The Challenges Facing Computer Science for the Next Ten Years

Download Report

Transcript The Challenges Facing Computer Science for the Next Ten Years

The Challenges Facing Computer
Science for the Next Ten Years
SCI 110
December 2, 2002
Dr. H.E. Dunsmore
Department of Computer Science
Ubiquitous Computing
"Ubiquitous" -- Existing or being everywhere, in all places at the
same time
We used to design software to run only on computers
Now software used in cell phones, microwave ovens, tv sets,
automobiles, photocopy machines, cameras, satellites, ....
Software for ubiquitous computing must be small, quick, robust,
error-free
Mobile Computing
Internet connectivity via Personal Digital
Assistants (PDAs), laptops, palmtops,
fingertops, cell phones, hearing aids, "glasses",
....
Challenges -Handling traffic of information flow in ether
Handling volume of information to servers
Providing up-to-date information
Data Compression, Bandwidth
Expansion
Every time we increase processing speeds and
transfer speeds, demand increases
Must find ways to do astounding job of...
Data compression -- Can I get a 75 MByte file down
to 150 bytes?
Bandwidth expansion -- Can I move data at
1,000,000 times today's speeds?
Programming vs. Creating
Systems
With volume of software necessary in future,
there will not be enough programmers
Analogy - telephone system made us all
telephone operators
We can all become programmers via high-level
software creation systems ... that will require
brilliant computer scientists to develop
Human-Computer Interaction
Windows, icons represent significant step
forward in human-computer interaction
Much more needs to be done -- natural
language commands, decision support
systems, expert systems, learning systems
Visualization and Graphics
Visualization systems are outgrowth of
computer games
Visualization systems can be used for training
pilots, doctors, dentists, large equipment
operators, ....
Visualization systems can also be used to slow
down fast time and speed up slow time
Bioinformatics, Computational
Biology
Progress has been made in this area (human
genome project, cloning) using sophisticated
software systems
Future progress will depend on databases,
blazing fast processors, better algorithms, ....
Databases -- Massive Data
Handling, Data Mining
Now possible to collect massive databases
Store and retrieve information
Correlate information
Sophisticated data mining
What have you done every minute of today? Where
are you now? What are you doing?
Security
Information in files and in transit must be
secure
Must develop better encryption systems
Must be very careful whenever data is not
encrypted
Terrorism, Vandalism
Need better mechanisms than accounts, passwords for access
control
Biometric access control (fingerprints, hand readers, facial
biometrics) shows promise
Terrorists might shut down Internet, interrupt e-commerce,
corrupt information (like account balances, charges), ....
What about turning off security mechanisms via network,
shutting down hospital monitoring devices, opening prison
gates,....?
Must make it impossible to deliver and execute viruses
Computer Science Educators
Computer Science been around since approximately
1970
30 years later many original CS faculty reaching
retirement age
Many faculty who started in 80s and 90s will retire in
next 5-20 years
Many students have opted for non-academic careers
Pipeline running dry
Educational Paradigm
Old educational paradigm: 16-week semester,
classes, labs, recitations, students in lock-step
Maybe this paradigm needs to be modified
Replace classes with digitized "explanations"
Replace labs, recitations with "chat groups"
Dump 16-week semester in favor of topic modules
Educational Paradigm (2)
Give students as much time as needed for
each module -- 1 week or 10 weeks
How do we give exams in this model?
Purdue CS student could take modules from
Stanford, Illinois, Carnegie-Mellon and students
from those schools could take Purdue modules
Administrative nightmare
Women in Computer Science
In mid-1980s women constituted 30% of the
professionals and students in Computer Science
That percentage now down to about 10%
Why?
Nature of the men in CS
Perceived long hours, stress, lifestyle
Women have longer-term, grander-scale perspectives
than men
Managing Email
Today's email systems will be remembered as the
dinosaurs of the email age
Manage the volume, respond using email triage,
delete un-needed messages, file needed messages in
retrievable place
Email all sorts of information (including audio, video,
documents)
Must get rid of (or control better) spam