"Too Much Information," Feature Article, June 2003

too much information

Organizing information—after gathering it in the first place—is the key to actually using it.

by Jean Thilmany, Associate Editor

As anyone with a Web browser and a little Internet-surfing experience knows, saying there's a lot of information floating around out there is like saying the ocean is a little bit wet. Tracking down the specifics of any given topic at a moment's notice—whether it's the ingredients for a chocolate velvet cake or Ronald Reagan's middle name—the early 21st century's birthright is the ability to find out just about anything without venturing too far from a personal computer.

But, as mavens of the Google search engine already realize, the unprecedented levels of information access can easily overwhelm. Does anyone really need to know within two minutes how many episodes of the television show Doogie Howser, M.D. were ever produced? (Ninety-seven episodes; 20 seconds to track down the number.)

Engineers aren't exempt from the need to walk a fine line between information access and information overload. Whether they think about it or not, they create reams of new information each day as they work on, analyze, test, and possibly discard designs.

In this Web-linked, data-rich world, engineers, like other professionals, must organize for easy access both the growing amount of information they create during the course of their day and the information they need to do their jobs. Indexing and linking documents and other information sources is an important first step, but engineers also must have a way to organize the information so they can find it via an easy search.

Some software systems can do just that. And, in the future, they'll be able to go much farther—to create a formal way for engineers to record their thought processes and early designs in a database, so that other engineers working on a similar project or tweaking a design can see what the original engineers were thinking. With the technology's help, engineers, in essence, could pick the brains of the original designers.

For an example of when such a system might be useful, look to the February crash of the Space Shuttle Columbia. Some investigators link the craft's external tiles to the cause of the crash. Investigators and NASA engineers use past design information to see what kinds of analyses were done on the tiles and the results. They're also looking at whether original engineers had predicted a scenario wherein the tiles might be dislodged at takeoff.

NASA already uses a host of ways to retain original engineering knowledge, simply because of the magnitude of its space program and the time it takes to ready a craft like the Space Shuttle. All that information can't be lost.

PICKING THE BRAINS OF THE PAST

The technical term for a method of digitizing the design process, including the brainpower involved, is called knowledge capture. Making sure the information is formatted and accessible is another can of worms called knowledge management. Managing information—figuring out how to best archive and retrieve it—is a relatively new technology field. It uses software to identify, capture, retrieve, share, and evaluate information such as databases, documents, policies, and procedures as well as the expertise and experience stored in individual designers' heads, according to Jose Luis Hernandez Martir, a professor in the electrical and computer engineering department at the University of Puerto Rico in Mayaguez.

As part of a research project, Martir built a database filled with already-solved design problems that included a step-by-step outline of how they were solved. He wanted to see if designers could look to the database when trying to find answers to their own design problems. Why reinvent the wheel if another engineer has already struggled with a similar design problem and found a way through? To standardize ideas, he used a software mapping system called concept maps, which offers a way to visually represent philosophical arguments, thought processes, and abstract ideas. Detailing a way of thinking through a concept map makes the thought process explicit and visual so it is easily examined and shared, Martir said.

A software service called Knovel organizes information from reference textbooks such as these, and makes it searchable on the Web.

"Concept maps can be used to collect a designer's knowledge and cognitive skills not available to other designers," Martir said. "This is the kind of knowledge and design experience not found in books and other written sources."

The concept map explains through a formal step-by-step visual representation how an engineer attacked a design problem, struggled with it, attacked it from another angle, and eventually solved it. The maps are archived in a database and can be referred to for help with similar problems, he added.
Engineers who search on past concept maps can find help that may guide them through complex problems like designing a robot or building a logic circuit or a myriad of other engineering challenges, Martir said.

FROM ONE CASE TO ANOTHER

The type of knowledge capture system that uses Martir's research as its backbone depends on case-based reasoning—the ability to look for answers to a current prob-
lem in past solutions. It's a useful way to drive software
because case-based reasoning mirrors the way humans think and solve problems.

"Like most humans when faced with a new problem, a case-based system compares the current problem with cases encountered in the past to see if one of the earlier experiences can provide a solution," Martir said.

A not-the-same-but-close case, when solved, is added to the case base for future reference, he added.

But case-based reasoning isn't the only way of archiving a designer's reasoning. Researchers are also looking at ways to record an engineer's thought process via computer as he or she designs a part, according to Heather Richter and Gregory Abowd, researchers in the College of Computing at the Georgia Institute of Technology in Atlanta.

A large amount of information generated during design—sometimes called design rationale—doesn't get recorded, which is a potential loss of important design knowledge, Richter said. Researchers have found ways to capture some of this extensive design information, but the attempts haven't become common practice, Richter said.

The usual ways of recording design rationale—asking engineers to record their thoughts as they design and include them in a database—are too time-consuming for designers. And simple Word documents can't really be standardized and searched. Plus, they don't really convey the thought power behind the design in all its complexity.

Richter said that engineers shouldn't have to think about recording their design rationale anyway. It should happen invisibly. Having engineers stop and write down their thoughts every so often creates a design meta layer, a sort of thinking about thinking, which calls attention to itself and interferes with thinking about the design—more or less the equivalent of being told not to conjure up a white elephant. Plus, engineers have enough to do already without worrying about whether or not they're recording their design thoughts.

Also, original parts are seldom designed by a single engineer who sits for hours in front of a computer. In reality, engineering design often takes place around a virtual water cooler as engineers collaborate and discuss their ideas and design changes back and forth via e-mail, and pass alterations among themselves.

Richter and Abowd experimented by videotaping about six hours of conversations between collaborating engineers and archiving the tape on a computer. Anyone wanting to view a portion of the video could click on keywords in a summary document that annotated the video timeline. The researchers are still looking at how the method of recording designers might be useful for much larger projects.

To archive design information for future reference, many engineering companies now rely on information like e-mail stored in product data management systems, which are used when engineers collaborate over long distances. Or they rely on a variety of other methods like spreadsheets and data sheets. But, more and more sophisticated knowledge-capture software like the application Richter and Abowd are studying is becoming available every day.

DIGITAL TEXTBOOKS

Of course, part of the abundance-of-information challenge is simply organizing it. Engineering libraries have long faced this challenge. The Kurt F. Wendt Engineering Library at the University of Wisconsin recently purchased a software service called Knovel from the company of the same name in Norwich, N.Y. It aggregates more than 450 engineering and scientific reference books, includes their complete texts in digital form, and makes the contents searchable, said Diana Wheeler, reference and instruction librarian at the library in Madison, Wis.

"If you consider what it's like to stand in front of several shelves of books in a reference section and wonder where you'll find a particular piece of information, Knovel allows you to see inside all those books in an instant and target what you're looking for," Wheeler said.

Books included in the Knovel database include popular reference texts for industries like aerospace, radar technology, chemistry, chemical engineering, food sciences, mechanical engineering, safety, semiconductors, and electronics. Pertinent mechanical engineering books include Finite Element Method, published by Butterworth-Henemann; Peterson's Stress Concentration Factors, published by John Wiley and Sons; and Strength of Materials and Structures, also by Wiley and Sons. Knovel has agreements with many reference-book publishers.

Engineers interact with data in ways that they can't with printed information. They can enter pertinent numbers into equations that would otherwise be static on the page and calculate answers.

"During the training, students started playing around with graphs and tables right away," Wheeler said. "One was pleased because she could pull out a couple of rows of data from the textbooks and juxtapose them. If you're just looking at a table in a book, you have to take it the way it is. This gives you a new way of looking at information, so you can tweak it and make connections and comparisons more easily."

At PPG Industries, the coatings, stains, sealants, and glass maker headquartered in Pittsburgh, Denise Callihan, manager of the PPG research and development libraries, looks to a day when internal information might be accessible and searchable in the same way the company uses Knovel to search reference books. The PPG libraries have used Knovel for two years, Callihan said.

"We try to manage our internal knowledge as best we can, and it is text searchable," she said. "We do a lot with our internal report information, but with anything generated in PPG, we don't have the ability to search through documents right now as we can with Knovel. But we're working on that all the time," Callihan added.

PPG chemists use Knovel when creating the material safety data sheets that outline boiling points, chemical makeup, and other significant features of the chemical compounds used to make coatings.

"Chemists can just pull boiling points and water pressures from the reference books at Knovel, whereas before they'd have to look around through books or ask us to do a search for them, and then wait for us to actually get it done," Callihan said.

Other research and development library tools at PPG include MicroPatent, which includes searchable patent information; the Encyclopedia of Materials: Science and Technology from Elsevier Science; and SciFinder, a method of searching and browsing scientific journals developed by CAS of Columbus, Ohio.

"There's tons of information out there, so we try to select which pieces would best serve the people who use our libraries," Callihan said.

PPG employs about 1,000 researchers and developers, and all 33,000 PPG employees have access to the research and development libraries. Callihan and her staff try to include Internet or computer-based information services in the mix of library tools, so that PPG staff in any location can easily access digitized information, she said.

The amount of information that engineers can access within minutes can be both a blessing and a curse. But as librarians and researchers know, the way to tame the information is the same way you tame clutter in your home: constantly tossing unused items and organizing the rest.