CERN’s Open Source Program Office
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 43 | |
| Author | ||
| License | CC Attribution 3.0 Unported: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/67114 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
FOSS Backstage 202433 / 43
8
11
00:00
Open setOpen sourceComputer programMusical ensembleLevel (video gaming)DemosceneConfiguration spaceBitPresentation of a groupData managementOffice suiteComputer scientistProcess (computing)System administratorComputing platformComputational scienceOpen sourceComputer programmingComputer programmingComputer animationLecture/Conference
01:13
Particle systemForschungszentrum RossendorfLine (geometry)Self-organizationPresentation of a groupUniverse (mathematics)Core dumpStaff (military)
02:18
Open setDegree (graph theory)Point (geometry)CurveField (computer science)AreaParticle systemNumberTrajectoryOrder (biology)Electric dipole momentMetreSpacetimeVacuumVirtual machineCollisionPhysicistInheritance (object-oriented programming)
03:39
Particle systemComputerEinstein field equationsVery-high-bit-rate digital subscriber lineNichtlineares GleichungssystemRight angleVirtual machineGroup actionMassLogical constant2 (number)CollaborationismMultiplication signState of matterInteractive televisionMedical imagingPhysicistTrajectoryFood energyComputer scientistPoint (geometry)Frequency
06:04
Open sourceMereologyTrajectory1 (number)Particle systemMixed realityInteractive televisionPoint (geometry)Computing platformBuildingOperator (mathematics)Operating systemTape driveProcess (computing)PhysicistComputer architectureSoftwareUniverse (mathematics)Structural loadMiniDiscDivisorInterface (computing)Local ringData storage deviceData center
07:55
Self-organizationNuclear spaceOpen sourceOpen setZeno of EleaSoftware frameworkComputer hardwareWeb portalWeb 2.0Self-organizationResultantBitAreaPhysicistWeb portalSoftware frameworkComputer simulationOpen setReal numberRow (database)Computer hardwareSoftwareData management2 (number)Open sourceOrder (biology)MereologyShared memoryRaw image formatSinc functionComputer animation
10:31
Gamma functionCompilation albumFunction (mathematics)Open sourceTask (computing)Open setCoroutineSoftwareCalculationWeb 2.0Connectivity (graph theory)Core dumpComputer hardwareData managementLibrary (computing)Open sourceComputer scientistAxiom of choiceMultiplication signBerners-Lee, TimLatent heatPhysicistComputerExpert systemChainPublic domainOrder (biology)IterationForcing (mathematics)Integrated development environmentMathematical analysisTape driveComputer simulationRevision controlForm (programming)Computer animation
14:33
Service (economics)SoftwareEuclidean vectorStandard deviationComponent-based software engineeringSoftware developerArchitectureOpen sourceData managementData storage deviceData analysisRootExploratory data analysisPhysicsFood energyComputer hardwareOnlinecommunityComputational scienceData centerComponent-based software engineeringSoftwareData managementService (economics)Product (business)Archaeological field surveySelf-organizationRoutingOpen setMultiplication signConnectivity (graph theory)Shift operatorTunisLibrary (computing)Data storage deviceOrder (biology)1 (number)Standard deviationSoftware developerMassPerformance appraisalFreewareSound effectKernel (computing)Computer hardwareWhiteboardCommunications protocolStack (abstract data type)Data analysisComputing platformOpen sourceConfiguration spaceProjective planeImplementationCollaborationismComputer animation
18:43
Open sourceFormal grammarTask (computing)Event horizonTrailSystem identificationService (economics)Data managementThermal expansionSoftware developerPhysicistDifferent (Kate Ryan album)WebsiteWhiteboardSelf-organizationFocus (optics)RootProjective planeMultiplication signPoint (geometry)SoftwareHeat transferForcing (mathematics)Service (economics)Self-organizationMechanism designData managementMoment (mathematics)FeedbackCASE <Informatik>Presentation of a groupOffice suiteTask (computing)Event horizonMachine codeMereologyWhiteboardExpert systemElectronic signatureProcess (computing)1 (number)Connectivity (graph theory)Disk read-and-write headData flow diagramPhysicistLattice (order)Open sourceUniverse (mathematics)FreewareGroup actionOperator (mathematics)Video gameComputer hardwareWebsiteGoodness of fitExperimental physicsSoftware engineeringBusiness modelComponent-based software engineeringComputer programmingKnowledge baseLevel (video gaming)Open setComputer engineeringSingle-precision floating-point formatComputer animation
27:21
Wave packetMeasurementLibrary catalogMetric systemSelf-organizationOpen setSoftwareLibrary catalogChaos (cosmogony)Projective planeWave packetRevision controlComputer hardwareComputer programmingWhiteboardData managementSet (mathematics)Metric systemConsistencyInstance (computer science)10 (number)Term (mathematics)Student's t-testPattern languageTraffic reportingSoftware developerSubsetOffice suiteMultiplication signMachine codeOpen sourceComputer animation
31:38
Computer-generated imageryOpen sourceOSI modelSlide ruleOpen sourceComputer programmingOnline helpPresentation of a groupOffice suiteBuildingComputer programming
32:15
System programmingComputerEquivalence relationSelf-organizationConfiguration spaceData managementCentralizer and normalizerSoftwareInternettelefonieData storage deviceVirtual machinePhysicalismCuboidMeeting/InterviewLecture/Conference
33:40
Presentation of a groupOpen sourceLecture/ConferenceMeeting/Interview
34:12
Inclusion mapLibrary (computing)Open setData managementMereologyBitGroup actionComputer hardwareOpen sourceWhiteboardProjective planeBusiness modelProper mapMeeting/Interview
36:04
Survival analysisSelf-organizationMomentumQuicksortAddress spaceDevice driverProfil (magazine)Event horizonDifferent (Kate Ryan album)EmailMultiplication signProcess (computing)Modulare ProgrammierungDomain nameOpen sourceEscape characterLecture/ConferenceMeeting/Interview
38:07
Touch typingQuicksortProjective planeOpen sourceSelf-organizationState of matterMachine codeBinary codeLibrary (computing)Process (computing)Point (geometry)View (database)Natural numberCASE <Informatik>Meeting/Interview
39:59
Control flowMusical ensembleLecture/Conference
Transcript: English(auto-generated)
00:08
Giacomo, hello from Berlin, can you hear us? I can hear you, can you hear me? Yes, we can hear you loud and clear. I would say the stage is all yours. Okay, thank you, good morning.
00:23
My name is Giacomo Tenaglia, I'm a computer scientist working at CERN. And while my daily job is like a system administrator and in the team providing the scientific computing platforms for the CERN community and the config management tools.
00:44
I'm also working part-time at the CERN Open Source Program Office. And so I'm happy to be virtually here with you today to present you a little bit what is going on, what has gone on behind the scenes.
01:02
So a bit of the backstage of our initiative to start an Open Source Program Office at CERN and a little bit of a future outlook. I'll start with a short presentation about CERN. So we are an international particle physics laboratory
01:22
located in Switzerland and France. So actually the dotted line you can see here is the border between the Swiss side that contains the lake and the French side. And we've been founded in 1954 and the core mission of the organization
01:42
is to provide, is to make like fundamental research to uncover what the universe is made of and how it works. So in practice, we study the very small to understand the very big, this is what they say. And we have around 2.5,000 staff
02:00
and 12,000 researchers from around the globe. So the idea is that we maintain and run a laboratory. And when you do your research at your organizational research center and then you come over to actually do the experiments live at our lab. So our flagship machine is a 27 kilometers
02:22
particle accelerator called the Large Hadron Collider. It's located around the 100 meters underground. And there we accelerate charged particles to nearly the speed of light. And we steer the trajectory with the large magnets that you can see in this picture.
02:41
So the blue one is the, there's about 1200 of them in the tunnel. And this is a dipole magnet that serves to curve the trajectory of the particles. There's a number of challenges with that. Mainly the fact that the particles go very fast so you need a very high magnetic fields
03:01
to change to bend their trajectory. In order to do this, you need superconducting magnets that need to be very cold, less than two degrees Kelvin. And so this is colder than outer space. And in the beam pipe, you need to have a vacuum otherwise the particles collide with air particles.
03:21
This, okay, so there's a number of technical challenges. And the interesting thing is that the two particle beams that you see in the picture, they cross over in four experimental areas in four points on the 27 kilometers tunnel where physicists have built gigantic machines
03:44
that are called detectors. And so I had to put, so I'm not a physicist, I'm a computer scientist, but I had to put the Einstein equation there because it's actually very interesting what happens inside these detectors
04:02
once we smash two beams of particles at a very high speed. So the Einstein equation says that mass is equivalent to energy because C is the speed of light is a constant, but it's a very large constant. So typically we see it in action from right to left.
04:22
When we burn some mass, we generate some energy, but we never see it on the other from happening from left to right because we live in a world at the low energy, let's say. So in the particle accelerator is one of the places
04:42
where you can actually see it happen from left to right. So when we smash two beams of particles that are running at high speed, high energy, basically you see new particles appearing for a brief period of time. So it's like accelerating,
05:03
it's like smashing together two apples and seeing one banana and one pear coming out, not just broken or pieces of apples. This is not a very, I apologize to the physicists in the audience. So basically what happens is that in those four interaction points, we built some international collaborations
05:22
that are LHC experiments, built gigantic cameras that can measure basically what happens. So the energy, the charge, and the trajectory of the new particles. So every piece of those machines is able to measure one specific thing. And this generates around 40 million images
05:43
of the state of the detector per second. In fact, most of the stuff is boring. And so it's not kept. There's like some clever electronics that are able to filter out stuff that is not interesting and is not kept.
06:04
So in practice, then it looks like this in the reconstruction. So basically you can see the interaction point in the center and then the various parts of the detector that are switched on by the particles traveling through them.
06:24
The particles that bend the trajectory are charged, the other ones are not, et cetera, because there's like big magnets also inside the experiments. So this, as you can see, generates a lot of data. So we've been operating
06:42
building and operating for the last 20 plus years, a worldwide distributed computing platform that is called the WLCG, the Worldwide LHC Computing Grid. And it is basically a network of data centers
07:00
hosted at research institutes, universities around the planet, and that are able to, that are distributing the load of processing and storing and analyzing the data produced by the LHC experiments.
07:22
And the physicists, they have a unique simple uniform interface to submit jobs that are then executed in various parts of the grid, depending on many factors. You can select the architecture, select the operating system, et cetera, data locality, and so on.
07:44
So there's around more than 1,000 petabytes of CERN data are basically stored also worldwide on disk and on tapes. We also still use tapes because they are cheap. One important part of the mandate
08:00
of CERN as an organization that is written in the convention, is that the results of the research have to be published and made generally available. So in order to understand this, we need to think about how the organization
08:21
was funded right after the Second World War and when Cold War started to kick in. And the idea was to build a laboratory for science for peace. So an essential ingredient was to be able to publish also all the results of the research to the whole planet and not keeping them secret.
08:43
So this has been really an important guiding principle throughout all the history of the organization. And this has led to initiatives that are then, you can see a little bit in this timeline from, I mean, one of the most famous one
09:01
was the invention of the web in 1989 to share documents between physicists and scientists. It obviously was a successful one. We also have been active in the open access area
09:22
since the, over the last like 10, 20 years in basically in lowering the, in getting rid of the paywalls for accessing the results of CERN research. On the open data side, we have launched
09:42
like an open data portal in 2015 where the experiments can publish their raw data for people to take a peek, analyze and run some simulations at home and this kind of stuff. And also to reproduce, it's very important to be able to reproduce
10:01
and reuse the data. So we have a dedicated framework that we developed. And in 2022, we got approved by the CERN management and open science, real open science policy that is covering all the aspects of open science including open source, hardware and software.
10:23
And for this, we got the OSPO approved last year. I want to just take also a step back and look at open source specific milestones. And we can see that the culture of sharing on the, at CERN started in the,
10:43
in the seventies we had the first iteration of the CERN School of Computing that is still running. And the idea was that computers started to really take over the scientific landscape and CERN thought we needed to educate scientists,
11:04
computer scientists and engineers and physicists to the usage of tools for high energy physics. And at the beginning of the eighties, while the world was starting to adopt free and open source software concepts,
11:22
CERN and other laboratories decided to share two very important pieces of the tool chain for running analysis. And one was a tape that was called the HEP-VM, high energy physics VM for VMS
11:41
that was helping, that was containing all the environment to set up simulations and run them. And the other one was CERN-LIB, which is the core library of routines used to actually run the scientific software and calculations. So some of those routines, interestingly enough,
12:02
are dating back to the, to 1966, 68. So this is really predating some of the, most of the world's maybe open source. In more recent times, after the web was released in 1989
12:21
in the public domain, and after Tim Berners-Lee left to MIT to go fund the W3C, the team at CERN decided that public domain was not good enough and re-licensed the web components
12:41
as open source with a dedicated CERN license. This is something that you would not do today, but it was deemed a good choice at the time in order to prevent companies from just taking the software and make it proprietary. Then this was later re-licensed with the MIT license.
13:06
And in the last, then after that, our colleagues making custom hardware for CERN accelerators. There's a lot of custom, of course, stuff that we have to design.
13:22
Decided, started to adopt also concepts similar to those of open source. So the four freedoms of software, but just adding the freedom to manufacture the hardware, to share their designs. And in 2011, we, they wrote the, Javier and his team wrote the CERN open hardware license
13:43
that is now at this second version, has been published a couple of years ago. And shortly after that, we got the blessing from the management in the form of a task force. So when we get an official task force, it means you basically made it. And that wrote recommendations
14:01
to release CERN software under OSI approved licenses. So this was the first official recommendation that we, that we had. And interestingly enough, the team that wrote the, the recommendation was composed of also Javier
14:21
from the open hardware and Francois Flueckiger, who was responsible for licensing the web in 1994. So some of the experts who've been doing open source for years. Then of course, in the last 10 years, things escalated quickly. And we ran a little survey last year about the service
14:42
on the, on the IT department and the services run by the IT department. And we, and service managers were declaring that basically 70%. So we are, we are a kind of a classic IT department running everything from network to,
15:00
to our own data center, orchestration stuff. Infrastructure as a service with OpenStack platforms, with OpenShift, we run many software as a service. So we run more, more than a hundred different services and 70% of, of those run, rely on a major free and open source software components.
15:23
So this is excluding like basic common components such as the Linux kernel, et cetera. So, and, and of those around two thirds have a significant certain contributions. So we could show to the, so it was clear,
15:42
it's clear that over the last 10 years, the, the adoption of, of open source as a, as a consumer, but also as a contributor has grown tremendously. And there's been also initiatives to move from homegrown config management and orchestration tools to more standard components.
16:01
We joined also the OpenStack and now the Open Infrastructure Board, of course now using Kubernetes, recently switching to Alma Linux. And, and there's a lot of, there's been a lot of active engagement with, with the user development communities. Another interesting development and side effect
16:22
of this massive adoption of open source has been what you call the MOLT project that started in 2019 to re, to rationalize the provisioning of software licenses and evaluate replacement of proprietary components with free and open source software ones.
16:41
We succeeded in some of the initiatives. We failed in some other initiatives. There were many, many lessons learned on accounting, eligibility, subsidization, especially in engagement with the, the user community at CERN and with the upstream
17:03
communities in order to better understand our needs and how we could convey them in the, in the upstream products. And at the same time on the, on the scientific, on the more scientific side,
17:20
there's been also a shift towards the reuse of CERN made stuff, software and hardware for beyond high energy physics. We have some of the, some examples are our conference management software that is now being used elsewhere
17:41
in other institutes and organizations. We have obviously storage, but also the data analysis and the scientific computing route, which is the successor of CERN LEAP that I mentioned earlier, and is being used also in financial institutes and around, around the globe.
18:01
And also on hardware, we, we actually discovered by chance that White Rabbit, which is a fine, a fine timing, fine tune timing protocol and implementation was used by, I think it's German, a German financial institute by chance.
18:23
So we set up also a collaboration to collaborate with them. And there's been in general, a lot of effort and to do an investment in, in making CERN products, let's say, go beyond high energy physics and be reusable elsewhere.
18:42
And in general, one common thing throughout the last, yes, 10, 20 years has been, there has to be like a more systemic approach to how we do open source. And there's been many questions coming up
19:00
that were slightly similar, how to, from the basic ones, how to license code. People were choosing different licenses, open sourcing in different ways. We needed, we were, we started to be asked from external entities, how to, how we were positioned as an organization in some,
19:22
with regard of some questions on open source. And, and we had some problems of, with the licensing components, how to accept contributions to certain open source projects,
19:40
how to actually contribute to an open source project that requires signature of a contributor license agreement that nobody, this, we spent like a couple of months trying to understand who was actually entitled to sign this and it's still unclear, but no, it's not still unclear. We finally cleared it out, but it's, it took some,
20:01
it took some time to understand and gather experience that is very much distributed across the organization. So we started discussing with a few, with a few experts and practitioners in open source a couple of years ago, the need for creating
20:26
something more formal and at the same time, something more recognizable by the, by the CERN community. And, and we started putting together a proposal
20:40
for an open source program office. And these things take a lot of time in a, in a consensus driven organization like CERN. So you really need to convince a lot of people and make your point, make your case and, and rework various proposals, et cetera.
21:00
So we started, I actually started some of the presentations in December and it was basically not very encouraging the feedback I got. It was particularly complicated to work with our, with our knowledge and technology transfer colleagues with some of the high management
21:21
that we are not really seeing the point in having to set up something for something that we can get for free. And, but fortunately we, that the people at the organization are amazing and they're super motivated. And we managed to have a, to have an official task force launched in October, 2022 with the, with the goal of,
21:44
with a specific task to write the mandate for an open source program office. It took around six months and many, many discussions and meetings. We got our, we got a couple of amazing colleagues from the tech transfer office that helped a lot in,
22:01
because basically it's important to understand that also in, in, in public organizations like CERN or like universities, the tech transfer office is responsible for managing the IP portfolio. And it's tricky to, to go there and say, listen we want to just publish everything for free. So it's, it's, it's challenge.
22:21
And, and I think it's, it's great to have people who are able to understand also open source business models and, and to try to, to have them on board and to, and to involve them. And that they were actually very, very instrumental in getting the mandate approved in, in May, 2023.
22:43
And we had the official meeting in September. Then we had a great logo design. You can see here some sketches. And then we had the launch event in, in November last year. So the, so this is the final logo that we chose.
23:05
And the mandate is, as most of OSPOs is divided in two parts, one internal and one external. So internally we are mandated to basically consult, advise and train, trying to consolidate all the knowledge
23:21
that is scattered across the organization in a single place and to have a single voice and in particular on open-sourcing CERN software and hardware. We are also mandated to track the dependency of the, of the, of CERN and on critical,
23:44
on free and open source software components in particular for critical IT services, critical to the operation of the, of the organization. And this, at the moment we are at this point, we are not at the point of how to do if we identify that the health of a critical dependency is not good.
24:03
So this is for, for a later stage and also to advise the CERN management in case there's open source questions or issues. In general, we try to avoid, one key part of getting the mandate approved was to avoid the other, putting advocating,
24:22
advocating for open source in the mandate. So of course we are all like outsource enthusiasts. So it kind of comes naturally, but the official mandate is excluding, is not specifically talking about advocating for open source. And then in the external mandate is more about showcasing
24:42
the contributions to open source made by the organization and facilitate partnership with external entities and promote the lab as an open source, good citizen in general. So in practice we are, we're a group of 12 people
25:03
who in real life do other things that are typically linked with open source. We have a person who was the project leader of the root software. We have Javier who is the open hardware guru. We have the head of the open science office.
25:22
We have many people doing, we have of course both kinds of physicists, the theoretical and the applied experimental physicists because they both use and produce a lot of open source. We have engineers, hardware engineers and software engineers. And we have our knowledge transfer experts.
25:40
We have been lucky enough to have been joined by a knowledge transfer legal advisor. And already this was last month. I already learned a lot about how to write things, what to write, what not to write in official documents. So it's actually very interesting that every one of us has around 10 to 20% of our time allocated for this.
26:03
We have a few interns as well who are a hundred percent fortunately who can help us. And they're actually a great driving force as well. And, but in general, we operate as a part of our job. So it's not, there's no one fully a hundred percent on this.
26:22
And at the moment we're working on writing a technical website that of course brings up issues of defining recommendations and policies and mechanisms. So how to implement the policies.
26:40
And we are doing this as we serve internal requests. So now of course people have noticed that we have an office and we were happy about this. And at the same time we get like kind of, I guess bombarded by requests all the time. So we're trying to answer them and at the same time
27:02
build a knowledge base and our technical website. So this is good, but it takes time because we also work as a consensus driven board and it's tricky to sometimes to discuss and agree but we eventually are close to publishing our internal website.
27:22
So in general, the, for this year, we plan to focus on three major access. One is to publish our software catalog. So there's around, we have an internal GitLab instance
27:43
with tens of thousands of projects. And we have a few hundreds. If you search on GitHub there's a few hundred organizations that are linked in a way or another to CERN. There's an official CERN organization with a few tens of projects but then there's many other organizations.
28:03
And so we would like to provide another catalog where we showcase basically where projects can submit can opt in to be part of and can be seen all in one place
28:21
and searched all in one place. So this is being developed and will be published later this year. Basically together with the new version of the open hardware catalog that has been worked on and will be published as well this summer.
28:40
Then we, of course, with the mandate, we have been asked to produce a yearly report. And so for this, we need something to measure and which is actually great because we in general like to measure things but also we would like to start providing projects some meaningful metrics and ways of measuring their project
29:05
and observing how they are doing in terms of both quality of their contributions to external, to other projects, or in general how they are doing in the community of their, if they are able to build the community,
29:23
if there's contributions to their project that they are maybe not noticing, maybe there is a pattern that we can notice of other organizations contributing to a set of projects. For this we've been, we are going to look at the chaos metrics over the last of the next few months
29:42
and hopefully integrate all this and deploy some tooling for measuring so we can provide it to the CERN community. And another request that we have outstanding is to set up a training program for our developers,
30:05
probably it will be like a short, like one day training on how to publish code, how to do things properly, how to accept contributions. So documentation is piling up. We are really starting to write a lot of documentation.
30:23
So it will be probably, there is material to write a little internal course to help people. One thing that is very peculiar is that in general development teams change a lot. So people come and go all the time. It's very rare that people stay
30:42
for more than a couple of years. Some people stay just for summer as summer students and they do like significant contributions to some projects, hardware or software. Some people stay for around a year and some people come and go every few weeks or months.
31:02
And so it's very important that practices are kind of consistent across the board. And this also can help us in identifying what's going on and showcasing to the society but also to CERN management that this actually matters and that they did a good thing,
31:21
deciding to approve the creation of CERN Open Source Program Office. So also the training of people, it's instrumental for this. So I think this is more or less everything I wanted to say
31:41
for today to present you about the CERN Open Source Program Office. And I really wanted to thank, I've been in contact with some people that I know are maybe in the audience and I wanted to thank them for their help. And I wanted to also thank my colleagues and my colleague Zuniya who couldn't be here today
32:03
for the opportunity to present and for their help in building this presentation. All right, thank you, Giurgolo. Raise your hand if you have any questions. Since there are no other questions, a quick one.
32:22
I was wondering how many people are involved in maintaining your infrastructure, your IT, all of the systems you mentioned in the beginning. So the systems I mentioned in the beginning are just the grid systems. And on my team, we are about 10,
32:41
but we do many other things. Of course, the grid computing is, I think we have around 15,000 boxes, physical machines. But also we provide the config management tools for the whole organization.
33:01
So we have some central puppets infrastructure and secrets management and many other, many other glues to make everything work. So we are, yes, we are around 10. I would say we're around nine full-time equivalent. The IT department that is developing storage systems,
33:24
conference management systems, running IP telephony systems, network, we are around 250. Yes, this is mostly the IT infrastructure. Then of course, there's IT everywhere. I mean, everywhere, everyone is developing software
33:44
and doing infrastructure stuff, but yeah. Yeah, thank you for the presentation. I was wondering how is the general mindset at CERN? Is it very open to open source or is it we don't really care about it or are they even more into towards closing down stuff
34:03
and in a sense of a competition with other research institutes? Can you say something about that? Yeah, so I think researchers in general are mostly, let's just put this out and publish it. So then open source sometimes is a bit of an afterthought.
34:28
Nowadays, things are changing because it's also great to show to the world that you're doing open source right, so on your GitHub. And if you can show that you contribute to some CERN projects, it's great.
34:42
So people really are pushing to have it done properly. So this on the, let's say on the grassroots side, on the management side, there is more and more like, yes, openness to be open. And of course, because it's also part of the mandate, there's been some resistance on,
35:02
in particular for the hardware side, because it's tricky to convince that we have a great idea that could be reused elsewhere. And it's a piece of hardware that could, we could really maybe like make a partnership and sell the IP. And in fact, we actually want to open source it.
35:22
So it's tricky, but finding, but our colleagues doing hardware are great at finding some business models for this, like launching foundations, that there will be a launch of the White Rabbit Foundation in a few weeks. That is a partnership with other institutes, public and private.
35:42
And in general, in the IT department, this is reality I know most, there is a lot of push for open source and doing it right and contributing back to upstream. So there's some of our group leaders that have been on the open infrastructure board of directors.
36:00
We had the colleague who was on the SEF board of directors and so yes, it's recognized as vital for the survival of the organization. Do you have trouble maintaining sort of momentum on the things that you want to do? You mentioned that the majority of your OSPO
36:20
are people who work in it part-time. So I'm interested in like how you get things done because that's a really difficult situation to be in. Yes, it is. Fortunately enough, we are on a very different domains. So typically conferences and big events happen at different times kind of.
36:41
So there's, but sometimes we kind of organize like one day hackathons to where we put everyone together and then you cannot escape and we get together and we get things moving. Like every couple of months, we had one last month. It was very productive and we, yeah.
37:03
So this was actually a great initiative that came from one of our interns who was proposing this and this helps. And yeah, because 10, 20% of the time is not that much. It's easy to either spend it answering emails or spend or just getting to Friday, 7 PM
37:23
and there's no 10, 20% that went in the OSPO work. So, but also the fact that many people are actually doing open source in their daily job, it's kind of a driver to improve things and to work on things.
37:41
For example, me, I'm working with a colleague who and we need to contribute to some puppet modules upstream and there was a question about whether I should use my private email address, my private GitHub profile, how to handle email addresses. And this was a need that I had in my daily work from my colleagues.
38:01
So we could just make it into a recommendation from the OSPO. All right, a final question from the online audience. You mentioned that when you OSPO became visible, you got lots of requests coming in. What sort of things were you asked in the beginning? So at the beginning, we've been asked
38:21
how to publish our code in general. And this is the most frequent question, which is also as normal. And because people realize that this could have an impact and maybe they were not thinking about it. And then you get, yes, requests to publish
38:40
like projects that can vary in let's say readiness states. So some projects are maybe more ready and some other projects ship maybe other libraries or binary lab view files that you maybe want to not to ship. So they have to refactor and then we have to iterate
39:02
and ask many questions. So after some of the requests, people were basically expecting, we realized that people were expecting kind of you just choose this license and you put it on GitHub on this organization kind of answer. But in fact, we structured the answer asking actually more questions
39:20
about how the project is done, how the project is, what are the dependencies of the project, where it is deployed and other things that people were not really expecting. So some people came back and they said, I was expecting lighter process. But in fact, asking all those questions helped in understanding the nature of some projects.
39:43
And then sometimes the recommendation was maybe it's not a good idea to open source at this point. You should do this, this, this before. And then maybe the team decided not to open source in that case. All right. Thank you Giacomo for your great talk.
40:00
We're gonna go now and take a short coffee break. Thank you. So thank you very much.