All entries for April 2005
April 21, 2005
The updated exam timetable software went live last Friday, the same day as the summer exam timetable was finalised (exactly on time according to plan). After some "real data" testing on Monday, we finally made personalised students' exam timetables available on Tuesday.
The student-facing system is essentially unchanged – the PDFs are generated on request from the database. Under the covers there was some improvement to the code for better maintainability. But most of the new development went into a new system for departments to nominate their invigilators via web forms instead of the old paper forms. We start piloting this tomorrow.
What surprised us was the huge demand for the personal timetables, much higher than last year. In spite of the complete exam list being published on a noticeboard last Friday, and in spite of us opening the system quietly at the very end of the working day on Tuesday, by 8am on Wednesday the system was overloaded and I had to restart it with more memory. The load was significantly higher than last year's peak.
Aside from the IT systems perspective, students seem to have been clamouring for exam timetables this year – I don't know why this should be different from any other year. A particularly immature student union officer decided it would be clever to organise a campaign of students emailing the exam office to demand publication. Several hundred students did so, although some of them apparently emailed to say they disagreed with the union's position and were quite happy with the current system.
There are two reasons why such a tactic could never work. Firstly, and very obviously, the task of reading hundreds of emails can only slow down the work done by an office, not speed it up. But secondly and more importantly, the exam office is an example of a central service department which depends heavily on academic departments. The main reason why producing an exam timetable is hard is that departments will not provide accurate information in a timely manner. They prevaricate, delay, provide wrong information and change their minds. The task of co-ordinating what exams need to happen is one of chasing up departments and meeting, in some cases, much less than full co-operation. It must be a frustrating job.
Module registration is subject to much the same kind of process: although modules for next academic year are supposed to have been finalised at this point, some modules are missing, some are still 'provisional' because departments failed to submit them to the approving committee in time, and one department has even completely opted out of the pre-registration phase for Online Module Registration because they intend to make so many changes to their courses (which should have been finalised by now) between now and October that it would be pointless.
It's a shame that students often perceive these issues as problems of the IT systems or the Academic Office. We can't do our jobs properly unless academic departments provide us with the information we need. Some departments are good at it, and some bad. It seems to me the University, on behalf of the students, ought to be a bit firmer with departments about what their information management responsibilities are. It's not a very sexy subject, and not one which academics are likely to be very interested in, but it's vital to being a smoothly run university.
April 06, 2005
I was interested to read about the new generation of the web sign-on service that Kieran's been working on. I had come across SAML a few months ago while looking into something else and made a mental note that it looked like a good, well-focussed standard.
What's also interesting is the fact that the new software has taken months of development effort, whereas I hacked together the first version in hours (admittedly re-using some code I'd previously written for something else). This seems typical of the way e-lab's work has changed in its 3 years' existence: in the early days we needed to get stuff up and running quickly, at least in prototype form, whereas now we're putting in the effort to 'do things properly', at least in areas we've identified as being worth the effort, for reasons of scalability and maintainability. My new project, the Academic Data Store, is another example. We have databases that do the various jobs, but the way they all connect to each other is messy, hard to extend and problematic to maintain, so we're going to pour resources into a proper meta-database that will fix that.
As someone who's always been interested in big, complex systems I like this trend. 20 years ago I was fascinated by operating systems, now there are worlds of fascination just within J2EE. I've never been very interested in small projects. Partly that's because I'm process-oriented rather than goal-oriented; I enjoy finishing projects, but it's the getting there that really interests and motivates me. Actually the designing and planning of projects is my favourite part; the blank slate is where there's the most scope for creativity.