Born to be geek!

IWESEP 2010 -- International Workshop on Empirical Software Engineering in Practice

2010-08-23T00:00:00+02:00

I have been kindly invited to be in the Program Committee of the 2nd International Workshop on Empirical Software Engineering in Practice (IWESEP 2010), that will be held in the city of Nara (Japan) next December 7 and 8.

Also, there will be a tutorial about Mining Software Repositories, for the first time in Asia, with lectures by Ahmed E. Hassan (Queen's University), Sung Kim (Hong Kong University of Science and Technology) and Thomas Zimmermann (Microsoft Research and University of Calgary).

For more details, visit the website of the workshop, or download the Call for Papers.

The deadline for abstract submission is October 4, 2010 ~~October 11, 2010~~, and for paper and demo submissions October 11, 2010.

Learning by doing

2010-08-10T00:00:00+02:00

"I hear and I forget. I see and I remember. I do and I understand". That's a very well known quote by Confucius, that is usually mentioned by pedagogues. Basically, it says that most of the times the best way to learn is by doing.

Since the end of July, and thanks to the generous funding of the UAX, I am visiting the SAIL research group at Queen's University. As part of my visit, last week I went to Waterloo to visit the rest of the team that works over there. Jack and Ahmed took me around the campus of the University of Waterloo, and I noticed that there was more campus activity than at Queen's. They explained to me that it was because of the "coop" students. At first, I did not really catch what they meant. What is that "coop" thing? At Waterloo, students have teaching terms and "coop" terms, when they go to companies and work there for six months, as part of their learning experience. They work and attend classes, during all the scholar courses. There are different options, like working only during summer terms, or doing it alternate, which means that some students have to attend courses during summertime. Basically, there is always a teaching semester, even in summer. And both attending classes and working count as part of their studies. Interesting, huh?

According to Wikipedia, Waterloo has the largest "cooperative educational" program in the world. There is a full article in Wikipedia about cooperative education, full of interesting references and experiences about the topic. It seems that it started in the United States at the beginning of the 20th century, with the goal of improving the learning process of engineering students. They realized that those students who had to work to cover their expenses were achieving better academic results and more job opportunities.

Actually, last weekend, while hanging out in Kingston, I found a group of students with a "Formula 1"-like car, and they explained to me that they built the car themselves, as part of their undergraduate studies on Mechanical Engineering. They actually have created a racing and engineering team, and were ranked first in Canada in the competition. The car even had some telemetry systems, that they used to tweak the car to the different circuits where they have to compete. They car cost about 50,000 CAD to build, with most of the budget coming from Queen's, but also with funding coming from different sponsors. I was amazed that a bunch of guys in their early twenties could have built such a car, which was by all means equivalent to the cars that we see in the Formula 1 racing, and above all, that they did it with so many constraints. They start have the car at the beginning of the course, and in the same scholar course, they have to participate in the competition. If they do that for undergraduate studies, imagine what endeavors will those students will be able to manage in their professional future.

So far, I thought that it was better to delay contact with the real world while studying. If you start to work, it will distract you, you will not get grades as good as you could, and even though you gain in experience, you will loose in grades, that are important for some future options (like applying for a research grant, at least in Spain). But having known about these initiatives, I have changed my mind. Grades, for themselves, are useless, They are just a carrot in front of students that make them concentrate in getting better grades, they are not a learning instrument. In engineering studies (and probably in other studies), learning is achieved by doing, and everything that separate students from doing is wasting the time of professors and students. In my experience as student, and short experience yet as lecturer, those students that go to a company for some time while they are still pursuing their undergraduate studies have to face more difficulties to pass their courses, and some times even the lack of understanding by lecturers, that want them to attend their classes. In summary, the policies that we use for grading, teaching and student practices, discourage students from going to industry to gain experience, because that is usually reflected negatively in the rest of their curriculum.

Next time we face a case of a student who cannot attend classes because she is working (either as a part of the curriculum, or just to earn some money), instead of punishing her, we should think about the goal of their studies, and therefore as our goal as teachers. Is it their learning or our grading? Is it more important to attend our classes or to be part of the real world of the profession out there? And should we discourage or encourage that? Do we want them to hear and forget, see and remember, or do and understand?

Data for Mining Software Repositories

2010-06-25T00:00:00+02:00

Last week, Daniel Rodríguez (Information Engineering Research Unit, UAH) visited our department to talk about how to start to collaborate in the field Mining Software Repositories, where to get data, what topics we could do join works on. I prepared a set of slides with practical information about datasets, conferences and journals, to be used as a facilitator for discussion. The slides are available in SlideShare:

Mining Software Repositories

View more presentations from Israel Herraiz.

The presentation contains some links to datasets that can be easily used for empirical studies, and that makes it possible to conduct replicable studies. Also, there is paper at MSR 2010 that describes the data sources used for the MSR Challenge; the paper is entitled Mining Challenge 2010: FreeBSD, GNOME Desktop and Debian/Ubuntu and contains description of the FreeBSD repositories, of FLOSSMetrics data about GNOME and of the Ultimate Debian Database. If you use the paper for your research, please consider citing it (download the BibTeX citation as text file):

@InProceedings{challenge_msr2010,
  author    = {Abram Hindle 
               and Israel Herraiz 
               and Emad Shihab 
               and Zheng Ming Jiang},
  title     = {Mining {C}hallenge 2010: 
              {F}ree{BSD}, {GNOME} {D}esktop 
              and {D}ebian/{U}buntu},
  booktitle = {Proceedings of the 
              7th IEEE International Working Conference 
              on Mining Software Repositories},
  pages     = {82--85},
  year      = {2010},
  publisher = {IEEE Computer Society},
}

The eye of the tiger: agile methods vs. architecture

2010-06-21T00:00:00+02:00

As part of the course on Software Engineering that I teach at UAX, we engaged in a research review and debate about how can agile methods and architecture coexist, and which one is better for particular cases.

To prepare the debate, we read a set of papers, most of them from IEEE Software, about software architectures and agile methods. From the list of papers that I preselected, the students decided to read the following:

We also read a IEEE Spectrum's paper about the Virtual Case File, a (in)famous example of software project that failed, partly because of the lack of flexibility that agile methods foster:

Who Killed the Virtual Case File?

The debate was quite intense, and we finally ended up concluding that there is no silver bullet, sometimes it is better to use agile methods, sometimes it is not, and it is always good to know that there are different approaches to software development, and that new projects should be faced without prejudices. I also tried to highlight that software development is a social activity, and that code, documentation, etc, are all important, but it is even more important to be open to the rest of people, and to communicate with the rest of the team.

I am very happy with the results, everybody enjoyed the activity, and I am very glad that they all read so many papers in English (being most of them Spanish, and one French), with a critical look, and using them as arguments for the debate. We also "discovered" the IEEE Software magazine, and other sources of information, that are available to anyone through the institutional suscription of the university, and that contain very useful practical information for the software professionals that they will soon become.

Code as design. Or what's the point of Software Engineering?

2010-04-06T00:00:00+02:00

As a part of the course on Software Engineering that I have started to teach at UAX, my students, some fellow lecturers and I will engage in a debate about the past, present and future of Software Engineering. In class, we have just finished learning about software architectures, and object oriented design, with all the heavyweight usual design process. In the next weeks, we will learn about agile methods, extreme programming, etc.

Last week we all read Code as design, a set of articles about how software design is bloated, and about using software code directly as design, without all the overhead documents produced while applying standard design and development processes.

In the next weeks, we will read some other papers, blog posts and other sources, that have dealt with this issue before. The paper by Tom DeMarco in IEEE Software Aug 2009 issue is worth noting. It triggered some comments about the death of Software Engineering, and how software design and development is rather craftsmanship than engineering, and wondering what's the point of Software Engineering.

Interestingly, this month's IEEE Software issue is entitled Architecture and agility, and it is about the coexistence of agile methods and architecture. I have not read the issue yet, but it looks like it advocates for a peaceful coexistence of both.

Most of the students have a strong opinion in this debate: they are on the side of heavyweight methodologies. I am probably a little biased about lightweight and agility, but don't get me wrong, this will not be a fight heavy vs. light methods. We will try to obtain a set of guidelines and recommendations about what methodologies are better in what cases, and how existing processes can be improved, taking in account existing successful experiences that would have been predicted as a failure using traditional engineering beliefs.

We will probably also add a grain of free software / open source software development, both on the side of heavy and lightweight methodologies, and also both for successful and failed projects.

We will add our conclusions and materials used during the debate to a wiki page, that will be published once we get finished (I will based part of my evaluations using the activity of that wiki, so it will not be public before finishing the course to avoid interferences).

If you want to help or collaborate, and know interesting examples to use for this debate, a enlightening paper or whatever, please add a comment to this post. Your help will be very much appreciated.