Correction of the exam

meichholz
Endlosschleifenbastler
Endlosschleifenbastler
Beiträge: 167
Registriert: 30. Mär 2016 08:28

Correction of the exam

Beitrag von meichholz »

Hi all,

we finished the correction of the exam. I will enter your final grades (including the bonus) into TUCaN and hopefully they will be available for you next week. The exam review will be approximately in two weeks.

I've sent an email to those who have submitted their solution for exercise 10 including the grading details. If you have any questions or think that you didn't get all points, please let me know as soon as possible. After the exam review it will not be possible to add missing bonus points.

Best,
Matthias

sbechtel
Mausschubser
Mausschubser
Beiträge: 56
Registriert: 17. Apr 2013 19:13

Re: Correction of the exam

Beitrag von sbechtel »

Dear Matthias,

could you give some preliminary informations on the correction? How much percent were necessary to pass the exam and how much were necessary to get 1.0? What was, from your viewpoint, the reason that the results are so bad (I mean 3.5 in avarage in an exam with a possible bonus of 1.0 is really really bad...). At least I cannot understand my result, since, when I consider the full bonus bound of 25point in the midterm exam as an equivalent of 1.0, I was 2 full grades worther in the final exam, of course with a much better preparation...

Best, Sebastian

meichholz
Endlosschleifenbastler
Endlosschleifenbastler
Beiträge: 167
Registriert: 30. Mär 2016 08:28

Re: Correction of the exam

Beitrag von meichholz »

Hi,
sbechtel hat geschrieben: could you give some preliminary informations on the correction? How much percent were necessary to pass the exam and how much were necessary to get 1.0?
you need to have at least 35 points to pass the exam and 71 to get a 1.0.
sbechtel hat geschrieben: What was, from your viewpoint, the reason that the results are so bad
By looking at the numbers I can see, for example, that a lot of students had difficulties with the last task on Software Design, but I don't really know the reasons.

Best,
Matthias

eichberg
Dozentin/Dozent
Beiträge: 448
Registriert: 25. Sep 2007 12:12
Kontaktdaten:

Re: Correction of the exam

Beitrag von eichberg »

I corrected the last task myself and it became very obvious that – except of a few students – no one "really" did the exercises related to the patterns... I was deeply shocked that only a very small fraction of all students even considered modelling the base structure using the composite pattern. Solving the other design tasks was then typically also not possible.

(I said it multiple times - very explicitly - that I expect everyone to know this pattern as it is one of the most fundamental patterns of them all and it is already taught in foundational course; additionally, we have seen it multiple times across the lecture/exercise.)

Given these very disappointing results, we made the decision that we will not offer SED&C in the same form again. (To avoid confusion: SED&C will be offered again, but the exercises will be structured differently.)

(BTW: Most students did not have nearly the complete bonus!)

Benutzeravatar
mr_composite
Neuling
Neuling
Beiträge: 1
Registriert: 6. Mär 2017 21:30

Re: Correction of the exam

Beitrag von mr_composite »

I did none of the exercises or bonus, but the composite pattern was very obvious and even I was able pass with 1.3 (but I have no Idea where I lost the 0.3)

majidazimi
Erstie
Erstie
Beiträge: 17
Registriert: 16. Jul 2016 12:25

Re: Correction of the exam

Beitrag von majidazimi »

eichberg hat geschrieben:I corrected the last task myself and it became very obvious that – except of a few students – no one "really" did the exercises related to the patterns... I was deeply shocked that only a very small fraction of all students even considered modelling the base structure using the composite pattern. Solving the other design tasks was then typically also not possible.

(I said it multiple times - very explicitly - that I expect everyone to know this pattern as it is one of the most fundamental patterns of them all and it is already taught in foundational course; additionally, we have seen it multiple times across the lecture/exercise.)

Given these very disappointing results, we made the decision that we will not offer SED&C in the same form again. (To avoid confusion: SED&C will be offered again, but the exercises will be structured differently.)

(BTW: Most students did not have nearly the complete bonus!)
My 2 cents for this disappointing result:

1. Exercises prepared students for math while the exam was about chemistry. If the student is supposed to recognize patterns in the final exam, then why he/she should implement a functional set(exercise 1 and 2) as an exercise?

2. I guess, by experience you already knew that most students are not even familiar with basic patterns, then what is the point of introducing type class pattern and then removing it from exam at the very last lecture? You are trying to teach calculus to students who don't know how to sum two numbers.

3. Time (True for 90% of exams in TU Darmstadt). Even though I got 1.3, but at least for me and couple of other students to whom I talked, the time to complete the last task was not sufficient. The end result is most students could not think deep enough to find the best fitting pattern for the problem.

4. It is not that much disappointing: Considering that most students are not professional designers, so it is obvious that they will fail to deliver a good design at their first attempt. Add exam timing/stress to it and holy smoke...

Benutzeravatar
Mark_G
Mausschubser
Mausschubser
Beiträge: 67
Registriert: 8. Okt 2008 23:07

Re: Correction of the exam

Beitrag von Mark_G »

I agree. Understanding foreign code and recognizing patterns has not been practiced at all. This is an entirely different task from implementing a pattern (when you already know which one) yourself.

I did all the exercises and have full bonus but only got 2.7. Quite disappointing.

Also time was indeed too limited, especially for the last task.

kasraq
Neuling
Neuling
Beiträge: 6
Registriert: 15. Mai 2016 20:35
Wohnort: Darmstadt
Kontaktdaten:

Re: Correction of the exam

Beitrag von kasraq »

Mark_G hat geschrieben:I agree. Understanding foreign code and recognizing patterns has not been practiced at all. This is an entirely different task from implementing a pattern (when you already know which one) yourself.

I did all the exercises and have full bonus but only got 2.7. Quite disappointing.

Also time was indeed too limited, especially for the last task.
majidazimi hat geschrieben:
eichberg hat geschrieben:I corrected the last task myself and it became very obvious that – except of a few students – no one "really" did the exercises related to the patterns... I was deeply shocked that only a very small fraction of all students even considered modelling the base structure using the composite pattern. Solving the other design tasks was then typically also not possible.

(I said it multiple times - very explicitly - that I expect everyone to know this pattern as it is one of the most fundamental patterns of them all and it is already taught in foundational course; additionally, we have seen it multiple times across the lecture/exercise.)

Given these very disappointing results, we made the decision that we will not offer SED&C in the same form again. (To avoid confusion: SED&C will be offered again, but the exercises will be structured differently.)

(BTW: Most students did not have nearly the complete bonus!)
My 2 cents for this disappointing result:

1. Exercises prepared students for math while the exam was about chemistry. If the student is supposed to recognize patterns in the final exam, then why he/she should implement a functional set(exercise 1 and 2) as an exercise?

2. I guess, by experience you already knew that most students are not even familiar with basic patterns, then what is the point of introducing type class pattern and then removing it from exam at the very last lecture? You are trying to teach calculus to students who don't know how to sum two numbers.

3. Time (True for 90% of exams in TU Darmstadt). Even though I got 1.3, but at least for me and couple of other students to whom I talked, the time to complete the last task was not sufficient. The end result is most students could not think deep enough to find the best fitting pattern for the problem.

4. It is not that much disappointing: Considering that most students are not professional designers, so it is obvious that they will fail to deliver a good design at their first attempt. Add exam timing/stress to it and holy smoke...
I agree. Time was really insufficient both for midterm and final. Not fair!

sbechtel
Mausschubser
Mausschubser
Beiträge: 56
Registriert: 17. Apr 2013 19:13

Re: Correction of the exam

Beitrag von sbechtel »

After today's exam review I have two points of critisism:

1) Since the exercises were programming tasks, our own chance to see whether we are able to solve exam questions was the mid term exam. One should be able to conclude from the mid term exam that one's performance in solving exam questions is sufficient or not. This was not the case! First, major requirements (identifying patterns and sketching a design, together worth much more than one third of reachable points) were not tested (at least I cannot remember...) in the mid term exam. Second, the rigidity in the correction style was not comparable! In the finial exam we had similar questions compared to the mid term exam (similar in the sense of style and difficulty) and I gave answers in a similar quality, but in the mid term exam I got a lot more points on my answers. Minor problems seemed to be ignored in the mid term exam, whereas in the final exam for every little mistake I lost a lot of points. E.g. in the task where one had to implement the fps signal I lost 2P because I forgot to invoke changed (this is a half grading step!) and I was told in the exam review that one could also remove 3P for this mistake...

2) In the lecture it was stressed that a design pattern has a purpose and that there are patterns which look similar but whose difference lies in differing purpose. Now we had this task where one should identify patterns and the task required to only identify those which are "clearly identifiable". If I only have an excerpt of code and I don't know which problem the code should solve, then I clearly cannot take the purpose of a pattern into account but have to identify things where the code looks like an application of something. How should I be able to choose the right one from similar looking ones if I don't know the context and the purpose of this pattern application? This is not possible. There were classes which are not used, therefore I have to assume that there is code not shown where this class gets used and whether this code uses the class as would be intended by a pattern X cannot be said. Therefore the task description doesn't make sense to me. There shouldn't be a schema for giving points on such an exercise, if there is a sufficient amount of succesfully identified patterns one should get full points in my opinion. In my case I write down 4 patterns and got 4 points (i.e. I lost 2 grading steps on that exercise!)...

Antworten

Zurück zu „Archiv“