https://www.um.edu.mt/library/oar/handle/123456789/14671 Thu, 16 Apr 2026 19:24:43 GMT 2026-04-16T19:24:43Z https://www.um.edu.mt/library/oar/handle/123456789/100892 Title: Detecting and managing knowledge risk in software development processes Abstract: Software engineering is a knowledge-intensive activity. In the case of software organisations, the most valuable assets are no longer buildings and machines but the knowledge held by their employees. Software engineers are not merely vessels of technical knowledge but are fully fledged knowledge workers. They are expected to form a deep enough understanding of whatever domain they happen to be working in such that they are able to apply their technical knowledge to build solutions which solve problems in the domain. In an industry with high staff turnover rates, this can be worrying. Organisations need to ensure that as employees flow in and out of the their employ, the intellectual capital they create and work with is somehow retained and leveraged to increase their competitive edge. Studies have shown that projects do not tend to fail because of developers' lack of technical knowledge, but rather for reasons such as requirements failures, communication failures and estimation failures. These failures can be traced back to inadequate knowledge management practices as a root cause. Software development processes tend to address knowledge management issues by prescribing documentation, a strategy known to knowledge management practitioners as codification. However, codification is but one of the techniques in the knowledge worker's arsenal. This work seeks to make a contribution to software engineering by carrying out research into the fields of software development processes and knowledge management with the goal of designing and evaluating a new software development life cycle that makes knowledge management an integral part of software development. Although work has been done to evaluate the effects of various knowledge management initiatives (e.g. introducing wikis or mentoring programmes) on software development, the concept of a development process built around knowledge management has not been found in the literature. Such a life cycle should not only aim to produce a quality product within reasonable timeframes, but also seek to achieve a healthy organisational knowledge landscape. The approach being proposed here involves the design and evaluation of a software development life cycle which is based on a knowledge management strategy that is tailored to software engineering. The strategy in question is biased towards personalisation and makes use of techniques from the engineering and cartographic schools of thought in knowledge management. The overarching goal is to make knowledge management part of the focus of employees' day to day job such that they can identify and expose knowledge assets, and expose and monitor knowledge risks so as to allow mitigating action to be taken when necessary. Description: PH.D. Sat, 01 Jan 2011 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/100892 2011-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/95521 Title: Reliable file backup system over p2p architecture Abstract: Decentralized storage systems rely on redundancy schemes and repair policies to construct a reliable solution. Various applications make use of erasure codes to provide the required redundancy level due to the fact that simple replication of objects suffers from storage inefficiency. However, one drawback of erasure codes is the large amount of bandwidth required for repairs incurred due to data loss. To counter this effect various designs have been proposed, one of them being Hierarchical Codes which we believe has got the highest potential to achieve its goals. The only drawback of this coding scheme is the complex algorithm used to determine the probability of failure. This project objective is to understand better the parameters involved in this coding scheme with the final result being a new way of structuring Hierarchical Codes which eliminate the use of such algorithm. This is achieved by anticipating the number of losses and building the Hierarchical coding schemes accordingly to completely mask temporary failures. Description: B.Sc. IT (Hons)(Melit.) Sat, 01 Jan 2011 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/95521 2011-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/95515 Title: Implementing proof systems for the intuitionistic propositional logic Abstract: Logical reasoning is something human beings consistently utilise to make decisions. The simplest calculus for proof search is the Natural Deduction. However, it is harder for an automated system to automate proof search using this calculus. At every step it has to decide the next step on the basis of: • which inference rule to apply at every decomposition level? • what sub goals should we prove in our pursuit to prove the final goal? • if the proof search is in a stuck state should we backtrack the proof search or terminate it? Analysis shows that if the proof search is carried out using the Sequent Calculus the number of choices required at each step are reduced. Hence our approach to minimise the high level of non determinism in Natural Deduction is to compute the proof search in the Sequent Calculus then translate it to the Natural Deduction. The advantage of this approach is that although the final proof search is expressed in the Natural Deduction Calculus the actual proof search is computed in the Sequent Calculus which reduces non determinism during proof computation. Furthermore, for the points highlighted above proof search in the Sequent Calculus eliminates the possibility of backtracking and restricts the number of inference rules applicable for certain propositions. Another advantage of this approach is that since the proof search is carried out in the Sequent Calculus, if the proof search in the Sequent Calculus is correct, by the correspondence between the Sequent Calculus and the Natural Deduction we are guaranteed that the same proof translated to Natural Deduction is also correct. Description: B.Sc. IT (Hons)(Melit.) Sat, 01 Jan 2011 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/95515 2011-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/95507 Title: Predictive recommender for day planning Abstract: THE PLANNING ACTIVITY, like planning a day round a city or a shopping plan requires that the user that is planning has a considerate amount of knowledge about the environment that the human activity will take place in. For example a user must have knowledge about a city to plan a tour round this city, and on the other hand a user plans a better shopping plan ifs/he has knowledge about what shops are available together with their location amongst other. Predictive Recommender for Day Planning (PReDaP) aims to make the process of planning (which in this thesis is seen as a number of decision making activities) less hectic to its users, where it recommends best places to visit. Recommendations are given by PReDaP, according to the knowledge that the system contains about a specific User and his/her preferences together with the knowledge it has about the environment, where the planning is to take place. Description: B.Sc. IT (Hons)(Melit.) Sat, 01 Jan 2011 00:00:00 GMT https://www.um.edu.mt/library/oar/handle/123456789/95507 2011-01-01T00:00:00Z