Too big for a term? Why terms?

I’ve reached the conclusion that a lot of courses have an unrealistically high number of evaluations. We have too many and we pretend that we are going to achieve outcomes for which we have no supporting evidence. Worse, in many cases, we are painfully aware that we cause last-minute lemming-like effects that do anything other than encourage learning. But why do we have so many? Because we’re trying to fit them into the term or semester size that we have: the administrative limit.

One the big challenges for authenticity in Computer Science is the nature of the software project. While individual programs can be small and easy to write, a lot of contemporary programming projects are:

1. Large and composed of many small programs.
2. Complex to a scale that may exceed one person’s ability to visualise.
3. Long-lived.
4. Multi-owner.
5. Built on platforms that provide core services; the programmers do not have the luxury to write all of the code in the system.

Many final year courses in Software Engineering have a large project courses, where students are forced to work with a (usually randomly assigned) group to produce a ‘large’ piece of software. In reality, this piece of software is very well-defined and can be constructed in the time available: it has been deliberately selected to be so.

Is a two month software task in a group of six people indicative of real software?

June 16: Remember to curse teammate for late delivery on June 15.

Yes and no. It does give a student experience in group management, except that they still have the safe framework of lecturers over the top. It’s more challenging than a lot of what we do because it is a larger artefact over a longer time.

But it’s not that realistic. Industry software projects live over years, with tens to hundreds of programmers ‘contributing’ updates and fixes… reversing changes… writing documentation… correcting documentation. This isn’t to say that the role of a university is to teach industry skills but these skill sets are very handy for helping programmers to take their code and make it work, so it’s good to encourage them.

I believe finally, that education must be conceived as a continuing reconstruction of experience; that the process and the goal of education are one and the same thing.

from John Dewey, “My Pedagogic Creed”,  School Journal vol. 54 (January 1897)

I love the term ‘continuing reconstruction of experience’ as it drives authenticity as one of the aesthetic characteristics of good education.

Authentic, appropriate and effective learning and evaluation activities may not fit comfortably into a term. We already accept this for activities such as medical internship, where students must undertake 47 weeks of work to attain full registration. But we are, for many degrees, trapped by the convention of a semester of so many weeks, which is then connected with other semesters to make a degree that is somewhere between three to five years long.

The semester is an artefact of the artificial decomposition of the year, previously related to season in many places but now taking on a life of its own as an administrative mechanism. Jamming things into this space is not going to lead to an authentic experience and we can now reject this on aesthetic grounds. It might fit but it’s beautiful or true.

But wait! We can’t do that! We have to fit everything into neat degree packages or our students won’t complete on time!

Really?

Let’s now look at the ‘so many years degree’. This is a fascinating read and I’ll summarise the reported results for degree programs in the US, which don’t include private colleges and universities:

• Fewer than 10% of reporting institutions graduated a majority of students on time.
• Only 19% of students at public universities graduate on-time.
• Only 36% of state flagship universities graduate on-time
• 5% of community college students complete an associate degree on-time.

The report has a simple name for this: the four-year myth. Students are taking longer to do their degrees for a number of reasons but among them are poorly designed, delivered, administered or assessed learning experiences. And jamming things into semester blocks doesn’t seem to be magically translating into on-time completions (unsurprisingly).

It appears that the way we break up software into little pieces is artificial and we’re also often trying to carry out too many little assessments. It looks like a good model is to stretch our timeline out over more than one course to produce an experience that is genuinely engaging, more authentic and more supportive of long term collaboration. That way, our capstone course could be a natural end-point to a three year process… or however long it takes to get there.

Finally, in the middle of all of this, we need to think very carefully about why we keep using the semester or the term as a container. Why are degrees still three to four years long when everything else in the world has changed so much in the last twenty years?