[00:00.000 --> 00:09.500]  So, thank you Paul, welcome everybody.
[00:09.500 --> 00:17.960]  So we will today present to you the Cortex platform, which is an open platform for research
[00:17.960 --> 00:20.200]  in social sciences.
[00:20.200 --> 00:26.280]  And it's a collective presentation where all the team is here, but we have three of us
[00:26.280 --> 00:31.240]  will be presenting today out of the team that you see on the picture.
[00:31.240 --> 00:33.680]  So what is the Cortex?
[00:33.680 --> 00:42.280]  It's an online platform built on top of an architecture and its main goal is to help
[00:42.280 --> 00:49.360]  the social scientists with a research question to actually find the benefit of a computational
[00:49.360 --> 00:55.240]  method to fit a specific question.
[00:55.240 --> 01:04.360]  So it has been founded in 2009 and it's driven by social science research and is supported
[01:04.360 --> 01:12.280]  by several French national institutes at the beginning, in particular INRA, and then later
[01:12.280 --> 01:18.520]  on some funding, some research funding and some European project called RIS, European
[01:18.520 --> 01:22.880]  project of infrastructure.
[01:22.880 --> 01:27.600]  We are now at our second version of this online platform called Cortex Manager.
[01:27.600 --> 01:31.760]  So this is the part that you can actually go online.
[01:31.760 --> 01:36.920]  It has been designed from the very beginning, as I said, for the social scientists.
[01:36.920 --> 01:48.800]  So we took the point of departure was that the social scientist doesn't have any IT capabilities,
[01:48.800 --> 01:57.160]  doesn't have any, on average, doesn't have any resource on the computer, especially 15
[01:57.160 --> 01:58.160]  years ago.
[01:58.160 --> 02:06.120]  So one of the main benefits of the platform would have to be empowering him with the power
[02:06.120 --> 02:10.480]  of computing and of methods.
[02:10.480 --> 02:17.840]  It's a collaborative platform, so a lot of projects can be shared between scientists.
[02:17.840 --> 02:23.680]  And there is, as a web interface connected to a workspace online, you don't have to,
[02:23.680 --> 02:30.160]  of course, have it on your computer, which was, again, quite new at the time.
[02:30.160 --> 02:33.520]  And it's very important to note that it's very permissive.
[02:33.520 --> 02:39.720]  So it's a bunch of methods that you can apply on all kinds of data that we will detail later.
[02:39.720 --> 02:46.080]  And this was one of the main foundation of our platform in the beginning, that all the
[02:46.080 --> 02:53.880]  methods are runnable on all the viable that you have at your disposal.
[02:53.880 --> 02:59.120]  So how is it going from ten, ten, ten, twelve years after that?
[02:59.120 --> 03:06.840]  So we have over 250 peer-reviewed publications that we are counting from now.
[03:06.840 --> 03:10.640]  And of them, 75 documents in 2022.
[03:10.640 --> 03:15.880]  So it's growing, and it's growing quite fast the last few years.
[03:15.880 --> 03:24.200]  So 50,000 analysis, so 50,000 jobs have been done on the platform by scientists last year,
[03:24.200 --> 03:28.800]  by over 1,000 active users, 400 institutions, and in 50 countries.
[03:28.800 --> 03:36.840]  So it's pretty worldwide issue-wide.
[03:36.840 --> 03:43.680]  The team we have now in 2223 is consisted of eight technical persons, so engineers mainly,
[03:43.680 --> 03:50.360]  two researchers, one trainee, and two close collaborators in companies, or independent.
[03:50.360 --> 03:55.760]  Very important is located at University Gustave Eiffel in Paris, near, in the greater Paris,
[03:55.760 --> 03:56.760]  so near Marne-la-Vallée.
[03:56.760 --> 04:01.560]  So you have Disneyland not far away.
[04:01.560 --> 04:10.360]  And this infrastructure is composed as a classic web-oriented application and infrastructure.
[04:10.360 --> 04:16.960]  So obviously we have the web interface that you can go to as a user, regular user, that
[04:16.960 --> 04:25.680]  encloses a lot of web services, so the main interface, but also an API that you can go
[04:25.680 --> 04:28.800]  and act upon from outside.
[04:28.800 --> 04:36.120]  So this allows us to open all our better-than-services to external applications, and specifically
[04:36.120 --> 04:39.840]  external projects that we are part on.
[04:39.840 --> 04:46.520]  We are built on an infrastructure of servers that are hosted at Marne-la-Vallée, so in
[04:46.520 --> 04:48.320]  our university.
[04:48.320 --> 04:49.960]  So everything is local.
[04:49.960 --> 04:57.800]  We don't have anything hosted online, which is 300 CPU, 3.5 terabytes of RAM, and 40 terabytes
[04:57.800 --> 05:03.360]  of storage, which is big and not big at the same time nowadays, but it's enough for our
[05:03.360 --> 05:05.440]  needs.
[05:05.440 --> 05:10.600]  We have two main services that we are providing.
[05:10.600 --> 05:19.480]  First is the storage of big database that we are collecting and curating, like for example
[05:19.480 --> 05:22.640]  the patents, the European patents.
[05:22.640 --> 05:33.200]  We have several other databases that we are putting inside some projects, so we can help
[05:33.200 --> 05:40.080]  specific projects with this data that we have stored in the infrastructure.
[05:40.080 --> 05:47.280]  And the second part is the processing that we offer, so the method, the scientist method
[05:47.280 --> 05:55.600]  that I will talk about later, that are run inside of projects on the platform directly
[05:55.600 --> 05:56.600]  by the researchers.
[05:56.600 --> 06:02.040]  So the researchers are actually actioning those methods directly from the web interface
[06:02.040 --> 06:08.480]  with the parameters and their data.
[06:08.480 --> 06:15.280]  And we of course have this monitoring of the whole infrastructure, so it stays online.
[06:15.280 --> 06:19.560]  So once you have an infrastructure, it has to do something for scientists, and that's
[06:19.560 --> 06:22.160]  the goal of the platform.
[06:22.160 --> 06:24.680]  So how does it do, and what does it do for science?
[06:24.680 --> 06:26.840]  First what?
[06:26.840 --> 06:30.800]  So this is the right part of the previous diagram.
[06:30.800 --> 06:35.240]  So we have a bunch of heterogeneous methods.
[06:35.240 --> 06:39.560]  So from science to metrics, the biometrics, the study of basically publications, scientific
[06:39.560 --> 06:47.080]  publications and tracks, natural language processing, so terms of traction, name-intensive
[06:47.080 --> 06:56.560]  recognition and stuff like that, social network analysis, so the study of all the interactions
[06:56.560 --> 07:03.840]  inside publications and texts in between actors, keywords, et cetera.
[07:03.840 --> 07:10.520]  Stochastic block models, then I will end the table somewhere in the room.
[07:10.520 --> 07:12.240]  We'll talk to you about later.
[07:12.240 --> 07:19.720]  Knowledge dynamics, so through time, the study of the knowledge through time, and the special
[07:19.720 --> 07:24.960]  analysis, which is a big part of our infrastructure now, because we have geocoding, geolocating
[07:24.960 --> 07:32.960]  inside publications, and geo-mapping framework.
[07:32.960 --> 07:41.920]  Okay, next we'll be, I will give, join you the mic to talk to you about how to cite
[07:41.920 --> 08:00.440]  cortex, because this is one of the big, I will try to note.
[08:00.440 --> 08:11.800]  Then as cortex is a research software being used mainly by researchers, then suppose the
[08:11.800 --> 08:22.400]  to be cited, besides the fact that there is no yet strong culture of citing software properly
[08:22.400 --> 08:26.840]  inside academic works, it's expected to be done.
[08:26.840 --> 08:37.120]  Then according to that, in 2022, we documented how cortex must be cited inside academic works,
[08:37.120 --> 08:43.280]  and here we have just an example, for example, if I am writing a paper, and if I cite cortex,
[08:43.280 --> 08:50.240]  this is how cortex will be handed at the end of the paper in the reference section, for
[08:50.240 --> 08:57.560]  example, and with this, we can give, for example, the credit to the developers, like, for example,
[08:57.560 --> 09:04.400]  Philip, Ale, and many others, who is contributing with this software for a long time, and must
[09:04.400 --> 09:11.400]  be recognized by this work academically speaking, then this is important, then that's what
[09:11.400 --> 09:23.000]  we did, and nowadays, we are lucky because a few years ago, we are missing the infrastructure
[09:23.000 --> 09:32.680]  to how to define, how to document, and how to cite software, and now we have, for example,
[09:32.680 --> 09:42.800]  the citation.cff, that it's a citation file format, that it's been adopted a lot, and
[09:42.800 --> 09:51.840]  I think it will be the standard way to do it, and on top of it, we have many tools.
[09:51.840 --> 09:58.640]  Here is one example, CFF convert tool, where we can convert the CFF file format, for example,
[09:58.640 --> 10:05.760]  to big tech, to bibliotech, or to APA format, and many other formats, then we can keep the
[10:05.760 --> 10:10.840]  automated data about the software important for citation in one single file, and from
[10:10.840 --> 10:22.040]  this file, we can derivate to many others, and it's really an easy way to do it.
[10:22.040 --> 10:31.520]  And then in Cortex, it's a solved problem, but we still have an issue about how to identify
[10:31.520 --> 10:35.640]  permanently this object.
[10:35.640 --> 10:41.280]  I say object as a digital object, okay, it's a software, but it's a digital object from
[10:41.280 --> 10:46.960]  the point of view of science, and how we can identify it permanently, for example, for papers,
[10:46.960 --> 10:53.400]  we have a DOI, that it's very well known, and so it's very well for papers, PDF, but
[10:53.400 --> 11:01.720]  for software, we have a problem to, we can need to cite software in many different levels
[11:01.720 --> 11:04.600]  of granularity, for example.
[11:04.600 --> 11:09.480]  We can cite the software, just the software name, or a specific version of the software,
[11:09.480 --> 11:15.000]  or a specific line of a specific file inside the source code of the software.
[11:15.000 --> 11:20.480]  Again, this is still an open question, because, for example, we have a proposition from the
[11:20.480 --> 11:30.440]  community to solve it, but there is no one single kind of permanent ID that covers all
[11:30.440 --> 11:33.560]  the levels of granularity, for example.
[11:33.560 --> 11:38.400]  We have the proposition from the Software Heritage Project, that it's the suite, the
[11:38.400 --> 11:45.120]  Software Heritage ID, that we can use it to cite software in the very specific code fragment
[11:45.120 --> 11:52.080]  to the snapshot of the full source code, but, for example, we can't use Software ID to cite
[11:52.080 --> 11:57.520]  software in a high level, like the software version of this, only this project name.
[11:57.520 --> 12:00.720]  For that, we should use another kind of ID, that's it.
[12:00.720 --> 12:07.560]  And it's an open question that we should work this year or next year, and that's it.
[12:07.560 --> 12:10.560]  Now it's the time for you, Juan.
[12:10.560 --> 12:11.560]  You?
[12:11.560 --> 12:14.560]  I don't know.
[12:14.560 --> 12:25.240]  So one question we've been asking ourselves is how to open the platform fully.
[12:25.240 --> 12:30.960]  What did we do to do that until now, and what to do in the future?
[12:30.960 --> 12:36.480]  So there is no platform, web platform, without free open source software today, and we are
[12:36.480 --> 12:43.600]  using a lot already, and I just put some of them on this page.
[12:43.600 --> 12:52.920]  You can see here, from any layer of the infrastructure, from the virtual machine, the containers,
[12:52.920 --> 13:02.000]  the scripts, the operating systems, et cetera, et cetera, until the actual methods and interface
[13:02.000 --> 13:04.680]  that you produce to the user.
[13:04.680 --> 13:09.120]  We are using free open source software, or at least open source software, because I don't
[13:09.120 --> 13:13.120]  want to start a debate here, but that's the general idea.
[13:13.120 --> 13:14.920]  Now this is not Cortex producing that.
[13:14.920 --> 13:22.480]  The Cortex is wrapping around this open source software, and it's open access for now.
[13:22.480 --> 13:27.960]  So it's free and open to everybody to use as an interface, but the code is not entirely
[13:27.960 --> 13:28.960]  free.
[13:28.960 --> 13:34.160]  We've been trying to do that a little bit more than the last few years, and hopefully
[13:34.160 --> 13:36.680]  in the next few years, even more.
[13:36.680 --> 13:42.680]  And Ale will talk to you about how to do that, and what are the challenges to actually produce
[13:42.680 --> 13:46.680]  a good open source software and methods.
[13:46.680 --> 13:52.520]  Hello, everybody, I'm Ale.
[13:52.520 --> 13:57.120]  I work with the guys and go in Cortex.
[13:57.120 --> 14:02.200]  So this is not really a talk about Cortex, I'm not going to do a demo or something.
[14:02.200 --> 14:07.800]  Because we thought of this talk as a moment to discuss what's going on with us, what is
[14:07.800 --> 14:12.880]  going on in the platform, because up to now, like Philippe just explained, everybody can
[14:12.880 --> 14:19.000]  come and use, but most of the software in Cortex has been developed without, well, with
[14:19.000 --> 14:23.800]  open source in mind, but not in practice, for practical reasons.
[14:23.800 --> 14:28.720]  So are we missing one slide?
[14:28.720 --> 14:33.680]  No, yeah, so everything from the methods, the interfaces, the back ends, everything
[14:33.680 --> 14:38.520]  has, so even our documentation is on Wordpress or Q&A also.
[14:38.520 --> 14:45.160]  So everything is based on open source, but up to now, we didn't have real open source
[14:45.160 --> 14:46.160]  practices.
[14:46.160 --> 14:57.000]  So how are we managing this transition from a project that is very aligned with open
[14:57.000 --> 15:02.200]  source community, but that doesn't manage to do open source, to start doing open source?
[15:02.200 --> 15:08.840]  So there was really a big part was just getting even more open source into the project.
[15:08.840 --> 15:17.120]  So we're starting to use project management tools that we were using, like some of them,
[15:17.120 --> 15:19.840]  but not as much, not as systematically.
[15:19.840 --> 15:25.000]  So this is a real common problem with research software, where researchers, they don't always
[15:25.000 --> 15:35.320]  have the reflex to go into GitLab or GitHub or whatever and use issues and use dashboards
[15:35.320 --> 15:41.040]  and project management, sometimes not even a thing in their head.
[15:41.040 --> 15:46.360]  And that's kind of the background that we were facing, like years and years of research
[15:46.360 --> 15:56.120]  and developing stuff, using open source software, but using version control because, well, it
[15:56.120 --> 16:01.400]  was still the infrastructure guys really needed that to work with, but it was really like
[16:01.400 --> 16:02.700]  the bare minimum.
[16:02.700 --> 16:06.240]  So how do you start to get people more involved?
[16:06.240 --> 16:14.400]  So we just, the first step was really pushing everybody to start using the usual tools,
[16:14.400 --> 16:21.040]  and sometimes first moving some projects and moving the discussion into issues on GitLab
[16:21.040 --> 16:22.160]  and all that.
[16:22.160 --> 16:29.040]  And then, okay, so, and then, and also adopting for our discussions more open source software,
[16:29.040 --> 16:36.480]  like Matrix, so it's like, I don't know if anybody knows Matrix here, maybe today everybody
[16:36.480 --> 16:40.920]  already knows Matrix, it wasn't really a given last year.
[16:40.920 --> 16:51.680]  And then, once we had like adopted the general workflow, the projects were still, most of
[16:51.680 --> 16:57.160]  the projects were still private, but they had the workflow of an open source project,
[16:57.160 --> 17:03.440]  so it helped get people into a better mindset and the people who would arrive and start working
[17:03.440 --> 17:12.040]  with us would like more easily recognize the pattern of working with open source.
[17:12.040 --> 17:19.280]  Then because we had this system that runs the jobs and everything, and usually in the
[17:19.280 --> 17:25.360]  old scripts, they would just use the structure that was in the system directly, and it was
[17:25.360 --> 17:27.960]  pretty hard to adapt and evolve them.
[17:27.960 --> 17:34.560]  So then we developed some kind of intermediate libraries so that it gets easier to, so making
[17:34.560 --> 17:40.880]  it easier to develop new methods in the platform.
[17:40.880 --> 17:44.960]  So we had to develop this intermediate, I won't have time to go into show the details,
[17:44.960 --> 17:51.800]  but just getting the idea that we have to get people into the workflow, develop libraries
[17:51.800 --> 17:59.800]  and interfaces and intermediate layers that help and automate some of the processes, give
[17:59.800 --> 18:06.040]  people more freedom to work, actually we also started integrating containers in the job
[18:06.040 --> 18:10.840]  manager and everything so that we could end with this intermediate layer so that people
[18:10.840 --> 18:21.840]  could have more freedom to make it easier for them to keep doing their non-standard,
[18:21.840 --> 18:27.720]  non-advisable, non-best practices stuff, but that works for them because their researchers
[18:27.720 --> 18:33.840]  are not coders, they're not always coders.
[18:33.840 --> 18:40.360]  And finally, and even for the people in the team that are more skilled programmers, of
[18:40.360 --> 18:44.920]  course, they want to have the possibility to do all this stuff.
[18:44.920 --> 18:50.360]  And finally, one thing is very important is at the management level, because this is
[18:50.360 --> 18:56.600]  like a kind of already intermediate kind of big-ish thing where we have hundreds of users
[18:56.600 --> 19:03.440]  everywhere, so we had to discuss this in the strategic planning so that our institutional
[19:03.440 --> 19:05.800]  partners could get it.
[19:05.800 --> 19:13.200]  So basically, just get it everywhere as soon as possible and then start, once you can show
[19:13.200 --> 19:22.520]  some of the benefits, also work at the institutional level.
[19:22.520 --> 19:27.600]  So these are some of the challenges.
[19:27.600 --> 19:32.400]  Most part of the team didn't have the know-how, part of the researchers we work with.
[19:32.400 --> 19:38.840]  People have limited resources, they have priorities, and I'm just reading out of there, but that's
[19:38.840 --> 19:39.840]  it.
[19:39.840 --> 19:41.340]  It's not very surprising.
[19:41.340 --> 19:45.840]  There's this learning curve of open source and everything.
[19:45.840 --> 19:50.320]  And there's always doubts about where is this going to lead us.
[19:50.320 --> 19:55.040]  There's licensing doubts, so there's part of, like, next thing I'm going to quickly show
[19:55.040 --> 20:00.160]  is that we're starting to open up stuff, and there's lots of stuff that hasn't been open
[20:00.160 --> 20:06.920]  yet simply because somewhere, we know that somewhere in some file, somebody used an extract
[20:06.920 --> 20:12.640]  from a software library that is not clearly documented, where it has been used, and we
[20:12.640 --> 20:19.360]  need to find it and check if the license is compatible.
[20:19.360 --> 20:22.240]  So we're always like, okay, where are we going?
[20:22.240 --> 20:29.840]  But getting things to as close as possible on every dimension is what's keeping us moving.
[20:29.840 --> 20:35.280]  It's keeping the feeling that something is moving, even if we're not seeing everything.
[20:35.280 --> 20:41.200]  But then we, well, through this effort, eventually, we managed to open source already a good part
[20:41.200 --> 20:52.200]  of specifically the science part, so the infrastructure, the dashboard, the interfaces, the front-end
[20:52.200 --> 20:58.160]  of the web application, and the job-managing part, or this part is not necessarily, it's
[20:58.160 --> 21:05.520]  taking longer because it's also another part of the infrastructure.
[21:05.520 --> 21:13.400]  But the main, most important, short-term part is to let the scientific part of it be available.
[21:13.400 --> 21:20.480]  So here's just some examples, I'm not going to, but yeah, so there's two parsers, an
[21:20.480 --> 21:29.840]  importer for a big database of scientific publications, a kind of generic parser that
[21:29.840 --> 21:37.960]  we use for, we're in a sociology of science lab, so even though the tools are more like
[21:37.960 --> 21:41.560]  general humanities and social sciences, we have a lot of interesting things that come
[21:41.560 --> 21:47.600]  from PubMed, data that comes from PubMed, and similar databases, so this is like also
[21:47.600 --> 21:54.680]  a parser for that, and here's one example of a project that's in progress because even
[21:54.680 --> 22:00.440]  though we reworked the project, parts of it still use some libraries that we're not sure,
[22:00.440 --> 22:08.600]  so we have to check, take the time to check if we need to change anything or replace.
[22:08.600 --> 22:14.960]  Just quickly show what some one of these things look like, well, this is just the repository
[22:14.960 --> 22:24.760]  for one of the projects, you have the source code, some documentation, it makes pretty
[22:24.760 --> 22:30.600]  graphs, and you have to use it like this, like this.
[22:30.600 --> 22:37.120]  So just getting things into what we all know about, what you know as an open source format,
[22:37.120 --> 22:53.520]  how do I go back to the presentation, F, no, no, there you go.
[22:53.520 --> 23:02.560]  What else there's to say, that's it, so thank you, I think, I mean, if you, if you, if you
[23:02.560 --> 23:10.480]  have any of you work in a similar institution where open source is not an evidence but it's
[23:10.480 --> 23:17.000]  something that you have to struggle with, we're very happy to exchange after, during
[23:17.000 --> 23:29.880]  or after the conference, did you skip the, no, no, you have no more time, thank you.
[23:29.880 --> 23:36.880]  Thank you.
[23:36.880 --> 23:39.880]  Any questions?
[23:39.880 --> 24:01.880]  Yeah, thank you a lot, I like your time because there is a lot of nice to choose, and you just
[24:01.880 --> 24:15.200]  wanted to share with us, and I'm very happy to hear that you are bringing up the gauging.
[24:15.200 --> 24:22.160]  Yeah, so the question was, what was preventing us to opening it from the beginning, and to
[24:22.160 --> 24:24.280]  make open source software from the beginning.
[24:24.280 --> 24:30.280]  So as Ale was saying, the first thing I think is just the lack of know-how, especially 15
[24:30.280 --> 24:36.720]  years ago when we started, the second thing I would say is that we, we basically didn't
[24:36.720 --> 24:44.480]  have a precise idea of what we were building, we just went basically script by script project
[24:44.480 --> 24:49.160]  by project, and this, this is how we built slowly the, the platform, then it became a
[24:49.160 --> 24:53.240]  platform, a proper platform, I mean, with the, with the resources and everything, and
[24:53.240 --> 24:58.480]  then, then it was already used, and, and we trained people on those methods already,
[24:58.480 --> 25:04.360]  so the problem of opening, of doing open source software is you have to think about it from
[25:04.360 --> 25:09.480]  the beginning, or at least refactor enough the, the code, so it's big, it's, it has no
[25:09.480 --> 25:14.880]  problem of, of license, and stuff like that, so this is, this is just a, you know, an ongoing
[25:14.880 --> 25:15.880]  streak.
[25:15.880 --> 25:32.880]  Yeah, speak loud, sorry, I can't, I can't hear you.
[25:32.880 --> 25:52.400]  Yeah, it's about, yeah, some, some people upload software to Zenodo or upload a documentation
[25:52.400 --> 25:58.760]  or, that's, that creates one kind of identifier, which you could also make a, some people just
[25:58.760 --> 26:03.040]  make a publication related to the software, so people can cite a DOI, Zenodo is going
[26:03.040 --> 26:05.960]  to give a DOI, so it's the same, pretty much the same thing.
[26:05.960 --> 26:10.480]  The problem is that, I think like Jenner was explaining is that sometimes you want to cite
[26:10.480 --> 26:15.680]  a specific version of the software or a specific file in a specific version, or you want to
[26:15.680 --> 26:21.360]  cite the software without mentioning a specific version, and, and you don't want to upload
[26:21.360 --> 26:34.680]  one version to Zenodo for every possible thing.
[26:34.680 --> 26:43.400]  Yeah, yeah, yeah, I, I don't know exactly deep what Zenodo offers, but I know that Zenodo
[26:43.400 --> 26:50.080]  offers a way to manage many different kind of digital objects, data, images, software,
[26:50.080 --> 26:53.000]  etc., and not only Zenodo.
[26:53.000 --> 26:58.400]  In our case, what is missing is to study all the options to see which fits better in, in,
[26:58.400 --> 27:06.520]  in our case, you know, is more to, to understand, because in, in fact, inside the, the, the software
[27:06.520 --> 27:12.040]  engineering community discussing this topic is still an open question, exactly how to
[27:12.040 --> 27:13.040]  do it.
[27:13.040 --> 27:17.880]  And when we, we search about this topic, it's quite difficult to find tutorials or someone
[27:17.880 --> 27:23.760]  teaching you how to, what kind of permanent indentifier you should use.
[27:23.760 --> 27:30.320]  For example, software heritage ID, for example, here is an example of the type of persistent
[27:30.320 --> 27:36.040]  indentifier that name it as intrinsic persistent indentifier.
[27:36.040 --> 27:42.560]  What means it doesn't, doesn't depend on, on, on a service registry where you need to
[27:42.560 --> 27:45.520]  ask a new ID to, to use it.
[27:45.520 --> 27:53.520]  It's like a hash, a hash ID of a commit is this kind of a intrinsic persistent indentifier
[27:53.520 --> 27:57.440]  that we can generate by the source code itself.
[27:57.440 --> 28:03.240]  Another option is to use maybe Zenodo or maybe any other thing to generate a DOI.
[28:03.240 --> 28:08.760]  Then for that, we are, we are going to ask to the registry, the DOI server to generate
[28:08.760 --> 28:10.440]  a DOI for us.
[28:10.440 --> 28:18.720]  And we have many options, we don't know yet deep and maybe Zenodo could be really interesting.
[28:18.720 --> 28:26.000]  I would be happy to, to learn with you, your experience because it's a lot of things to,
[28:26.000 --> 28:27.000]  to learn.
[28:27.000 --> 28:28.000]  Sorry.
[28:28.000 --> 28:41.200]  Okay, thanks for coming back to stop here.