ÖйúAƬ

On top of Darrel Ince's proposal (MM, March 10) to disband computer science departments came news of the poor showing by those departments in HEFCE quality assessments. Readers leapt to the subject's defence.

In his vast diatribe about computer science teaching, Professor Ince suggests that students need to be taught that time to market is often seen to be more important than quality. I believe that most of our students notice that the way to early market share is to produce shoddy trash that sometimes works. They just have to look around. In case they do not look we tell them - but with a regret that Professor Ince presumably would not feel.

We shortly have a re-accreditation visit from the Institution of Electrical Engineers, so I have been revisiting their guidelines. Telling students how to produce hasty rubbish to make a quick buck does not seem to come very high at all in the desiderata for the formation of a professional engineer. Of course some parts of the commercial market do this, and some always will. That does not make it a seemly subject for university education. Imagine a course: "Defect Latency Engineering" - or, in plain English, how to make things seem to work until the cheque has cleared.

Roger Needham

Professor of computer systems and head of the computer laboratory University of Cambridge