The work was meant to consist of three parts: Research, the Knowledge Base Web Portal and Software Development.
The project is research – oriented, with research activities conducted in such a way as to benefit both CNZ and us, the students doing the project. Keeping in mind the aim of the project: to help us cross over from academic learning to learning in the workplace, we conducted the research working as virtual staff members of the CNZ Research and Development Division.
We selected fields in the IT industry at the cutting edge of technological development like DDR2 memory, Windows 64 bit, Web Services and so on. We researched each one of them in depth, by focussing on different facets of that particular field: not only the technologies behind it, but how it works, front – applications, deployment issues and how it can be integrated with the existing systems in an organisation, in order to get a more complete perspective on the technology as a whole.
At the end of each research activity, we wrote reports detailing our findings. The reports followed a format specified by TN Chan. The reports had the following scope:
1. Vertical Scope of Research Activity
Ø Objective and definition of the technology
Ø Functionality, scalability, availability, security, ease of use
Information flow block diagram of related technologies (the big picture), coverage within and relevance to the BTECH syllabus
Ø Standards, approving organization, chronological information of endorsement, Maturity of standard and technology
2. Horizontal scope of Research
Ø Product comparison, such as Email Systems – Groupwise 6.5, Notes, Exchange
Ø Deployment information worldwide and in NZ
Ø Approaches used by System Integrators, especially in NZ. Find out their Service strengths, Organization weaknesses, Customers Profile and relationship.
3. Collation of Research Results
Ø All sources of information identified
Ø Credibility of source identified
Ø Every report version controlled by increment
Ø Only moderated reports become version 1 or above
Ø Latest version of reports will be catalogued in Knowledge Base
My first research activity was the Web services sector of the IT industry. I started off with J2EE, went on to the technologies behind web services: SOAP, WSDL, ebXML, and so on; and finally, I researched front-end applications of web services technologies. These reports can be found in their entirety here.
CNZ has a Knowledge Base that is a continuation of the work by 7 BTech students over the last two years. The Portal is a MySQL based database at the back-end and is accessed by a front-end, Apache based website using PHP.
Our work consists of establishing a Technology Scoping Document for each class of target customers. Next comes the identification of the state of each technology (related standards, who the suppliers are, and so on.) as well as a categorization of the existence of skills in the New Zealand IT industry, with a Knowledge Database of Technology Disciplines, in addition to a Database of Potential Customers, as the deliverables.
We will also be involved in improving the portal so that the scoping document can be expanded to show up to 2 levels of knowledge areas. It should also be able to search for sources of information and knowledge as well as register the key points and guidelines for each. The resulting Knowledge Base Web Portal will be easier to use, faster and more efficient than the one currently in existence.
Edmond Chan, who is in charge of the maintenance and the upgrading of the portal is currently busy working towards his Masters at the University of Auckland. As a result of this, he was not able to oversee our improving of the Portal, which is why it was shelved in favour of the Linux project for Jeff Zhu, Quality Assurance work for Fei Xu and Software Development work for myself.
The software-programming segment of the project is with AZUL Limited, the software division of CNZ. AZUL is the codename of a database-software that runs under Windows. It includes accounting, inventory control, customer relation’s management (CRM), after-sales service tracking, staff resource management and knowledge base, all in one unified format and window. The current approach is to develop a separate version of AZUL for each customer, with heavy customisations in each of the deployments, reflecting the individual specifications by each client. The move for the future is towards a single deployment of AZUL, with additional modules available, depending on what is required by a particular customer.
The current approach by the software architect is to develop a separate module of Azul for every customer because they needs vary. Each deployment has heavy customisations to meet the clients needs.
As the needs change, new versions of each module are developed and each version needs to be maintained. This is because Azul programmers need to have access to every version of the software in order to identify the source of the problems that arise and to fix them.
That's where we come in.
The Software Development section of the project has 2 parts, Quality Assurance and Crystal Reporting:
The QA work primarily consists of documenting the code behind AZUL, which is aimed at pinpointing any dependencies in it; and D-Unit testing, which involves analysing and testing the software to identify any shortcomings that it might have. Maintaining the different versions of AZUL under a versioning control system, CVS is also required. This is because the AZUL programmers need to have convenient access to every version of the software in order to pinpoint the source of any problems that arise in order to fix them.
To maintain the different versions of Azul properly, we're starting to check them into a version controlling system called CVS. My part is to learn CVS and formulate a set of draft procedures for its deployment in the organisation. This document will be made into the official CVS manual for use by the Azul programmers and is also a Quality Assurance requirement when Azul Limited applies for International Standards Certification in 2005.
The second part of my work is Delphi unit testing and involves analysing the software, module by module and documenting the subroutines that are used by each unit. Each of these subroutines then needs to be tested. This process is aimed at identifying any possible shortcomings in the code and also pinpointing the underlying dependencies of the code.
The final part of the Software Development part of the project involved Crystal Reporting. My work involved learning Crystal 9 and familiarising myself with the Azul databases at the same time. The next step was building new reports, as required. I was also required to simultaneously catalogue the existing Azul report base. This involved, first and foremost, developing a set of guidelines for the designing of Azul reports and building a template based on these guidelines. Then, a report naming convention also needed to be arrived at as the existing names did not intuitively tell you what the reports did.
After these rules, guidelines and templates were formulated, the existing reports database had to be modified so that each report conformed to them. Since the existing database had reports that were not catalogued properly, i.e., not organised under the proper categories, the underlying reports database did not reflect the way the reports were accessed via the Azul RMS. Each of these reports (530 in number), had to be opened and the connections to the database reforged. These reports were then summarised, converted from Crystal 7 to Crystal 9, their name and layout modified to conform to the new templates, and any reports that were duplicated were gotten rid of.
Finally, this section was meant to include modifications to the Azul RMS interface. The purpose of this activity was to improve the efficiency of the RMS as well as incorporate certain features requested by our clientele. Early in September, the Azul Software Development decided to change over from Crystal Reports to Delphi Report Builder as our report-writing tool. It was because of this that the final section of the project was deemed to be unnecessary and was not completed.