All entries for Friday 07 May 2004
May 07, 2004
It's been a mixed week of both troubleshooting and continuing the work on LoadStudents. Attempted to rope Julie in to use Oracle admin tools while the process was running, but she said she wouldn't have time to look at it for at least a week. So attacked it from the Java end, with advice from Libelis, and eventually had success in making it run in optimistic-locking mode. Between locking mode, refactoring, sorting and some cacheing, speeded it up from 453 mins to 100 mins, a speedup factor of 4.5, which I will call good enough and leave it at that.
The opening of OMR for students to poke at always brings a swathe of problems to investigate for the Academic Office. There was one bug, but as usual, all the rest turned out to be data quality issues.
It's a pain that we have to spend time working out data problems that they have to fix, but it's a difficult problem to solve. The troubleshooting generally involves seeing something that's going wrong on the OMR web interface, and tracing this from the OMR frontend through the OMR database to the SITS UWTABS tables to the SITS product tables to one of scores of possible screens on the SITS frontend. This is such a complex procedure that the AO staff would have no chance, nor should they have to do this. But it's not clear to me how we could bypass it either.
7833 students (out of 9412) have used the Exam Timetable system since go-live two weeks ago. We've heard of less than a dozen who had a problem with it (all of them unsure how to display PDFs) – that's under 0.1%. A pleasing success rate.