FLOSS for Science

In episode 21, we interviewed Juan Pablo Carbajal, an Argentinian physicist currently working as a postdoctoral researcher in the Department of Urban Water Management at the ETH domain in Switzerland. We had a great discussion about GNU Octave and how it can help scientists. We compared its core functions and its expandability through packages to its commercial equivalent Matlab and its toolboxes. An interesting feature of GNU Octave that we explored with Juan is the possibility to migrate code from Matlab directly to GNU Octave and to a certain point maintain code compatible with both. Juan shared with us that since the introduction of an integrated GUI in 2015, he noticed a continuous growth in popularity for the project. We then discussed about a few of the reasons why companies are interested by GNU Octave and why universities should teach using free/libre software. Before asking our usual quick questions, Juan talked with us about the reasons why FLOSS is important for science and the importance of exposing non-FLOSS users to the benefits of FLOSS.

00:00:00 Message to our listeners

00:00:29 Intro

00:00:45 Introducing Juan Pablo Carbajal

00:01:32 30 seconds elevator pitch for GNU Octave

00:02:20 How does the Octave programming language compares to other common programming languages

00:03:23 Compatibility between GNU Octave and Matlab

00:06:29 Matlab's toolboxes compared to GNU Octave packages

00:07:31 Simulink models with GNU Octave

00:09:06 Parallel processing with GNU Octave

00:10:40 The issue with CUDA in GNU Octave

00:11:48 How GNU Octaves differs rom other open source Matlab equivalents

00:13:34 Syntax compatibility to ease transition and reusing code from Matlab

00:15:11 Resources to start using GNU Octave

00:16:40 GNU Octave's graphical user interface and the old QT Octave GUI

00:20:14 GNU Octave's graphical user interface compared to Matlab

00:22:11 Why GNU Octave and not simply Octave

00:23:06 GNU Octave licence

00:24:01 How often he uses GNU Octave

00:24:18 Juan's numerous contributions to the project

00:25:27 GNU octave for companies

00:27:45 Arguments for teaching with GNU Octave instead of Matlab

00:29:32 How many are involved in the project?

00:30:37 Communication channels within the project

00:31:34 Is the project actively looking for developers?

00:32:11 Skills required to contribute

00:33:14 The two-level language dilemma

00:34:59 Juan's vision about FLOSS and its importance for science

00:37:09 Possible negative impacts of FLOSS and converting non-FLOSS users

00:40:17 The most notable scientific discovery in recent years

00:41:46 Juan's favourite text processing tools

00:42:38 Things we forgot to ask about

00:43:57 Anything else to share?

00:44:25 How to contact Juan

00:44:50 Outro

In episode 20, we interviewed Arfon Smith, Editor-in-chief of the Journal of Open Source Software (JOSS) and Head of Data Science at the Space Telescope Science Institute. We talked with him about the creation of JOSS and its role in peer reviewed publications of research software. He described to us how to start a peer reviewed scientific journal and the challenges and competition that he faces. We chatted about how, by using automation, reusing existing tools and staying nimble, JOSS manages to publish with low operational costs. We also discussed about the submission process and what is reviewed during the peer review. Finally, we had a discussion about the importance of FLOSS and in science and how science and FLOSS could inspire each other.

In episode 19, we interviewed Glenn Cooper the Head of experiment computing department at Fermilab about the history and future of the linux distribution Scientific Linux. The discussion started with a brief overview of current research activities at Fermilab. When then extensively talked about Scientific Linux, its goals, the reason why it was started and what made it successful. Glenn made a good argument about the need for a stable software platform in science. We then switched topics and discussed about the recent announcement in regard to the end of the project and the motivations for that. We then talked about the transition to CentOS at Fermilab and CERN and the upgrade path for current users. We concluded the interview with our usual quick questions.

In episode 18, we interviewed Jesse Allison an Associate Professor of Experimental Music and Digital Media at Louisiana State University. We had a great discussion about his work to expand the experience and interactivy of music and sounds. He shared with us his view on how sciences and arts intersect in his projects. He listed some of the tools he uses to create sounds and how these can be applied to create sonic intervention. We had a chat about his past experiences regarding open source software and running a business based on it. We also questioned him whether or not anyone could or even should start creating music. The episode concludes with a small audio sample from his 2013 TEDx talk.

In episode 17, we interviewed Adrian Serio the Scientific Program Coordinator of the STELLAR group about the C++ Standard Library for Concurrency and Parallelism (HPX). We started with a general discussion about parallel computing, where it comes from, where it is going and what can we still expect to gain. We then clarified what are C++ standards and how HPX is developed to be standard compliant. HPX was compared to other parallelism libraries such as MPI and we learned that HPX is a foundation to develop other software for domain specific applications. Adrian informed us how HPX can be used to take advantage of hardware accelerators such as Intel Xeon Phi or GPUs. We looked at the inception of the project and the sources of contributions to the project

For episode 16, we interview the Research Associate Professor Karl W. Schulz. The episode starts with a discussion about High Performance Computing and how OpenHPC facilitate the managment of computing ressources. We then open the discussion towards open source tools, how they became so important for HPC and the their importance for open science. We also discussed about the inception of the OpenHPC project and its governance structure. We end the interview with our usual question in addition to a totally new one.

For episode 15, we interview the Associate Professor of Archaeology Ben Marwick. We start our discussion with an overview of some FLOSS tools he uses and how much FLOSS are used in archaeology. He shares with us his experience in regard to working completely in the open with GitHub and his hope that open science will become the norm in the future. We also discuss about rrtools and his propositions on how to greatly improve the reproducibility of science. As a closing though he shares with us his arguments why early career researchers should invest time to learn and transition to FLOSS tools.

In episode 14, we interview Pat David a Free Software advocate, occasional photographer and engineer about the GIMP project. We talked about how GIMP can be used by scientists to enhance their images for their publications. Also, Pat shared with us his strong opinions regarding scientific communication and why free software matters. You will also learn a few interesting trivia about the origins of the GIMP project, including the content of the original announcement email.

For episode 13, we interviewed Martin Owens, a British Free Software contractor based in Boston Massachusetts, during the 2018 Kiel Inkscape Hackfest, about the Inkscape project. After a brief presentation of Inkscape, we talked about the svg file format and the compatibility of Inkscape with other commercial proprietary offerings. He presented us a list of features planned for Inkscape 1.0, the next major version. We discussed about Inkscape in science and the wider importance of free software and open science. Follow us to be informed when we will release our next episode about the GNU image manipulation program aka GIMP.

In Episode 12, we interviewed Kenneth Hoste, who is an HPC system administrator at Ghent University in Belgium. We talked about the problems that arise when installing scientific software on HPC systems and how the EasyBuild project that originated at HPC-UGent helps dealing with those problems, the impact of open source software in scientific research, as well as the yearly FOSDEM open source software meeting in Brussels where he is actively involved in various ways.