Moving to a new place

Finally decided to host my own WordPress and other stuff, so, this blog is now re-located to: http://blog.dizquierdo.es

Are traditional publishers moving to Open Access?

Looking around the Empirical Software Engineering journal from Springer, I realized that there is an option about licensing the work using a libre license (let me use the term “libre”  as  free as in free beer and as in free speech, as usual). At least this is what I understand from the open access website of Springer (hopefully, I’m not missing anything ):

Anyone is free:

• to copy, distribute, and display the work;
• to make derivative works;
• to make commercial use of the work.

In fact, in the current volume: December 2011, the last paper is accessible by everyone.

However, as far as I know, the open access journals usually ask for money to the authors since part of their business model is missed by the “openness” of the papers.

Does anyone know how this works in the case of Springer if I want to license my work under a creative commons  license?

Distributed source code management system patented?

Has been the Git / Mercurial / Bazaar idea patented?.

Some days ago I was surfing Google and Google scholar to look for academic references regarding to distributed software development. However, it was quite surprising to find a patent from the USA describing the way a distributed source code management system work.

Specifically, this patent was born in 1997 and can be found as a Google search.

After having a look at the document, it is still not clear if a piece of software such as Git or Mercurial are kind of “illegal” because of this patent in the USA…

Patents and the mobile market

Quite interesting the post in Google’s blog about their point of view of software patents and how the Android technology is being “attacked” by specific companies that are buying more and more patents.

A couple of sweets:

Patents were meant to encourage innovation, but lately they are being used as a weapon to stop it.

Unless we act, consumers could face rising costs for Android devices — and fewer choices for their next phone.

So, probably Google is exactly pointing to the problem: software patents and how they were supposed to be, but how they really  are.

Let’s think about how this may affect the business model of  SMEs which can not reply as Google, Samsung or HTC do…

Thanks to  Xataca for the pointer.

Educación en el siglo XXI

“Al entrar en el aula casi parece que el conocimiento sea algo escaso, difícil, prácticamente imposible de obtener a no ser que tengamos a un adulto debidamente cualificado en frente” – Curtis Johnson (La manera disruptiva de aprender – Redes)

Creo que esa frase resume básicamente lo que se encuentra uno en la actualidad en cualquier aula. El problema de fondo es: ¿cómo se puede mejorar?.

En el vídeo, altamente recomendable, se habla de que poco a poco, la educación irá cambiando hacia un modelo donde los estudiantes dejan de ser actores pasivos y pasan a ser actores totalmente activos. ¿De qué servirá que un profesor dé una fecha concreta para un acontecimiento importante como la Revolución Francesa, si ese mismo alumno es capaz de buscarlo por sí mismo en Internet o a través de una enciclopedia digital?. Al fin y al cabo, cualquier niño sabe buscar en Internet.

El rol del profesor por tanto, debería tender a cambiar hacia un rol  que ayude a  desarrollar en el alumno competencias específicas, como es el trabajo en grupo, empatía personal hacia situaciones de otras personas o situaciones similares y no a ser meramente un difusor de conocimientos.

El vídeo explica como  el uso de plataformas online permite acercar a aquellos que se encuentran lejos y que incluso permite tener una relación mucho más directa entre el profesor y el alumno ya que activamente se están requiriendo respuestas .

Aunque breve, creo que es una pequeña reflexión que toda aquella persona que actualmente esté trabajando de algún u otro modo de profesor debería hacerse. Cómo mejorar el sistema actual que esté usando para sacar el mayor provecho de cada una de las personas que están atendiendo. De hecho, no todas las personas aprenden de la misma manera, ni a la misma velocidad. Aunque esto último es un tema que debería tratarse por separado.

 

 

QualOSS: Interviewing Companies

This project started around four years ago. I just started in LibreSoft and this was my first project in terms of acquiring experience and working in an international environment. This was in somehow my starting point in real software engineering field, working with companies and academic partners and having real responsibilities.

And here we go: QualOSS basically started as this kind of project start: looking around and checking what could be useful. In fact, one of the main purposes of the project was to have a real industrial application, what implies start talking to companies.

Several companies were interviewed by the QualOSS consortium. Between two or three per partner and main questions were derived from the current state of the art during those days in software quality, but also trying to deal with the idea of open source in all the skeleton.

From my personal point of view, the main goal that we achieved with this approach was to realise that there were kind of three prototype of company:

• First of all: Those companies which do not care open source. If this is useful for the company, good news, if not, bad luck.

• Those companies which base their business model in open source.

• And a mix of the two aforementioned points.

Deepening in the second point, I could say that those companies were pretty interested specifically in the licensing process and the community around a product. In other words: perhaps the product is a great product, but they want a set of good people supporting it in terms of community. Thus this was a key factor.

So, after all, we created a preliminary definition of the main quality attributes addressed in the project (Robustness and Evolvability) based on the interviews:

As we can see, the community quality was addressed as being part of the quality models by some companies, what means that open source does not mean just free software (as in free beer) for companies, but also some other factors are being raised as important enough to be studied.

More information can be found at deliverable 1.2 from the QualOSS consortium.

BlameMe: Storing the diff information

That was some weeks of effort, first of all, to understand how the unified diff format works and then to deal with it and store all that information in a database.

BlameMe comes to help me in the PhD. This tool will analyze all the revisions and for each pair of them, there is always a difference in terms of modifications. That information is stored in a MySQL database.

This is now working with Git, which is highly recommendable to mix with the CVSAnalY results. In the case of Mercurial, there is another script (still work in progress) which stores extra information that it’s not directly retrieved from the diff, such as the committer and the date.

In any case, information for each of the lines is stored (a line could be added, modified or removed) and that information could be found among different files and committed by a committer in a given date).

Finally, it is worth to mention that the selected license was GPL v3.

For further information, please refer to our libresoft-tools mailing list.

Enjoy!

ps: please, don’t be too cruel with me, still in beta version of the tool .

QualOSS: Quality in Open Source Software

QualOSS is a FP6 project funded by the European Commission whose main goal was to go a step ahead in terms of the state of the art with respect to quality in open source software.

This project finished some months ago, but I’d like to keep in mind some interesting ideas that were developed there and that could be useful in a near future:

- Companies took part of the definition of the quality model: what provides other scope different from the academic one.
- There are four main quality aspects to be measured: product, processes, documentation and community.
- For each of them, and following a Goal-Question-Metric approach several metrics were raised to answer all of them.
- We had an industrial validation from a couple of projects that worked with us.

As lessons learned, I’d say that measuring quality is a hard task (wow, what a great conclusion! ¬¬). Even after the project I frankly think that we missed some good stuff, but we didn’t have enough time. We obtained interesting results, but my impression is that we could have done something a bit better, just a small step. However once you are in the diary work is totally opposite your perception and you are pretty busy everyday .

In any case, I’d like to personally thank all the partners that I’ve worked with for their time, patience and professionalism.

ps: after reading again the post, I think that this is more a personal thought about the project than the project itself. I promise to write something more technical in the following!.

Repositories with Public Data about Software Development

I am proud to announce the new edition of the International Journal of Open Source Software and Processes where the WoPDaSD (Workshop on Public Data about Software Development) was invited to present a special issue.

Three papers have been accepted plus one from the organizers (Jesús Gonzalez-Barahona, Megan Squire and me).

Specifically the abstract related to the paper entitled Repositories with Public Data about Software Development is next:

Empirical research on software development based on data obtained from project repositories and code forges is increasingly gaining attention in the software engineering research community. The studies in this area typically start by retrieving or monitoring some subset of data found in the repository or forge, and this data is later analyzed to find interesting patterns. However, retrieving information from these locations can be a challenging task. Meta-repositories providing public information about software development are useful tools that can simplify and streamline the research process. Public data repositories that collect and clean the data from other project repositories or code forges can help ensure that research studies are based on good quality data. This paper provides some insight as to how these meta-repositories (sometimes called a “repository of repositories”, RoR) of data about open source projects should be used to help researchers. It describes in detail two of the most widely used collections of data about software development: FLOSSmole and FLOSSMetrics.

Thanks to Stefan Koch, editor in chief of the Journal, the WoPDaSD PC and all the authors that have helped to have this issue ready .

Mercurial to Git

Why is this important?: Well it depends on your own situation. In my personal case I have developed some Python scripts that work with Git so far. It means that instead of implementing a new backend for Mercurial or other source code management system, I’d prefer to have other tools to do it for me.

However, as usual, I do not know how reliable is this tool, what brings in context new questions related to the reliability of my results.

All in all, using the hg-fast-export tool, developed by different authors, has been a great discovery since it’s really easy to migrate from one kind repository to another.

The steps are next (based on information explained at Minimum Air Induction’s blog):

1- Clone the git repository of the fast-export tool: $ git clone http://repo.or.cz/w/fast-export.git
2- Create your new git repository (where the mercurial repository will be migrated) (e.g. $ mkdir $HOME/new_git_repository)
3- Initialize that git repository ($ git init)
4- And finally, run your hg-fast-export.sh command with the -r option to indicate the mercurial repository and the destination ($ hg-fast-export.sh -r "HG REPOSITORY" "GIT REPOSITORY")