Profiling is the term generally used for the creation and maintenance of profiles of skills. This is important, among other reasons, for employability: employers often want to know about the skills, particularly transferable ones, offered by potential employees, and they want evidence to back up claims of competence at a skill. Students may want to record the evidence they have for ability in these areas, which is not only in terms of qualifications, but could also be from testimonials, witness statements, or demonstrable products, gathered in the course of work or other experiences. Much of this is of relevance to writing CVs, but profiling as a whole goes beyond CVs. Students may want to know about how to develop their skills, or they might want to make plans for doing this. Even more broadly, they might want tools to help with planning their lives, while keeping skills and employability in mind.
These have become highly topical recently, particularly in the light of the Dearing Inquiry (e.g.
http://www.leeds.ac.uk/ncihe/) which states (recommendation 20 ):We recommend that institutions of higher education, over the medium term, develop a Progress File. The File should consist of two elements:
- a transcript recording student achievement which should follow a common format devised by institutions collectively through their representative bodies;
- a means by which students can monitor, build and reflect upon their personal development.
We are developing and testing a student profiling system, called LUSID, the Liverpool Universal Student Interactive Database. The process of developing this system has raised all kinds of more or less interesting and important questions and issues.
The questions and issues of importance, which need to be taken into consideration in the design of any student profiling system, include these.
There are a number of student profiling systems currently in existence, including
Both these approaches have implications in terms of the questions above.
There are arguments for the system being on the Web, unlike the above systems. Here are some reasons.
The overall system design needs to take into account the different classes of user.
A key aspect of this system is that it is conceived as highly flexible and customisable. Different universities will want different arrangements of the data, perhaps have their own definitions of skill areas, and will have different resources to offer. This may well mean that they want to write their own interface to the system, or at least to modify the original one easily. This requirement led us to the decision to segment the system between
In principle this is not unlike other systems that perform similar functions by embedding commands into web pages, such as Microsoft's Active Server Pages. However, for LUSID, we aimed at a language which was much better tuned to the needs of this particular application. We chose to do this in XML, and we are developing a DTD for our language as an extension to HTML.
Given that the main target user population is students, the interface to the users needs to be both reasonably simple and interesting. We decided that any compulsion beyond what was strictly necessary for the system to function would be unacceptable. The assumption is that users own their own data, and may actively give permission for other people to view it.
A major factor in our choice of technology has been the client systems that will be run by the users. In this case, we could not rely on Java, or even Javascript, as some machines might be running Windows 3.1 with Netscape 2: some users may even want to use Lynx as a browser. This means that we could not rely on any data processing being done at the client end, so we had to go with server-based processing. However, we wanted to reserve the capability for transferring functionality to the client side as that became available. This immediately suggested using Java as a development environment. The unknown quantity was the performance of the system.
LUSID has been developed in parallel on three sites: one running HP-UX, one Novell Netware, and one NT. Different databases have been used at the different sites. This has ensured that the system being developed has had to be easily portable to a variety of environments.
Data that needs to be stored in a profiling system can include
Whereas there is no insistent need for a relational database approach to this, Java includes the ability to interface with a wide variety of relational database management systems, and these provide the ability for future flexibility in for viewing the data in whatever way might be required. This seemed the best choice as a basis, rather than using flat files. A relational database imposes no onerous restriction on the way that data can be held, but also provides no help in deciding what data structures are necessary.
XML definitions are also used for other system-related data. This extensive use of XML is intended to facilitate future developments, in terms not only of XML editors, but also of portability and communicability between different installations.
The issue of database design is highly salient, being the interface between the real world and the computational system. The current state of Web-based CV databases reflects the fact that there are no generally agreed definitions of, for example, how to represent a qualification, or a period of work, much less a skill and an individual's level of skill. We have developed our definitions and structures from first principles, always remaining open to the actual or potential requirements of other systems with which we may want to interact in the future.
For a system such as this to have a long-term future, much will depend not only on how useful it is in the trial stages, and by the developing institutions, but also on how generalisable the various parts of the system are to include other related functionality. We are in a position where if we are able to develop standards for the representation of this data, related to personal skills, careers, and life, the possibility will open of working towards a more universal system where someone can easily and quickly transfer data between various institutions and installations. In this way we hope to provide the basis for a tool of very great coverage, much in the spirit of the Web itself.
We are developing a student profiling system which, on account of the use of the Web and other related technology, will be a candidate for rapid dissemination and widespread uptake, which will put more meaning and functionality into the Web for a range of purposes of importance to many people.