[NSWI004] Regarding Project Activity Points

[Ondřej Roztočil]

Hi Peter,

you misunderstood me, but I take it as my responsibility for not writing clearer.

My remarks were not about difficulty and not really about total time demands. Clearly, any Matfyz student can do greater things than write 2000 lines of code given 20 free days. (Although, I am not sure you have done the math on how many weeks there are in the rest of the semester.)

Regarding "fairness" of the metrics, to simplify my question: Why should someone who does the same or better job in 1000 LOC and does it on 10 separate days receive less points than someone who does it in 2000 LOC on 20 days.

Then there are technical questions and issues. How does it work with Git, what code is taken into account, etc.

To your points:
> It isn't even clear whether the metric should measure your effort or the quality of your work.
Obviously, this metric is intended to measure (and ultimately promote) effort - how could it measure quality in the first place?

> Both of these are hard to even define properly, not to mention the difficulty of finding easily testable tools to measure them with.
Yes, that was part of my point. However, what I am wondering is why there has to be a quantifiable metric in the first place. The output - i.e. the assignment solutions - is already tested for correctness and then manually reviewed (through considerable effort on the reviewer's part). Why not avoid the potential issues and either make all the assignments individual or leave it up to the teams to handle fair division of work internally (only stepping in if some teams become truly dysfunctional)?

Anyway, I should have more clearly separated the practical questions from the 'why' questions, as the first are more relevant for the work at hand.

Best regards

O.

From: NSWI004 <nswi004-bounces at d3s.mff.cuni.cz> on behalf of Peter Fačko <pfacko1 at gmail.com>
Sent: Wednesday, October 21, 2020 7:31 AM
To: Operating Systems Course <nswi004 at d3s.mff.cuni.cz>
Subject: Re: [NSWI004] Regarding Project Activity Points 
 
Hi,

I am sure that if your goal is not to game the system you will write 2000 LOC in at least 20 days no problem. 

Any objectively quantifiable metric one would choose for this would be artificial.
It isn't even clear whether the metric should measure your effort or the quality of your work. 
Both of these are hard to even define properly, not to mention the difficulty of finding easily testable tools to measure them with.

On Wed, 21 Oct 2020, 00:11 Tomáš Kubíček, <tomas.kubicek69 at gmail.com> wrote:
I would add one more thing to those points. Point per 100 lines of code will in my opinion encourage us to write ugly looking code, because why would we write something compact and easy to understand when we will get penalized for that.

On Tue, 20 Oct 2020 at 23:28, Petr Tlapa <tlapik123 at gmail.com> wrote:
I agree, I think these points you mentioned are quite solid and a really good ground for some kind of discussion. 

út 20. 10. 2020 v 23:02 odesílatel Tomáš Kubíček <tomas.kubicek69 at gmail.com> napsal:
I have similar doubts about this system and would appreciate discussion about this topic.

On Tue, 20 Oct 2020 at 22:26, Ondřej Roztočil <roztocil at outlook.com> wrote:
Hi,

the mailing list seems to be quiet lately. I would like to use this opportunity to post some questions and unsolicited critical thoughts about the concept of 'Project Activity Points' (see 00-common.md in grading repo) which is going to be relevant for students soon. I understand that the teachers have thought about this more than I have, so they probably had their reasons for setting the rules like this. However, I expect more students to have similar doubts and it might be useful to have some discussion to dispel these doubts.

I will try to be brief:

1) Using number of days on which students make a commit as a metric for points seems rather unfriendly towards their schedules. If a student communicates and plans with their team well, I don't see why they should be penalized for only having time to work on the assignments once a week, rather than twice a week, for example. This can of course be "gamed", but that seems silly and promotes bad Git practices.

2) Number of submitted lines of code is also a questionable metric. Generally, the shorter the code is (while retaining readability, etc.) the better. Or do you expect that everyone will need to write 2000+ LOCs anyway? Further, how does it work with Git history and what code is actually taken into question? Should teammates refrain from refactoring each other's code or they risk taking away their points?

3) Finally, more theoretical point. Maybe I am missing why there are teams for this course in the first place, but it seems to me that these 'Activity' rules teach the wrong lesson about team development: that writing code is the only valuable thing a person can do. Research, design, all sorts of communication - to give few examples relevant even for our limited project - are just as important as coding. Has a system without artificial metrics, where it is solely up to the teams to make sure everyone does a fair share of work, been tried in the past and it proved too problematic?

Sorry for not structuring this better and sorry to my teammates if I tanked our chances for passing the course.

Best regards

Ondřej
_______________________________________________
NSWI004 mailing list
NSWI004 at d3s.mff.cuni.cz
https://d3s.mff.cuni.cz/mailman/listinfo/nswi004
_______________________________________________
NSWI004 mailing list
NSWI004 at d3s.mff.cuni.cz
https://d3s.mff.cuni.cz/mailman/listinfo/nswi004
_______________________________________________
NSWI004 mailing list
NSWI004 at d3s.mff.cuni.cz
https://d3s.mff.cuni.cz/mailman/listinfo/nswi004
_______________________________________________
NSWI004 mailing list
NSWI004 at d3s.mff.cuni.cz
https://d3s.mff.cuni.cz/mailman/listinfo/nswi004