relate back to the strategic importance of a certain country," Pennywitt said. "Basically the operator can traverse the information in non-linear fashion."
HyperTech will allow users to update databases while they are on location or in the field, so they are acting on the most current information possible. In addition, this particular interface avoids one of the main problems hypertext users encounter frequently: following so many paths within the information that they become lost.
"Our system provides a navigable map of all the data elements of the system," Pennywitt explains. "Pieces of information are represented as boxes, and the links are presented as arrows going in and out of them, so you can see which elements are linked to others, and how."
HyperTech also provides multi-user support so more than one person can access information at one time. Multimedia support for text, graphics, video and sound are standard features. The researchers are using off-the-shelf products whenever possible, Pennywitt says.
"That's easier for users to maintain, and is generally a more economical way of developing things," he explains. "Because it's based on a relational database system, we have a theoretical capacity of about two billion data elements. That is considerably greater than most other hypertext systems."
If those two billion data elements each consisted of a half-page of text, this would equal over a trillion characters, enough to fill 1,500 encyclopedias.
AMONG THE CHALLENGES the researchers faced in developing the interface was a lack of commercial tools for developing hypertext Unix-based programs. For example, they had to develop many components for the text editor to support embedded graphics, different fonts and style characteristics.
By March 1996, the researchers hope to make HyperTech capable of automatic link generation, solving the problem many users face -- time to create all the links they need between databases. They also plan to develop customizable views of the data. The user can filter the information presented to fit individual needs; for example, the user can limit the number of links presented by limiting linked documents to those "related to Somalia," "containing digitized photographs," or "larger than 32K," for example.
The researchers also plan to optimize the HyperTech's interface's performance based on user feedback and additional usability testing.
"Hypertext is a constantly evolving research area," Pennywitt says. "We would like to incorporate as many exciting and useful new developments into the program as possible."
Send all questions and comments to Webmaster@gtri.gatech.edu