Course analysis for DD2459 Software Reliability

Period 3, 2012

Author: Karl Meinke, Nada

^ Up to Course home page period 3, 2012.

Below is a course analysis for the course Sofrel, Software Reliability. This was a new course, and no previous historical evaluation data is available for this course.

Course data

CourseDD2459 Software Reliability. (7.5p)
Course period Period 3, 2012
Lectures 14 hours
Exercises 8 hours
Labs (timetabled supervised) 6 hours
Coursebook P. Amman and J. Offutt, Introduction to Software Testing, Cambridge University Press, 2008, first edition.
Number of Active students 50
Passed exam (end of Period 3 2012) 48 students = 96%
Completed Course (end of Period 4 2012) 44 students = 88%
Course teacher Karl Meinke
Lab Assistant project marker Siavash Soleimanifard

Course Evaluation Form

29 students completed a course evaluation during the last lecture. (Paper format). This represents 58% of students taking the course, so the results should be reasonably representative. The course evaluation form was designed to test alignment between published learning outcomes and actual learning outcomes from the students own perspective.

Comments

Here are a selection of comments,

The difficulty (of the labs) was good for the time you have for each lab"

Not a very difficult course, lectures were excellent, it comes across very clear that it is an important subject

I'd like to see a lecture on agile testing methods.

Constant fear of having to present your solution in class.

I liked the theory behind pairwise testing.

New concept that is good. Lot of time for labs is good since it makes it more easy in combination with other courses.

After this course I will know white/black box testing, model-based testing and have tested NuSMV. That is more than before the course.

The lectures are a little abstract for the first learner to understand.

The lab tasks is good chosen (sic).

There are too many things in software testing which can not cover in one course. So there should be following further course in this area. (sic)

(Lectures were) a little bit too fast with too much information.

Some things were inclear (until now) : finite/infinite counterexamples (how to distinguish).

More exercises and examples in PLTL. PLTL task sin the exam was among the most time consuming tasks.

The form of the exam we had (48 hours to submit the solved exam) is highly appreciated. You have enough time to sit and think and it is more "real-life" simulation than traditional examination form.

Too many abbreviations ... would help with an abbreviation list in the course web site.

(Lectures) ... very good presentation of the material.

(Exercise classes) ... we are actively contributing in the exercise -> very good learning method (we are enforced to study).

Learning to rationalise and thinking structurally helped a lot.

More examples in the lessons would be good.



Summary

The exam performance for the course was very good, although a somewhat easy paper was set since the course ran for the first time. Attendance at lectures was very good (rooms were sometimes too small!). Attendance at exercise classes was less good. There appeared to be several scheduling clashes with other courses. The percentage of students purchasing the course book is very low, even though (in my opinion) the course book is modern and very good. The greatest source of dissatisfaction would seem to be the exercises, so perhaps these can be improved next year.

There was considerable trouble installing commercial testing software. The computing service failed to achieve this after 3 months of collaborating with the software vendor, for reasons that were never clear.

Goals

The subject matter of this course concerns modern and practical techniques for software testing and reliability modeling, suitable for the working software engineer. Software testing concerns the problem of analyzing and evaluating software products to reach some conclusion about their fitness for use. For safety critical, mission critical and enterprise critical systems development it is nowadays essential to have some form of software quality assurance measures in place within an IT project. Quality assurance measures may involve both managerial and technical procedures. However, the primary focus of this course is on technical procedures to predict, discover and diagnose errors in software systems.

The intended learning outcomes are published on the course home page.

Changes made to the course

Not applicable, first time the course was run.

Reflections on this years course

  • This course seems to have started well. Student numbers and student satisfaction levels are good.
  • Course participation and completion rates are good, indicating a good level of motivation.
  • Lab work in the end went smoothly, and there is some virtue in the students not using existing tools, but developing key components of these themselves.
  • Intended Changes for next year

    ^ Up to Course home page period 3, 2012.


    Sidansvarig: <karlm@nada.kth.se>
    Last changed 13 June 2012.
    Tekniskt stöd: <webmaster@nada.kth.se>