It happens.

I had compilers cancelled twice on me during my tenure as an undergraduate. It's simple economics: when courses become more specialized (as they tend to do in 4th year), applicability drops. When applicability drops, fewer students will enroll in a course. If too few students enroll in a course, it isn't economically viable to run the course. Further, courses in 4th year tend not to be *required*, hence it is more difficult to ensure a healthy enrollment...

I suppose the theory is that if the course is then postponed, a percentage of students in their 4th year who would otherwise have taken the course will come back for a 5th year (for whatever reason), while 3rd year students who also would have taken the course will still be eligible to do so the following year, compounded with new 3rd year students who are also eligible to take the course next year.

Of course, if for whatever reason a disproportionately high amount of students who would have enrolled in the course graduate and present 3rd years opt not to take it in the next offering, then the technique somewhat backfires and the course gets postponed again for similar reasons.

It's a shame really, but I don't really see a viable way around it.

There may be other "unorthodox" reasons for the courses rescheduling, but I can't begin to speculate on those.

Though not quite the same, I wonder if you could look into doing a CIS*4500 project along the same lines?