The amazing evolution fuelled by the introduction of the computational element has already changed our lives and continues to do so. Initially, the fast advancement in hardware partially enabled an appreciation for software potency. This meant that engineers had to have a better command over this field that was crucial in the solution of current and future problems and requirements. However, software development has been reported as not adequate, or mature enough. Intelligence can help closing this gap. This chapter introduces the historical and modern aspects of software engineering within the artificial intelligence perspective. Also an illustrative example is included that demonstrates a rule-based approach for the development of fault management systems.
TopIntroduction
Since its earlier days the audience was attracted to the notion of ‘electronic brain’ that potentially could aid, replace, and even improve human-level intelligence. Expectations from such artificial intelligence, in improving the daily life, now had to be extended to aid the very techniques that would develop the artificial intelligence. Software Engineering could definitely benefit from such leverage while software demand has grown exponentially as ever and corresponding offer – i.e. software development, could never catch up with. Fortunately, before the end of its half a century long quest, after experimenting with improvised notions the inevitable incorporation of intelligence is in the process of being established.
It is desirable to replace the human in the software process, as much as possible. Linear improvements through better techniques but still targeting the addition of the next single block at a time, to build a huge system, are not sufficient. The development technique needs to be automated and automation is an open invitation to Artificial Intelligence (AI). Software engineers are under pressure for addressing organized ways for utilizing new approaches such as Component Orientation, Service Orientation, Aspect Orientation, Software Product Line Engineering, Model Driven Development, and related approaches. If not due to a conscious planning that is aware of the necessity and the opportunities, this consideration is at least being motivated by such technologies getting more and more popular. Such approaches assume the role of “enabling technologies” for the “Compositional Era”. Today’s software engineer should expect to deliver a system through locating and integrating sub-solutions in contrast to creating them. AI will help the engineer in automating the search and integration of the sub-solutions. Consequently, our methodologies that are aligned towards defining the code to be written and eventually writing it one line at a time, are changing towards locating and integrating blocks. AI will help the paradigm shift from code writing to composition.
Our History in the AI Perspective
Manifesting itself after the introduction of the modern computer, the older dated field of Computer Science (CS) surfaced with the goal to reduce the “semantic gap”. This was simply to convert problem representations from the conventional ways we developed for human understanding, to another, that computers could understand (execute). Figure 1 depicts the semantic gap concept.
This clearly states the main issue as that of Knowledge Representation (KR). KR in return, is a fundamental AI notion. In other words, the computation field is basically trying to enact AI. Among the many sub-fields only these that concern the higher-level regions of the semantic gap are somehow regarded as related to AI. One lowest level tool, covering the bottom region of the gap is the compiler. Neither a compiler, nor an embedded software that drives a dishwasher for example, are considered as artificially intelligent. When the behavior of the artifact mimics human intelligence, then it seems to earn the “intelligent” attribute.
However, when human-level intelligent activities are compared to the Central Processing Unit (CPU), the shier amount of intellectual activity does not seem to be sufficient in differentiating them on the basis of intelligence: It is easy for a human to recognize the face of another, among thousands of faces. While such a capability is very difficult to attain by the computers, multiplying millions of pairs of floating point numbers in a second is easy and yet that is unthinkable for the human brain. Yet the brain is intelligent and the CPU is not. Both require computational power. Repetitive execution of smaller granularity actions does not qualify as intelligence, unless they are organized around a structure to combine the massive calculations towards deducing a higher-level result such as recognizing a face. Otherwise, if even they required the same amount of operations, one-million independent floating point multiplications conducted within a second do not deserve the intelligence attribute.
The AI filed has gained and lost popularity few times in the relatively short history of CS. Nevertheless, there have been very satisfactory applications. Success of expert systems has been operationally proven even in the medical domain where doctors use them for diagnosis etc. If we were to make some approximation to accept all computer applications as Artificially Intelligent, ignoring the common perception for their identification as discussed above, it will not be wrong to state that AI is now being used to develop AI. Software Engineering discipline is now in an effort to utilize AI for organizing the way to build software. AI applications are software therefore we are in the era where AI is developing AI.