| Robert Amor's Project Ideas |
|---|
I have listed a variety of areas that I have interests in. As you will gather, the common theme is information (or content) management (often on the Internet). Each of these areas could support projects at very different levels, from fully specified undergraduate projects (COMPSCI 380 or SOFTENG 401), to postgraduate projects (COMPSCI 780), through to Masters and Ph.D. theses. Please contact me (trebor@cs.auckland.ac.nz) if any of these ideas interest you, or to explore your interests in related areas.
I have broad interests in almost any IT project related to Architecture, Engineering and Construction (A/E/C) and also Facility Management (FM) areas. I have been researching in this area since 1985, and spent 5 years at the Building Research Establishment in the UK doing applied research on a wide variety of construction specific projects. The kinds of projects I'd be interested in include:
A large number of applications need to exchange information with other services as part of their service. However, the majority of services structure their information in differing formats (data structures). How can we describe and manage the mapping of this information between the varying structures? There are restricted solutions in RDBMS, XML, and even a developing ISO standard (EXPRESS-X). But no one seems to have defined what the scope of mappings are, and hence have no way of describing what types of mapping are required for different systems. Specific project areas include:
When a range of services are swapping overlapping information we reach a situation where the consistency of the information needs to be maintained. In RDBMS areas a transactional approach is used, but this is very restrictive and doesn't work for applications requiring long transaction cycles (e.g., 1-2 days of design analysis for a portion of a building system). Differing approaches can be developed in this areas, and the impact of different approaches calculated for particular types of applications.
For a search engine to be able to serve a user with pertinent information it needs to identify relevant web sites for a user's search. One way of doing this is to accurately classify the content of a web page or site. However, a human user couldn't possibly classify every web page or site on the Internet, or even within a country. Here we look at how to automatically classify a web page to a known classification system. This will give us the ability to identify the country a web service is serving (e.g., is this .com targeting a NZ audience), the type of product being sold (e.g., aluminum window manufacturer), the type of service offered (e.g., construction, installation, or demolition).
How can we support an individual user in providing information suited to their needs? There are many approaches to automated profiling of users based on their web habits, and occasionally they throw up some interesting information. Can we improve how users' information needs are identified, specified, and managed in specific domains, and in general.
Game engines provide very efficient 3D visualisation and low bandwidth immersive collaboartion between multiple participants (game players). How can we hijack this technology to do some useful work? With the School of Architecture we have been developing design critique tools for student building designs. Projects in this area range the gamut from: information visualisation within an immersive environment; interfaces to the environment; modification of designs within the environment; etc.