A collaboration between
the computer science department and University Libraries has put
students to work on real, state-of-the-art software challenges with
extraordinarily successful resuts.
In one case, the work done by the
students in the graduate-level software engineering class of
computer science professor Barry Boehm will almost certainly
go into service at the library later this year. And the work
done on others will be a boon to libra ry fund raisers in coming
months.
Stereoscope cards, medieval manuscripts,
antique maps, archived planning studies, clunky government databases,
movie stills, student films, videotaped engineering lectures,
biological photographic images, Latin American pamphlets and
more - students have produced rough drafts of multimedia solutions
for these hard-to-manage collections, at essentially no out-of-pocket
cost to the library.
The students, who in previous years
in the course chafed at assignments requiring them to impersonate
client librarians or fire department dispatchers, have been
delighted to have a chance to work on real-world issues. Boehm
has had a chance to develop hi s ideas on methods to simplify
and rationalize the creation of software for specific users.
Meanwhile, the library has received
services that would have been prohibitively expensive on the
open market.
Photo- IRENE FERTIK
Barry
Boehm, professor of computer science. Students in his graduate-level
software engineering class are working on innovative projects
for the library.
Stereoscopic Card Collection
The university has an extensive collection
of more than 2,000 stereoscope images that originally belonged
to the C.B. DeMille studio library. The pictures, taken before
World War I, were originally marketed as a parlor amusement
during pre-television days : Edwardian families could sit at
home and see 3-D pictures of Paris or the Panama Canal.
At the dawn of the 21st century,
these images are potential treasures for scholarly researchers,
illustrators, historians and even the public. And new multimedia
web techniques could make them more accessible than ever - even
presenting them in stereo to remote viewers.
But John Ahouse, curator of the collection,
wondered what would be the best way to catalog and present such
images in multimedia form. What would be affordable, in terms
of library resources, for digitizing images and entering data?
Similar challenges faced each one
of the 13 teams that Boehm assembled, and the resulting student
proposals drew raves from library officials.
"Every one of the 13 projects succeeded,
at least to some extent," said Denise Bedford, the special projects
librarian who, with science and engineering librarian Julie
Kwan, initiated the project. "At this point, I think it can
be called an unqualified s uccess," Bedford said.
"Every single project that came out
has fund-raising possibilities," said Kwan, who had been the
first to start discussions with the computer science professor.
"This collaborative process has significant
potential for everyone involved," said Jerry Campbell, dean
of university libraries and chief information officer.
Risks were involved, Boehm acknowledged.
"When we went in, there was a lot of apprehension on both sides
that we would waste each other's time. I was really extremely
happy with how it came out."
Photo- IRENE FERTIK
Six
of the librarians who took part in the computer science/library
project with a stereoscopic holder and pictures that will
become part of the Internet information project. Front, from
left: John Ahouse, Julie Kwan, Barbara Robinson and Jean Crampon;
back row, from left: Ruth Wallach, Charles Phelps and Dace
Taube.
How Collaboration Began
Boehm, a member of the national academy
of engineering and internationally known expert on software,
is holder of the TRW Professorship in Software Engineering and
director of the Center for Software Engineering. He teaches
CS 577, a four-credit course in software engineering that consistently
attracts more than 100 students at a time, the largest enrollment
of any graduate course in the department.
The germ of the collaboration occurred
when Kwan learned that Boehm was having his students develop,
as exercises, database access systems that were extremely similar
to the one the library was using to update its venerable HOMER
electronic card catalog. In fact, Boehm's students were practicing
their software design skills on a simulated library system,
with students not-very-convincingly playing the roles of client
librarians.
The conversion of HOMER to the new
SIRSI system, a proprietary product produced by a company in
Huntsville, Ala., was well underway. (HOMER actually changed
over to its new, multimedia "Unicorn" interface system early
this year.) But, Kwan knew, the libra ry had dozens of special
collections of art, video, photographs, sound recordings, public
reports and other hard-to-manage materials.
It was Bedford who came up with a
concrete proposal. What if Boehm's class could use as raw material
the actual collections that were candidates for digital delivery
of multimedia materials, drawing on the expertise of professional
archivists and libraria ns?
Boehm was extremely re- ceptive;
he had, in fact, been hearing from students that the exercises
assigned as classwork weren't sufficiently realistic for fully
effective education. This would be a chance for his students
not only to organize challenging da ta, but also to work with
real customers.
"What we're trying to do in software
engineering is help the students understand some of the softer
aspects of the engineering system - such as the people aspects
and the economic aspects," Boehm said. Having the students deal
with actual clients, not com puter specialists, was a "tremendous
learning experience for them," he said.
Boehm and computer science department
chairman Ellis Horowitz quickly began to meet with Kwan, Bedford
and other library staff.
An August memo by Boehm set forth
the situation. The proposal was to create between 15 and 20
teams of six students each. Each team would work with a librarian
on a specific collection presenting unusual access problems
to create a software architecture p ackage for the desired capability.
Boehm said that this would take a
"large amount" of library staff time, and testing and bringing
up to speed new software would also use library computer resources.
And, he admitted, some of the solutions brought forth by computer
science students might be "off target." But, he noted, "With
15 to 20 projects, even if some fail, there will be a good number
of useful solutions."
Against the downside, Boehm laid
out the other benefits:
Above all, the collaboration would
give the library the luxury of exploring possibilities. As Bedford
pointed out, "To have a professional even look at the software
problems of many of these special collections is tens of thousands
of dollars or even more ."
Just the process, moreover, of having
librarians talk to student software engineers was doubly beneficial.
The librarians would learn about the parameters of what is possible
with state-of-the-art software designs. The students would gain
expertise in a s pecific area of software design. Additionally,
the project would "create a talent pool of USC-trained software
engineers with interest in and familiarity with library applications,"
Boehm said.
Launching the Partnership
With trepidation laced with optimism,
the project designers went forward. The first hurdle was to
see if the curators of the libraries' special collections would
be willing to invest the necessary time working with students.
A recruitment plea went out by e-mail
- and the response came back. Eleven librarians - Kwan and Ahouse,
Jean Crampon, Ken Klein, Robert Labaree, Sandra Joy Lee, Charles
Phelps, Barbara Robinson, Caroline Sisneros, Dace Taube and
Ruth Wallach - agreed to work with the students. Some volunteered
to double their time commitment by working with two teams of
students on separate projects.
"The participation of library people
was crucial," Kwan said, "and I think I can say without exaggeration
that every one who worked to prepare a project genuinely cared
about their project."
The projects illustrate the change
that multimedia has brought to library sci- ence, and highlight
the variety and richness of collections at the university.
"Librarians," Bedford said, "traditionally
are comfortable dealing with formal, secondary information,
final published compilation, the finished result, two levels
up from primary research materials - lab books, analytical data,
readings, photographs and the like."
Libraries have traditionally maintained
archives of these materials, but for researchers "they are extremely
inaccessible," Bedford said. "And not just here ... During the
project, I got on the phone with the Smithsonian. They had the
same problem we do."
Multimedia technology offers extraordinary
possibilities for solving some of these problems. For example:
archivist Barbara Robinson at the Boeckmann Center maintains
a remarkable collection of Latin American pamphlets, which contain
valuable information for specialists but which are old and fragile.
New optical character recognition
technology makes it possible to put the contents of the pamphlets
on-line. Scanning techniques can reproduce the covers. The entire
collection can be put on the net for researchers anywhere in
the world - and computer sci ence students created a working
demonstration of how it could be done.
Such a demonstration is critical
for library fund-raising for special projects. As Bedford says,
"A fair amount of money is available from various sources for
library archiving. But if you go and ask for the money, they
will tell you, 'Show me how you can do this, show me a proof
of concept.'
"What this project has done is give
the library 13 proofs of concept, which we can - and will -
take to funding agencies."
Project to Go On-Line
Of the 13, the project on the fastest
track is one developed at the Crocker Business Library by students
working with Caroline Sisneros.
Almost all business students routinely
must access the so-called Edgar Database of Corporate information
compiled by the U.S. Department of Commerce and available as
text (ASCII) files - but which typically consists of tabular
information that must be ref ormatted.
Crocker librarian Sisneros worked
with two separate teams to find ways to let students access
Edgar data more quickly and eliminate time spent manicuring
garbled files - and time spent by Crocker staff teaching students
how to do so.
One of the solutions produced by
the class was more than promising: the Crocker is planning to
implement it this year. During the second, spring semester of
CS 577, students are finishing the prototype, adding refinements
and enhancements.
"I think it worked out quite well,"
Sisneros said. "I enjoyed working with students, who gave me
a different perspective on my problem. I think it was nice for
them, too, to examine a problem outside their usual realm. I'd
have no hesitation getting invo lved with this project again."
Bedford adds one other advantage
for library personnel in working with students: "The semester
time-frame was wonderful. What slips in system development in
the commercial world is delivery date: programmers can almost
never deliver things by the date the y say. But these students
had to finish to get grades - they were motivated."
Boehm singled out teaching assistant
Alex Eygyed as a crucial element in the mix. "He was always
there to bridge the gap, to answer questions, to do what had
to be done. I think he did a spectacular job."
Students, too, found the process
rewarding. Egor Elagin worked on accessing the cinema-television
moving-image archive. "I learned that the computer sometimes
is not the best answer. My conviction before the class was,
digitize the whole world and dump in on the web - what's the
problem? But I found, for example, why it doesn't make sense
to digitize the movies they have at the archive and why they
have to spend money on a huge storage with fancy climate control
to keep the rolls of film instead of videot apes.
"I believe our project, once completely
finished, will be a great help to raise the money for the archive,
because it will be an advertisement on the web," Elagin said.
"It was a marvelous idea to have us help the school and do something
useful. It was le ss technical than most of the classes, at
least for me, but I don't regret that I took it.
A "Win-Win" Solution
Of the 13 projects completed in the
fall 1996 semester, six are continuing in intensive development
in spring 1997. Boehm said he doesn't see any reason why the
collaboration won't continue next year. "We've built up a strong
relationship with the library people, with Denise Bedford and
Julie Kwan, who have additional possible projects they've already
mentioned. This was definitely a win-win."
Ellis Horowitz, the department chair,
added his own perspective: "It is rare when a course has direct
impact on a community outside of the students who attend it.
I am pleased that one of our courses has managed to reach beyond
the confines of the classro om and has actually had an impact
on computer technology consumers. The learning experience for
the students was tremendously enhanced by their interaction
with the librarians."
Dean Campbell agrees: "These student
projects carried out in Dr. Boehm's class are exciting and important
- exciting because they bring the fresh perspective of a new
generation to the challenges of managing human knowledge; important
because our ability to succeed in this endeavor will be a key
ingredient in our continued intellectual growth in the next
century. Dr. Boehm and his students and the library faculty
members who worked with them are to be commended for creating
a learning environment where en gineering skills are immediately
applied to real problems."
