FOSDEM is the biggest free and non-commercial event organized by and for the community. Its goal is to provide Free and Open Source developers a place to meet. No registration necessary.


Interview: Juan David Gonzalez Cobas and Javier Serrano

Juan David Gonzalez Cobas and Javier Serrano will give a talk about "A strategy for managing diverse equipment in the CERN controls group" at FOSDEM 2012.

Could you briefly introduce yourself?

I am responsible for the low-level software development activities at the Hardware and Timing section of the Control Group at CERN. Javier Serrano, who will share floor in this talk, is the section leader. Our work at CERN, in connection with our talk, is twofold: first, we provide hardware and software support for the control system behind all the accelerator complexes at CERN; second, our section is responsible for providing the timing. This is the heart whose beats orchestrate every action in the system: when actions like the ramping of a magnet, the acquisition of data at sensors, or the extraction of the beam, have to happen.

What will your talk be about, exactly?

We must first say that the scope of the talk is very broad, and we are showing the work of many people involved in it. The core subject is the development of a sizeable collection of components for the accelerator control system of the next decade (maybe decades), the design decisions taken in connection with this, and how we deal with them all, from a technical point of view. You may identify three core axes in the talk: the Open Hardware initiative, which is the backbone on which all our developments rely; the White Rabbit timing network, and the hardware toolkit that supports acquisition of data, timing reception and other control system features, with the software architecture associated to it.

What do you hope to accomplish by giving this talk? What do you expect?

To raise interest around the Open Hardware initiative; to show how it works, how we manage to work in collaboration with companies (and with ourselves :D); to get more people interested and hopefully involved; and, last but not least, to exchange ideas with the Open Software community, still the model for the idea of Open Hardware.

The Open Hardware Repository was inspired by the success of open source software. But there are fundamental differences between hardware and software, such as tangibility and manufacturing costs. How does developing open hardware differ in practice from developing open source software?

The main difference between hardware and software is that producing hardware takes more resources and time than compiling software. One consequence is that companies are an even more vital ingredient in Open Source Hardware (OSHW) than they are in FOSS. All considerations relating to freedom of both developers and users apply without any change, so all we have learned from FOSS in that context is very relevant. Free as in free beer is impossible in hardware, so it's easier to focus on freedom and avoid misunderstandings. OSHW development in places like proceeds in a very similar way to typical FOSS projects. There are repositories for designs, mailing lists, etc. When it comes to producing circuits, one or more companies do it. It is important to find companies which understand the advantages of working in such an environment, and we have had no problem in that respect so far.

Last year CERN launched its Open Hardware Licence (OHL). Why was a new license needed and in what ways does this license differ from the GPL for software?

GPL does not say anything about manufactured goods. The same applies to Creative Commons licences. The closest hardware licence to what we needed was the TAPR OHL, but CERN is an International Organization and there are some important things we needed to cover in the way disputes are settled. In addition, we decided to open the drafting process so that everyone could give their input and make the CERN OHL the best possible Open Hardware licence. You can see progress towards version 1.2. of the licence in the OHR itself.

Can you tell us something about the openness of the development of the White Rabbit and ZIO projects? How many developers are working on it, from how many companies are they coming, how do they collaborate in practice and what are their business models?

White Rabbit is a multi-company, multi-lab collaboration whose aim is to extend Ethernet to solve problems in distributed real-time systems. The current list of potential users can be found at its wiki. Many of these users are developers as well, or will become so in the future. Companies involved in White Rabbit get paid for their design time right away, i.e. they do not count on future sales to cover their current expenses, but there is no doubt that they will be in a privileged position to commercialize White Rabbit solutions. You can see a list of involved companies in the main White Rabbit project page.

As for ZIO, it is a project hosted in the OHR repository, and it follows the same, standard pattern. I would not say it differs essentially from any other Open Source project. The main contributors are currently Alessandro Rubini and Federico Vaga. As it is essentially a kernel framework for analog and digital I/O, the usual culture and rules of kernel development apply.

Have you enjoyed previous FOSDEM editions?

No, this is the first one we attend. We are very excited and have great hopes for this conference, because the more people know about, and get involved in, what we are doing, the better. Thanks a lot!

Creative Commons License
This interview is licensed under a Creative Commons Attribution 2.0 Belgium License.