Software Evolution

Course code: 2IMP25

Time and location: Wednesday, 15:45-17:30 (Paviljoen L10) and Friday, 10:45-12:30 (Paviljoen J17). The course has been videotaped, see below.
Target audience: Master (CSE,ES,BIS) students interested in Software Engineering; mandatory for CSE-students following the Software Science stream and recommended (stream elective) for CSE-students following Systems Science stream. Students intending to work on their graduation project at the Software Engineering & Technology specialization are specially encouraged to participate. Students following other programs at TU/e are welcome to join the class, subject to the regulations pertaining to the program followed and my approval.
Whether you would like to become a software developer or a researcher, a quality manager or a software architect, a requirements engineer or a consultant, this course is for you! (see also How to continue)

Background information on the course:

Nowadays change is often considered as the only constant factor in software development. Successful software systems are, therefore, those systems that can adapt to the ever changing requirements of the environment. One can, thus, compare this process of adaptation to the Darwin's "Survival of the fittest" principle. This course will explore issues related to software evolution: why it is difficult, how we can cope with this difficulty and what can we learn from the past. Specifically, the course will look at advanced tools and techniques proposed by the research community to understand, ease and automate software evolution. In this course, you will have a chance not only to learn about methods and tools of software evolution, but also to apply them to assess software evolution of existing software systems.

Prerequisites: An undergraduate software engineering course. Up-to-date knowledge of Java. Knowledge of scripting languages is an advantage.

Learning objectives:

After completing this course, you will be expected to be able to

list important challenges associated with software evolution;
discuss methods and tools addressing these challenges, their advantages and disadvantages;
apply the methods and tools to existing software systems;
interpret the results obtained in a scientifically responsible way.

Assessment

As opposed to the previous years, this year the assessment is based on the homework assignments and the final exam.

The assignments are available on Peach. You can access Peach with your usual university login; once you have registered in OASE you should be able to access the assignments. Keep in mind that the deadlines are strict and no late submissions will be accepted (either by Peach or by me).
This year we will have three assignments that can be carried out in pairs. If you cannot find a partner yourself, I can send an email to all the students to suggest them to contact you if they also do not have a partner. However, I will not assign partners.
Deduction policy and grades (per assignment):

Anonymous feedback on the assignments can be provided using the online form.
Final grade for the assignments part A is determined as w₁*A₁ + w₂*A₂ + w₃*A₃, where

A₁, A₂, A₃ - grades for the three assignments
w₁, w₂, w₃ - weights for the three assignments; the lowest grade is multiplied by 0.2, the remaining two by 0.4.

Clarification (Feb 20, 2016): student that fails to submit an assignment gets 0 for this assignment.

Since this is the first time the exam is given, no previous year exams are available. To give you an impression of what can be expected from the exam, please check the mock up exam.

To pass the course, you should score at least 5 both for the assignments and for the exam. Let A be the assignments grade calculated as above, and E be the exam grade. Then, the final grade is calculated as follows:

0.6 * A + 0.4 * E rounded to the nearest integer, if A ≥ 5 and E ≥ 5.
min(5, 0.6 * A + 0.4 * E) rounded to the nearest integer, if either A < 5 or E < 5.

Special cases:

If a student registers for an examination but fails to appear he/she will be deemed to have failed the examination and the examination result will be marked as a "no-show" (NV). (Program and Examination Regulations CSE 4.5.4).
If a student has cheated, the examination result, in accordance with Article 5.3, paragraph 6, will be deemed "failed" (ON). (Program and Examination Regulations CSE 4.5.5)

Those who fail should submit the replacement assignment (if A < 5), retake the exam (if E < 5 or "no show") or both (if A < 5 and either E < 5 or "no show"). The replacement assignment will be announced on Peach after all the assignment grades have been published.

Fraud policy

As any course offered at Eindhoven University of Technology, it is subject to the TU/e Educational Fraud Policy. At the very basic level, it means do not copy your solutions from the Internet or from other groups; if you need to cite a scientific paper or a website do this in the appropriate way. Your submissions might be subjected to the plagiarism detection tool.

Check the IEEE Plagiarism FAQ.
The text cited should be clearly visually separated from the surrounding text (e.g., using quotes).
Visual elements (tables, plots, figures) taken from the external sources should contain reference in the caption.

Please keep in mind that copying is considered fraud, and students who copy may be reported to the examination committee.

Resources

The course has been videotaped in 2012. You can find the lectures on http://videocollege.tue.nl/, (2) Computer Science, 2IS55 (old course code). Limitations of the video:

the assessment has been changed: exam has been introduced and different assignments have been introduced. Always check Peach for the most up-to-date information about the assignments and submission deadlines.
Relative importance of some topics have been changed: we will spend less time than in 2012 discussing code duplication and testing; more time discussing repository mining. Check the lecture plan of 2012 and compare it with this page.
Guest lecture of prof. Vinju is absent from the video.

There is no mandatory textbook for this course. Should you be interested in software evolution, you might like to consult the following books (I've used them when preparing the lectures):

Tom Mens, Serge Demeyer (eds.) Software evolution, 2008. Read via SpringerLink or purchase.
Tom Mens, Alexander Serebrenik, Anthony Cleve (eds.) Evolving Software Systems, 2014. Read via SpringerLink or purchase.
Nazim H. Madhavji, Juan Fernandez-Ramil, Dewayne Perry (eds.) Software Evolution and Feedback: Theory and Practice, 2006. Purchase.
Paolo Tonella, Alessandra Potrich Reverse Engineering of Object Oriented Code, 2005. Read via SpringerLink or purchase.

SpringerLink will work only if you are at the university network, use a VPN connection or the library proxy server.

Continuation opportunities

If you like the topic of software evolution, consider one of the following options to work on a research assignment in this area:

2IMP00 Seminar Software Engineering and Technology (Q2, 5 ECTS). You will work on a small research project supervised by me or one of our PhD students, and learn how to conduct research and report on the research results orally and in writing to a group of peers. Seminar is mandatory for CSE students.
2IMP05 Capita Selecta Software Engineering and Technology (not scheduled, 5 ECTS). Individual research assignment (by invitation only). You can always indicate that you are interested in Capita Selecta and if your assignment grades are sufficiently high, I will invite you. "Not scheduled" means that the project can take place whenever we agree upon.
Honors project (not scheduled, extracurricular). If you are an Honors program student, you can conduct one of your Honors research projects under my supervision.
Internship (not scheduled, 15 ECTS). Internship can be conducted in academia or industry, in the Netherlands or abroad. In particular, if your assignment grades are sufficiently high I would recommend you to go for a more research-oriented internship. Internship might also be an interesting option if you are looking for an international experience.
2IMC00 Master project (not scheduled, 30 ECTS).

Lectures

The list of lectures and topics is subject to change. Powerpoint presentations will be made available after the class.

Location	Date	Topic
PAV L10	Feb 3	Introduction. S-, P- and E-type systems. Lehman's laws of software evolution.
PAV J17	Feb 5	Requirements Evolution.
	Feb 10	No class. Carnaval.
	Feb 12	No class. Carnaval.
PAV L10	Feb 17	Architecture reconstruction (reverse engineering) of structural UML models.
PAV J17	Feb 19	Architecture reconstruction (reverse engineering) of behavioural UML models. Architecture Description Languages.
PAV L10	Feb 24	Guest lecture. Dr. Jurgen Vinju. Source code analysis in Rascal (Tips and tricks).
PAV J17	Feb 26	Code duplication
PAV L10	Mar 2	Software repositories. Program differencing.
PAV J17	Mar 4	Repository mining (I): Social aspects.
PAV L10	Mar 9	Repository mining (II): Files and bugs. Software metrics (I). Metrics and scales. ?LOC.
PAV J17	Mar 11	Software metrics (II). Structure and modularity metrics for imperative and OO-programs
PAV L10	Mar 16	Software metrics (III). Structure and modularity metrics for OO-programs. Package metrics. Metrics aggregation. Churn metrics.
PAV J17	Mar 18	Software metrics (IV). Metrics for cost and time estimation. Tests.
PAV L10	Mar 23	Refactoring.
	Mar 25	No class. Good Friday.
PAV L10	Mar 30	DB migration.
PAV J17	Apr 1	No class.
PAV SH1	Apr 14	9:00-12:00 Exam.
To be announced	June 30	18:00-21:00 Resit.