There were two really interesting suggestions buried in the meeting
that I thought had a very real potential to make teaching significantly better.

The first was from (I think) Richard Buckland, and related to a
pre-requisite exam in the first week of session. This would show the
lecturer whether the assumed pre-requisites had in fact been met by the
subjects that taught them. Some kind of multiple choice exam is probably
the cheapest way to do that. Arthur Ramer pointed out that students
don't like exams and especially extra ones.

The second was from Aleks Ignatovic, who suggested that a reference
textbook be selected for common content and that a part of the subject
objectives is that students know the content of a particular chapter in
the textbook at the level that it appears in that book. For example 2011
teaches students quicksort at the level of detail in the Cormen chapter.

Some CSE courses seem to make incorrect assumptions about what students
have been taught in pre-requisites, resulting in either the same material
being taught multiple times, or assumptions being made that students
know and understand material they have never seen before. In addition,
such a scheme would enable the results of the teaching of particular
subjects to be reviewed by an academics peers.