DOD'S HPCMP director loves the cutting edge

Kay Howell director of the Defense Department's High Performance Computing Modernization Program (HPCMP) has shaped her career around her love of cutting-edge technology and an appreciation of what users want from it. Howell took her first job out of college as a computer specialist with the Naval Air Systems Command (Navair) because the agency was installing its first VAX 780 Digital Equipment Corp.'s 32-bit minicomputer. A few years later in 1986 she started working with the world's fastest computers at the Naval Research Laboratory (NRL) because she thought the Cray X-MP a supercomputer made by Cray Research Inc. "was a hot machine."

"I wanted to work with that box " she said. "I thought I was one of the luckiest people in the world."

Now Howell is in charge of assembling the computing infrastructure that supports the military's top scientists - systems that make possible complex simulations needed to produce better weapons more reliable communications and improved troop deployments.

But it is the users and their research more than the machines themselves that Howell thinks about daily. She said she has begun recognizing many cases in which automation can act as a barrier to problem-solving because the systems require so much effort to use.

"You don't own these systems for your own benefit " she said. "You own them to empower the end user. It's important never to lose focus of that."

She has reached that practical conclusion after years of working closely with people who were not computer scientists. Before obtaining her current job in 1996 she spent seven years with the NRL's user services branch and three years as the director of the lab's Center for Computational Science where she headed a user group that advised her predecessor in the HPCMP office.

Howell says she did not plan a career in government or information technology and ended up as a manager because colleagues encouraged her. Growing up in the small town of Kinston N.C. her neighbors did not think much of the federal government and the only federal worker she knew was the postman.

She entered East Carolina University as an accounting major but switched to computer science after taking a computer course designed for math students. She signed up for the class because the computer courses for business students were full and she did not realize how different the classes would be.

"I just fell in love with [computers]" she said. "They were nonjudgmental. They didn't care about your gender or your physical appearance. If you're willing to learn the rules they treat you very fairly."

A Visit Begets a Career

In need of some practical work experience before she graduated she chose an internship with Navair in Washington D.C. because she wanted to visit the nation's capital. She then considered a public-sector career for the first time because she "was so impressed [by] how smart people were there and how hard they worked " she recalled.

Navair hired her in 1982. After moving to the NRL she eventually became a manager because she found she was "was good with users " and they viewed her as an ally she said.

DOD will spend nearly $150 million on high-performance computing this year. Howell said investments in the technology over the years have produced dividends for commercial users as well as for government. Private companies are for example buying software modeling tools for data mining and other applications that were not available a few years ago.

"If we'd not been pushing it with vigorous appetites it would be much slower in happening " Howell said.

There have been practical benefits of this advanced research she said. For example when U.S. peacekeepers in the former Yugoslavia needed to cross the flooding Sava River into Bosnia in 1995 the Army Corps of Engineers already had an application it could use to collect weather data from the field model the height of the river and send the results back to commanders who used it to decide when to move forward.

"When the groundwork for this was laid it was all cutting-edge technology " Howell said. "They were so far along they were able to go out there and use it."

She already is looking toward what she thinks is the next big computing challenge: developing tools that enable users from different disciplines to collaborate on scientific problems. "Few problems we're trying to solve anymore are just physics or just chemistry " Howell said. "We're going to solve some very important science problems that just a short while ago we never thought we'd be able to tackle."