Testing Complex Applications - Software Testing: Testing Across the Entire Software Development Life Cycle - page 253

combination would be the remote testing of autonomic systems that are physically
inaccessible. Physical inaccessibility is used here to mean the system is operating in
an environment hostile to human beings. Two examples quickly come to mind.
The fi rst example is outer space such as earth orbiting satellites. Self-healing
communication of navigation or scientifi c satellites seem very desirable when con-
sidering the alternative cost of sending humans into orbit to repair a satellite. Some
of the software testing challenges would be suffi ciently valid testing before satellite
launch and a suffi ciently robust test monitoring after satellite launch.
The second example is inner space such as ocean going submersibles. Self-
healing exploration systems seem very desirable when considering the cost and time
required to bring a submersible vessel to the ocean surface from a depth of several
miles. Two of the software testing challenges would be suffi ciently valid testing be-
fore submersion and suffi ciently robust test monitoring during a dive.
One fi nal prediction arises out of a movie the late 1960s of “2001-
A Space Odyssey.” [53] One particularly interesting sequence shows the hero Dave
deactivating a series of memory cells in a runaway computer system named HAL
that control's Dave's spacecraft. As each successive memory module is deactivated,
HAL degenerates from a sophisticated, voice-activated, chess-playing, 3-D object
recognizing, and lip-reading mega computer to a nursery rhyme singing desktop
computer.
What challenges the imagination is the implication that HAL “learned” all his
supercomputer abilities starting from the simple ability to sing the tune “Daisy,
Daisy.” There are a number of articles in the Artifi cial Intelligence research arena
that propose how such “learning” might occur. No overtly successful efforts have
been reported to date. Showing a bit of optimism in the hardware and software devel-
opers' genius, we expect that “learning” computers will come into existence. At this
time, we can only wonder at the testing challenges posed by software that “learns.”
15.6 PUTTING FUTURE TESTING DIRECTIONS
IN PERSPECTIVE
Technology professionals always have a nagging question in the back of their mind,
“What is the useful lifetime of my current technical expertise?” Some technolo-
gies tend to grow and mature over time, offering the experienced professional a
long, prosperous career with appropriate continuing education. Other technologies
become a dead end because something leapfrogs them and becomes dominant in the
industry.
Software testing clearly falls in the former category of growth and maturity.
We see vast opportunity for basic software testing skills because the current and
foreseeable software development methods remain highly reliant on correct human
behavior. When (hopefully not if) software development methods truly mature, then
software testing professionals will have vast opportunity for developing more ad-
vanced software testing skills to match the challenge of new technology arenas.
Clearly, the software testing profession has a very promising future.