Real world evaluation of aspect-oriented software development : a thesis submitted in partial fulfilment of the requirements for the degree of Master of Science in Computer Science at Massey University, Palmerston North, New Zealand
Software development has improved over the past decade with the rise in the popularity of the Object-Oriented (OO) development approach. However, software projects continue to grow in complexity and continue to have alarmingly low rates of success. Aspect-Oriented Programming (AOP) is touted to be one solution to this software development problem. It shows promise of reducing programming complexity, making software more flexible and more amenable to change. The central concept introduced by AOP is the aspect. An aspect is used to modularise crosscutting concerns in a similar fashion to the way classes modularise business concerns. A crosscutting concern cannot be modularised in approaches such as OO because the code to realise the concern must be spread throughout the module (e.g. a tracing concent is implemented by adding code to every method in a system). AOP also introduces join points, pointcuts, and advice which are used with aspects to capture crosscutting concerns so they can be localised in a modular unit. OO took approximately 20 years to become a mainstream development approach. AOP was only invented in 1997. This project considers whether AOP is ready for commercial adoption. This requires analysis of the AOP implementations available, tool support, design processes, testing tools, standards, and support infrastructure. Only when AOP is evaluated across all these criteria can it be established whether it is ready to be used in commercial projects. Moreover, if companies are to invest time and money into adopting AOP, they must be aware of the benefits and risks associated with its adoption. This project attempts to quantify the potential benefits in adopting AOP, as well as identifying areas of risk. SolNet Solutions Ltd, an Information Technology (IT) company in Wellington, New Zealand, is used in this study as a target environment for integration of aspects into a commercial development process. SolNet is in the business of delivering large scale enterprise Java applications. To assist in this process they have developed a Common Services Architecture (CSA) containing components that can be reused to reduce risk and cost to clients. However, the CSA is complicated and SolNet have identified aspects as a potential solution to decrease the complexity. Aspects were found to bring substantial improvement to the Service Layer of SolNet. applications, including substantial reductions in complexity and size. This reduces the cost and time of development, as well as the risk associated with the projects. Moreover, the CSA was used in a more consistent fashion making the system easier to understand and maintain, and several crosscutting concerns were modularised as part of a reusable aspect library which could eventually form part of their CSA. It was found that AOP is approaching commercial readiness. However, more work is needed on defining standards for aspect languages and modelling of design elements. The current solutions in this area are commercially viable, but would greatly benefit from a standardised approach. Aspect systems can be difficult to test and the effect of the weaving process on Java serialisation requires further investigation.