School of Computer Science

University of Birmingham

Edgbaston, Birmingham

B15 2TT, UK

In a modern University, traditional teaching methods are hard to sustain. There are an increased number of students from a wider spectrum of intellectual ability taking more courses. Educators have increasing pressures on their time to both prepare and present material, whilst institutions have resource limitations. To deliver the appropriate material, we require educational techniques that are student-centred, offering flexible learning, that are involving, and maximise the exposure of students to different approaches to information dissemination and presentation. The web offers itself as a medium, tool and platform to deliver such new material, and has seen increasing use over the past few years. However, to take advantage of this technology, we require new tools that can help us exploit it to the full. This paper explores the design and use of a few simple aids that all share the properties of providing dynamic information and content in the web environment to aid its use as an educational medium. The tools are not themselves complex since the focus has been on providing functionality as easily and as quickly as possible; however, they have proved themselves useful over the past few years of testing.

The web offers obvious opportunities to offer computer based questioning of students, particularly in the form of multiple choice questions, but the effort involved in creating such questions with any decent format is at best time-consuming, and at worst confusing and impractical. MCQGens aim is to resolve this problem, and is a Perl script that parses a plain text file with a simple set format. The text file takes the form of a question followed by arbitrary number of answers, coupled with notes on the answers to give more information to the student. There can be an arbitrary number of questions, answers, and notes. MCQGen produces a set of HTM pages, which are titled and contain the questions with the answers beneath. Each of the answers is hyperlinked to a separate page which relates to question, the chosen answer, and then provides the notes as well. A return link to the questions in the appropriate place is also given.

MCQGen makes it exceptionally easy to create web based questionnaires, and it can also be integrated with server software to score these multiple choice questions. However, we have not done this in an educational setting for a number of reasons. It is difficult to make such a system resilient to cheating, though are the primary reason for not marking such questions automatically is that this was not of the educational aim in the first place. The intention is to get students to play around with ideas and concepts and discover the primary features of the work they are doing by exploring their knowledge and feel for the subject. The system has had a lot of use over the past five years, making it easy for educators to create sets of multiple choice questions that the students can interact with in their own time. They are simple for the students to use, and feedback from them suggests that they find it very helpful in identifying the key concepts in any new topic.

In teaching a neural network course, there is a strong need for the student to experience the practical aspects of network design and performance. To achieve this, the students do a self-motivated mini-project, which gives them both practical experience in neural networks, as well as the experience of tackling real world problems, working in groups, integrating their ideas with other course work such as databases, and practising their communication and leadership skills. For this aspect of the course, we used the web as the focus for reporting progress and results. The students are provided with an area for the work, which is outside their usual quota of disk space - this provides a strong incentive for them to use this particular area. By putting all their work on the web, it becomes available for all to see, and is very easy for the students to update. This general updatability is necessary given that the groups maybe geographically distributed throughout the region and my only meet up physically once or twice a week despite working on the project practically every day. In addition, having the material on the web makes it easy to mark, and saves enormously on paper.

Some students may do very little in this scenario! The problem is how to identify this early enough to do anything about it. The solution is a programme called ‘lazy monitor’, or ‘zzz-mon’. This programme is run as a cgi-script, and the project areas, following links into students home areas if necessary. It checks the last modified date of file, which is usually sufficient although a more intelligent version compares the file contents to those from earlier in order to ensure that there has been a real change. This is because some students becoming aware of the monitoring techniques and ran simple programs to alter file modification dates! The script produces a web page that has all of the projects on it not modified with them the last few days, and these are ordered according to the severity of their laziness. The page provides links to the email addresses for easy contact with the students, which are dynamically found by running other processes on the results of the lazy monitor script. One particularly useful aspect of the programme is a single button that emails a pre-prepared snotty message to all the lazy students asking them to put in more effort into their projects!

The script is security checked in that only my machine is able to run it. After zzz-mon, it was relatively easy to modify it to produce an active monitor, which produces a separate web page identifying those projects that are busy. These simple tools, basically shell scripts that call simple Unix processes, have been used over the past few years to manage many projects, which have meant that I have been able to give detailed attention to potential problems for classes of that least 150 students. It has been interesting to see that students do not object to the reminders that they get - in particular, a dynamically created web page available for all to see that has a ‘Hall of Fame’ and ‘Hall of shame’ appears to work particularly well. The critical effect from an educational point of view is that it allows potential problems to be spotted very early on.

In the human-computer interaction course, there is also a strong need to introduce practical elements into the course to reinforce the theory. We have adopted the studio concept, where we attempt practical based small group teaching despite the fact that we have a large group situation of the least 150 students. A series of related topics to be explored are given, which are combined with lectures, web-based material, and web pointers for background reading. For the practical work itself, detailed programming is out of the question, as it puts too much effort into non- HCI issues, and the students have to much of a diverse range of abilities to be able to achieve effective results. Some of the rapid prototyping environments would be ideal, but these are expensive and machine specific. Because of this, we used the web as a design medium. HTML is pretty straightforward for people to learn and understand, and cgi scripting gives extensive programmatic behaviour to more advanced students who can exploit it - it is now also possible for people to include Javascript and Java programs within the web pages without too much difficulty. Using the web offers us access to multimedia - movies, animation, and moving text - it is also cross-platform, and most importantly for an educational establishment, it is free!

Using the web as a design medium poses a number of problems, however. There are limited text for formatting abilities, and limited control of detailed layout. The web has a shortage of conventional interface controls, and suffers from an effectively unknown set of design guidelines. Iner-platform differences compound the problems. However, we teach the students that these are not problems but challenges! All interface design has constraints imposed upon it by the underlying technology and the application - using the web as the design medium simply means we have more of these constraints that we have to work with them, and therefore have to be more creative in our solutions.

Prototyping on the web has proved hugely successful within the course, with students relishing challenging the constraints that the web imposes on them. The lack of explicit guidelines for what an interface should look like encourages creativity, and the students have managed to create dynamic pages that offer a good impression of what a fully-fledged interface would look and feel like. In addition, by being easily accessible on the web, peer review helped poor students improve their performance. We also allow the pages to be viewed externally, and the incentive of being able to show their friends in other universities what they created encourages the student to greater efforts. We have even had students from the USA doing an evaluation exercise on pages that our students have developed within the UK – inter-continental HCI collaboration at its best!

In an educational setting, the web offers a powerful, flexible medium for both material delivery and for working in, and it provides the opportunity to offer real student-centred learning. By using a number of simple tools to either create pages, or to provide management information served up by a web pages, we are able to make maximum use of the medium, and provide students with a more effective, easier to access, and more interesting learning resource. We have been able to run better courses within the same budgetary constraints which would not be possible without the use of dynamic tools doing the monitoring of students and providing active content to web pages.