Council meeting minutes from 2024-10-23

minutes of the students' representative council for study programs of computer science

Nikofete

Lukas reminds people that the Nikofete is coming up.
For this we usually collect donations a local charity organization. Lukas ask everyone to think of possible organizations to support. We want to make a discussion about this next week. If you have a suggestions, bring it up in the meeting. Please remember to have a specific sugeestion.

Lukas will also sent around a poll to select a suitable date.

SE-Lab

Last semester break the SE-Lab happened. We had several guests that wanted to bring up their inputs on that matter.

Manuel reiterates that we already had complaints in the last zear and had numerous talks on what could be improved. Some things have been adjusted in this iteration, eg clarification on tests and passing requirements.

In this meeting, we wanted to ollect issues and complaints on this years iteration.

One students noted that from their perspectives it seems that the situation for the group phase has improved. It was positively noted that there is now a clear number of tests to be passed. However they generally still think that the SE-Lab is too demanding in general.
It was critizised that some people had exams in parallel to the lecture week, which ends with the entry exam. Manu comments that an overlap with exams, especially advanced lectures will be very hard to avoid.

Daniel brought up some complaints that he received via text. Appareantly there were some cases in which behavior was tested which was not clearly specified in the specification.
In addition there were long waiting times in the office hours, especially in the individual phase. This resulted in some people waiting for the whole office hour and not getting a spot. Manuel already had a discussion with the SE-Chair. In their discussion it came up that it didn't reached them that significantly amount of people were cut off at the end.

There was a discussion on whether it would be reasonable to have pre-allocated time slots that students can book via CMS. The chair has made bad experiences with that previously, where people would book slots as soon as they came out and then had no specific questions at all, thus blocking time for students with valid questions. There were various discussions if and how this can be realized in a meaningful manner.

There were complaints that the tests sometimes test edge cases which are not obvious. There was also a point made that it was not clear how and why tests were selected as mandatory and optional. The optional tests are supposedly for debugging your code. One student reported that they passed (nearly) all optional tests, but not enough mandatory tests, in which case the optional tests did not help in debugging.

There were calls for specific examples in which the specification did not match the tested behaviour. One student reported that they discovered one case were the reference implementation was inconsistent. Another point raised was that the specification was vague in some points in general.

Another point that was brought up is that getting answers in the group phase was much easier as there was more close supervision by tutors. In the individual phase, there was a forum to ask for clarifications. While posts were answered relatively quickly, the forum in general was overloaded with information and sometimes there were some contradicting answers. It would be nice to have a single ground of truth, however this results in a sync issue to transport all information in a timely manner to all students and tutors.

Alex has asked for specific examples in which the specification is unclear and students can only find out the specific behaviour through writing system tests. People with specific examples of that should give them to us in writing.

Manuel asked for clarification how Mutants work, as he heard several complaints before. Students mentioned that they seemed it was easier to find mutants in the individual phase but still there apparently were bugs in the reference implementation which made finding mutants hard. This resulted in some students brute-forcing several variants of a behaviour to determine the implemented behaviour in the reference.

One issue that multiple students mentioned that the refueling behaviour was especially confusing and unclear, which lead to many students failing. We will discuss and note down the exact details for later discussion.

Max summarized that it might be more helpful to have more descriptive test descriptions and to find a way to streamline the communication process.

One suggestion of a student was to improve the specification situation by either reusing old projects in which a lot of issues might have been fixed in the previous iteration or use real world specifications. Reusing old specifications is currently not feasible as there were only two iterations using the current (simulation) approach to implement. For real world specifications it can be hard to find projects that are suitable for second semester students.

Daniel also brought up that there were few students, who could not take re-exams during the SE-LAB due to the workload. Another student raised the same issue we already discussed last year, that the the organization of SE-LAB during the semester break makes it harder for students who have to work in parallel. Alex commented that while this is problematic, the university is a full time study. If this would happen to part-time students, which are enrolled as such, this would be a huge problem. However we are currently unaware of any students who are enrolled as part-time students and had issues.

A student raised that the entry exam included a question that requires Kotlin knowledge, which many students only learned a few days prior.

We continue our discussion with potential mitigation tactics:
- Do we want to keep SE-Lab?
- Can we fix SE-LAB?
- If so, hat are our proposals?

Alex claims that while he is a theory guy, he would like to keep SE-LAB as a practical project. He proposes to for now to communicate our complaints and see if there are improvements. If not, we need to push for an alternative.
Mona raises the point that abolishing SE-LAB is not feasible, but maybe we can push for substantial organizational changes, e.g. removing/adapting individual phase.

Manuel mentions that from his point of view, there is no feasible way to convert SE-LAB into a regular course that spans the whole semester, simply because the study plan does not allow for another 9 CP lecture during any semester.
Manuel thinks it is bad that there is mandatory attendance during the group phase and critisizes that groups pass and fail as a whole.

Manuel has summarized some options that he discussed with the SE chair:
- Re-Using Specifications
...

Manuel also talked with another person about experiences from similar lectures at different universities and shared their experiences. IN these other courses there seemed to be a more detailed supervision by tutors to further evaluate group contributions. This approach would most likely require more tutors (and thus, funding)
Manuel thinks its highly problematic that there are good students who have continuously failed the SE-LAB. In some occasions there are students who are only missing the SE-LAB. This may indicate that there are systematic issues with how it is organized.

Students reported that they could not use their full potential during the individual phase because they were exhausted from the group phase. They also mentioned that the communication atmosphere of the chair towards the students could be more positive.

Simon explains, how there is not much room regarding the timing of the course. A student mentioned that maybe some deadlines could be converted to soft deadlines to reduce the stress on students.
A proposed approach could be to recognize partial passes in future iterations. Manuel claims that this would maybe hurt students more, as they have to learn about a whole new project in the next year.
A break between the group phase and individual phase was proposed, but there does not seem to be time for that.

We collected thoughts on the idea to drop the individual phase and replace it with a big exam at the end of a course. Manuel thinks that would make passing the course easier and would be less frustrating as e.g. goose chasing mutants.
Daniel asked whether it would be reasonable to lower the required tests to pass. Manuel mentions that we don' t really have a say in that, as it falls under freedom of teaching. Ferdinand further mentions that students do not have a right to an easy exam, but to a fair exam (As in, comparable to how other students are graded in the same course)

Generally it feels to some people that SE-LAB is increasing to become a pressure point that should be addressed and discussed. Simons suggest to request insights to the QUALIS evaluations of the past couple years to see if there is a trend.

Manuel wanted to collect points on what we want to address:
- The students council should make a decision whether we want to propose a fundamental change to SE-LAB. Daniel mentions that this would also require us to make a soecific proposal on how this can be done.
- Do we want to keep monitoring the changes? And where do we make a cut?
- Manuel feels like the general sentiment has moved from angry to depressed, which is not recognized by the SE Chair at the moment.
- Manuel will ask the students' council in the next meeting to discuss that in detail