Born to be geek!

Popularity bias in bug datasets

2011-11-01T00:00:00+01:00

In recent times, the replicability of Software Engineering empirical studies has become a main concern in the research community. One way to achieve replicability is by reusing datasets, so everybody base their results on the same data. However, if these datasets contain any kind of problem, they could cause more harm than benefits.

In the case of software defects, there are datasets that are known to contain bias, mainly when referencing a fix to a particular bug report.

We have studied a different kind of bias: popularity bias. A software project with less bugs is of higher quality. However, in open source software development, more bugs may mean more quality. Why? Because more found bugs imply more people looking for those bugs. This is, if you have no bugs it is because nobody is using your software and reporting them. If you have more bugs, it is because your software is popular; should your software be less popular, the number of bugs would be lower. We have studied this effect in the case of Debian, using the Ultimate Debian Database, and we indeed find that only very popular Debian packages will present a very high number of bugs, and that non-popular packages get very few bug reports.

If you want to know more, read our WCRE 2011 paper, entitled "Impact of Installation Counts on Perceived Quality: A Case Study on Debian". A tag cloud of the contents of the paper:

created at TagCrowd.com

To cite this paper, there is a BibTeX file available, or you can copy from below

@InProceedings{debian_wcre2011,
  author =       {Israel Herraiz and Emad Shihab and Thanh H.D. Nguyen and Ahmed E. Hassan},
  title =        {Impact of Installation Counts on Perceived Quality: A Case Study on {D}ebian},
  booktitle = {Proceedings of the 18th Working Conference on Reverse Engineering},
  year =         {2011},
  publisher = {IEEE Computer Society},
}

IJSODIT - Call for papers 2012

2011-09-29T00:00:00+02:00

The International Journal of Social and Organizational Dynamics in Information Technology (IJSODIT) calls for papers for its 2012 issues.

The mission of this journal relates to social issues in information technology. Social issues are those research topics most aligned with the human factor in terms of information systems planning, development and utilization. This journal includes all aspects of social issues that are impacted by information technology affecting organizations and interorganizational structures. This includes the conceptualization of specific social issues and their associated constructs, proposed designs and infrastructures, empirical validation of social models, and case studies illustrating socialization success and failures. Some key topics may include:

Ethics
Culture
Relationships
Human interaction
Security
Design
Building relationships
Diversity in the IT workforce

This journal follows a full blind peer review process.

More details about the call for papers.

The interplay between businesses and open source

2011-09-08T00:00:00+02:00

This month, the IEEE Software magazine comes with an interesting article about the impact and possibilities of different open source licenses on business models. The paper is available at the IEEE Digital library: Matching Open Source Software Licenses with Corresponding Business Models. From the abstract:

Scores of software producers have turned toward open source licenses
to improve service for their customers. For these companies, choosing
the correct license determines business success. When the available
open source stack and licensing options grow, so does the need to
understand the interplay among licensing, sourcing decisions, and
business goals. A model of license choice emphasizes different
licenses and rationalizes the choice of an open source software (OSS)
license. This is crucial for smaller companies and start-ups that
don't have the tools and knowledge to perform a thorough investigation
of all the consequences of their license choice every time they employ
OSS.

Furthermore, the Computer magazine also brings an interesting article about how to manage open source projects, from a point of view of a software firm. It is also available in the IEEE Digital library: Controlling and Steering Open Source Projects.

Please bear in mind that the IEEE Digital library is a paywall. Send me a message if you want to have a look; I can send you a copy if you cannot access the papers.

Software and the game of life

2011-07-29T00:00:00+02:00

When I read The Recursive Universe I discovered that there was much more than I thought behind the Conway's game of life. The game itself is very simple: draw a diagram in a 2D matrix where each cell can have only one of two different states (dead and alive), and let it evolve according to the following rules:

If a cell has less than two neighbours, it dies
A cell with two or three neighobours survives
If a cell has more than three neighbours, it dies
A dead cell with three neighbours revives

The rules are applied again to the resulting matrix, in a recursive manner.

With these simple four rules, you can obtain a large amount of patterns, some of them very complex. Some patterns are still forms, that is to say, static diagrams that do not change no matter how many iterations are executed. Some patterns oscillate. Some patterns move and live forever. Some of them just die after a few iterations (called generations in the terminology of the game).

The complexity that is obtained in some of the patterns is very surprising when someone first knows of the game of life. Yet it is just the consequence of the recursive algorithm.

The interesting point is that the game of life is a universal Turing machine, and the 2D diagrams are programs that are executed by the Turing machine. Therefore, the game of life can perform any computation that can be done by a Turing machine, that is to say, any computation that can be expressed in a Turing complete programming language.

Is there any relationship between the recursivity-born complexity of the game of life and the complexity and problems to estimate software in the real world? Some software engineering researchers (myself included) are trying to solve the holy grial of software estimation, to try to cope with the problems that complexity conveys in software management, planning and estimation (I call it holy grial because I doubt it will be ever solved, but if it is solved, it would have a big impact in the way we engineer software).

By looking at the complex patterns that the game of life produces, it is very difficult to infer the four simple rules that are behind the game. Are we doing that in software engineering research? Is software engineering trying to deduct the rules by just looking at the patterns produced by the software process? Is the programming activity a recursivity-born complex process? And if it is, how can we unroll this recursivity so we can discover the simple rules that are behind the programming process?

What's the distribution of software size?

2011-07-20T00:00:00+02:00

The statistical distribution of software size can be used to estimate a priori the effort and size that a project will take. In the research literature, the software size distribution has been found to be lognormal. However, we have found that the lognormal distribution systematically underestimates the size of software, and that the double Pareto distribution seems to fit better the size of source code files.

If you want to know more, read our ICSOFT 2011 paper. Before reading it, you may want to have a look at slides of the presentation at the conference, or at the tag cloud generated from the paper text:

created at TagCrowd.com

To cite this paper, there is a BibTeX file available, or you can copy from below

@InProceedings{herraiz2011:software_size,
  author    =    {Israel Herraiz 
                  and Daniel German 
                  and Ahmed E. Hassan},
  title     =    {On the distribution of source code file sizes},
  booktitle =    {International Conference 
                  on Software and Data Technologies},
  year      =    {2011},
  address   =    {Seville, Spain},
}