WEBVTT 00:00.000 --> 00:05.000 Thank you. 00:05.000 --> 00:13.000 I'm going to try to speak loudly, which hopefully will be OK, and hopefully not too bad for people online as well. 00:13.000 --> 00:17.000 So, yeah, so thanks for the opportunity to be here. 00:17.000 --> 00:22.000 In this endeavor, I'm going to talk about research software engineering, kind of in one stream, 00:22.000 --> 00:29.000 and then what's been going on at the University of Illinois or Banish campaign with regard to research software engineering kind of at the same time. 00:29.000 --> 00:34.000 So, I guess first, just how many people know research software engineering? 00:34.000 --> 00:38.000 OK, this is probably an unusual room here. 00:38.000 --> 00:41.000 That's hard to tell. 00:41.000 --> 00:46.000 So, wait, now I feel like maybe I should just skip the first six or seven slides. 00:46.000 --> 00:48.000 Well, we'll try and see what happens. 00:48.000 --> 00:51.000 So, maybe there will be something that you don't know. 00:51.000 --> 00:58.000 OK, so first, I'm going to start off with basically just defining research software, which I think probably everybody here has some idea. 00:58.000 --> 01:01.000 I don't know how many people have a formal definition. 01:01.000 --> 01:08.000 And so, in this group that more end-grimbeder led with a few of us over a couple of years ending in 2021, 01:08.000 --> 01:11.000 we tried to define this through a discussion. 01:11.000 --> 01:16.000 And the definition that we came up with is that research software has a bunch of different forms, 01:16.000 --> 01:21.000 but it's things that were created during the research process or for research purpose. 01:21.000 --> 01:27.000 And there are other pieces of software that are used in research, like operating systems and libraries 01:27.000 --> 01:33.000 and dependencies and package management systems and other things that are that weren't created for research, 01:33.000 --> 01:36.000 and they're not really intended for research. 01:36.000 --> 01:41.000 And so, we said that those should be called software in research and not research software. 01:41.000 --> 01:44.000 And this varies between disciplines. 01:44.000 --> 01:50.000 Sorry if you heard me say this yesterday, but if you are, I'll make a different example. 01:50.000 --> 01:57.000 If you are a computational linguist, and you have some code that you've written in Fortran, 01:57.000 --> 02:00.000 let's make this a really strange example. 02:00.000 --> 02:04.000 You need a compiler in order to actually run that code. 02:04.000 --> 02:09.000 And so, the compiler that you're using is not research software for you as a linguist. 02:09.000 --> 02:14.000 But, on the other hand, if you're a computer scientist who is studying Fortran compilers, 02:14.000 --> 02:16.000 maybe there's such a thing still. 02:17.000 --> 02:20.000 That same compiler, if you would be research software. 02:20.000 --> 02:23.000 So, it's a context question. 02:23.000 --> 02:30.000 We can argue about this definition for a while, and in this process, we did argue about it for about 18 months. 02:30.000 --> 02:34.000 And we came up with this as the best thing that we could agree on. 02:34.000 --> 02:36.000 It's not necessarily the right thing. 02:36.000 --> 02:41.000 It's not necessarily what any one person thought was the right thing, but it's what we got to. 02:41.000 --> 02:45.000 And lately when I've been saying this in talks, people want to have this discussion again. 02:45.000 --> 02:50.000 They say, can we have a community group that goes through and comes up with a better definition? 02:50.000 --> 02:53.000 And my answer is, sure, let's try it and see. 02:53.000 --> 02:56.000 I don't think we'll come up with a better definition, but if anybody wants to try it, 02:56.000 --> 03:00.000 I think it's worth doing, and maybe we will. 03:00.000 --> 03:01.000 Okay. 03:01.000 --> 03:06.000 So, research software then typically comes from research. 03:07.000 --> 03:10.000 This has happened from the start of computing, more or less. 03:10.000 --> 03:17.000 Software started appearing around 1948 as best I know, somebody can correct me if they have a better definition of better term. 03:17.000 --> 03:25.000 Research software started appearing in the early 1950s, just a few years later, and it was initially weather software as far as I can tell. 03:25.000 --> 03:35.000 And then another roughly 20 years more or less until we got to software engineering and starting to think about how do we do this in a reasonably professional way as a discipline. 03:35.000 --> 03:47.000 The people that are writing research software who are researchers, sometimes faculty, sometimes other people generally don't know good software practices still. 03:47.000 --> 03:55.000 And we have software engineers who really do know good research, sorry, good software practices, but they don't understand the context of research. 03:55.000 --> 04:04.000 And so you get situations where software engineer comes in and they say, please write down the requirements and then I will build software that meets those requirements. 04:04.000 --> 04:12.000 You say, I don't know the requirements, this is research, and so there's that kind of question about how do we go back and forth between those things. 04:12.000 --> 04:23.000 At universities and other places where research happens, you often do have students and postdocs who initially don't know good software practices, but they kind of pick them up over time during their career. 04:23.000 --> 04:30.000 And they get to a point where they're actually fairly decent at this sometimes, and then they leave and they go somewhere else. 04:30.000 --> 04:36.000 Some of them stay around and they become the staff at an institution and that's great if that happens. 04:36.000 --> 04:43.000 Sometimes they're called a postdoc for a very long time and they're not officially ever called a staff member, but that's a different problem. 04:43.000 --> 04:50.000 And and eventually we do get to a point where we have staff who have an understanding of research as well as software engineering skills. 04:50.000 --> 04:52.000 And I think that's where we are today. 04:52.000 --> 05:03.000 And so one one way of looking at this is the difference between the the crafts person and the scholar where we can say that a scholar is a researcher who's really driven to understand things. 05:03.000 --> 05:10.000 They find intellectually demanding problems, they work on them until they really understand them and then they pass on that understanding through teaching. 05:10.000 --> 05:18.000 And then we have crafts people who are driven to create and and leave behind an artifact that really. 05:18.000 --> 05:23.000 That that really says what they understand that that is the understanding in some sense. 05:23.000 --> 05:27.000 This is people that feel pain when they make something that's fragile or ugly. 05:27.000 --> 05:33.000 They prefer to make things that explain themselves although they certainly often will understand the need for documentation as well. 05:33.000 --> 05:38.000 And this is work that requires patience and pride in doing a job well. 05:38.000 --> 05:44.000 And so scientific software research software needs people that can understand and do both of these things. 05:44.000 --> 05:49.000 And I just want to point out this is from blog post from 2012. 05:49.000 --> 05:56.000 So this is not a new idea at all. This is something that's been around for almost 15 years at this point. 05:56.000 --> 05:57.000 Sorry. 05:57.000 --> 06:02.000 So if we then think of a research software engineer and an effort to try to define this a little bit. 06:02.000 --> 06:08.000 We have a space which I'm going to define with two axes because we're in a research context so I can do this. 06:08.000 --> 06:11.000 We have a software engineering axis and we have a research axis. 06:11.000 --> 06:15.000 And so we have people that are software engineers that are really great at software engineering. 06:15.000 --> 06:18.000 We have people who are researchers that are great at research. 06:18.000 --> 06:26.000 We have people who are researcher developers who are kind of somewhere in the middle that are quite good at research but also are decent at software. 06:26.000 --> 06:33.000 And the idea is that we have research software engineers who are really more in the middle and can do both of these things. 06:33.000 --> 06:39.640 And there's not a point to there's kind of a cloud of which we have research software engineers somewhere better at research somewhere better at software engineering 06:39.800 --> 06:42.840 But they're pretty good at both of these things and that's kind of the key here 06:43.720 --> 06:51.440 So in 2012 in the UK I believe in Oxford, but I'm not 100% sure if I'm right about that 06:52.800 --> 06:54.800 There was a collaborations workshop and 06:55.980 --> 07:02.160 Breakout group at this workshop was talking about this problem and what they realized was that there were a lot of people that were doing this work 07:02.160 --> 07:07.200 That were doing research software with an understanding of research and trying to apply good software engineering practices 07:08.000 --> 07:15.600 But they didn't have a common title and in the UK they had 194 different job titles that they found in job listings that were basically doing the same work 07:16.560 --> 07:21.000 So they said let's come up with a common title. Let's call this research software engineer RSE 07:22.360 --> 07:28.520 They decided that there wasn't any community of these people all these people with different jobs and different places different titles had 07:29.480 --> 07:34.880 felt like they were alone often and so they created an association that's turned into a society and there's now 07:34.880 --> 07:41.640 Others societies and associations and other places and they realized that this wasn't a profession and so they started working with their 07:41.640 --> 07:48.520 Universities to develop career paths and a and a formal structure by which people could be hired and could get promotions and and 07:48.680 --> 07:51.640 Could be evaluated based on their work in this in this job 07:52.640 --> 07:55.600 So 13 in plus years later 07:55.600 --> 07:58.920 We have again this this movement this term that was born in the UK 08:00.040 --> 08:04.400 The UK RSE Association formed with about 50 people in 2013 08:04.520 --> 08:08.400 It's now a professional society a non-profit do's paying 08:09.360 --> 08:15.240 Organization with 700 official members and almost 7,000 unofficial members 08:16.240 --> 08:21.080 And there's other associations in a lot of other countries at this point as well 08:21.600 --> 08:25.960 Belgium Germany Netherlands Nordic countries Australian New Zealand Canada Chile 08:28.440 --> 08:32.960 In the US we have a US RSE organization that has about 3,800 members now 08:33.480 --> 08:37.480 So we're kind of growing a little bit behind the UK, but catching up hopefully 08:38.240 --> 08:44.800 There's an RSE like group in France that's called devlog that actually started before the RSE movement and it has kind of 08:44.800 --> 08:49.120 Its own history and does things all a little bit differently, but there's a lot of overlaps as well 08:49.840 --> 08:56.440 And there's new associations that are forming or at least trying to form in in Africa and in Asia 08:56.840 --> 09:02.480 But primarily I would say in South Africa and India at this point, but trying to expand beyond those as well 09:03.720 --> 09:07.720 And these associations then can work on local issues collectively 09:07.920 --> 09:11.920 So so in the US working on issues of despisic to the US 09:11.920 --> 09:19.000 But then also we can coordinate when there's an effort to try to work together across countries to try to change culture to 09:19.000 --> 09:21.000 Change practices at a higher level 09:21.000 --> 09:24.000 Okay, so that's the RSE side of things 09:24.840 --> 09:30.720 If I go then on to the side at the National Center for Super Computing Applications at the University of Illinois 09:31.280 --> 09:33.280 What we've been doing 09:34.400 --> 09:40.880 NCSA started in 1986 so 40 years ago actually like a week ago, or so it was our 40th anniversary 09:40.960 --> 09:50.240 Professor Edil and I submitted a proposal without being requested to without submitting a two particular call to create a US 09:50.440 --> 09:56.160 High Performance Computing HPC capability the next year, sorry in 1985 and 09:57.240 --> 10:08.080 After a lot of discussion from the funding agency they decided to seem like a good idea and it's something they should do and they supported five national centers that were formed in NCSA. It was one of those five 10:08.080 --> 10:14.840 We ended up then I would say catalyzing the internet and the web under this support 10:16.120 --> 10:22.040 In the first year we created NCSA Telnet which was in some ways intended to help people connect to these super computers 10:22.600 --> 10:28.560 We created HTTBD to serve websites so that people could then share the information that they were creating 10:29.160 --> 10:31.160 We created Mosaic 10:31.160 --> 10:36.720 a few years later, which was the first graphical web browser that I would say in a lot of ways kind of started started a lot of the 10:36.960 --> 10:38.960 The web is 10:38.960 --> 10:44.760 Well initially as a scientific activity, and then eventually into more of a commercial activity 10:46.080 --> 10:48.640 And then we got to a point where there were some problems 10:49.160 --> 10:53.960 So the core funding that we were getting from the funding agency ramped down over a few years 10:54.080 --> 10:58.600 So we had this core funding that supported everybody for ten years and then five years later 10:58.600 --> 11:05.720 We didn't have any of this core funding and so that led to a lot of challenges in terms of what do people do and how flexible are they and how do they work? 11:06.720 --> 11:14.520 All of the funding became competitive so in order to get funding we had to write proposals and compete with people in every other university every other lab 11:15.880 --> 11:21.000 Some large projects one some of these competitions as well as some small projects 11:21.880 --> 11:33.440 But then some people left because they were unhappy with this competitive landscape and they were looking for other things and and we also wrote some proposals that didn't exceed and so then there wasn't money for everybody and some other people left 11:34.400 --> 11:39.880 And we ended up at a situation where we had multiple groups in software that had overlapping skills 11:40.400 --> 11:44.840 We had internal competition were different groups were creating projects that competed with each other 11:44.840 --> 11:47.080 So we were developing three different workflow systems 11:48.440 --> 11:54.040 We had internal competition on grants about who would be allowed to propose to particular things if there was a limit 11:55.520 --> 12:01.360 Internal competition for collaborators right a good computational scientists in some department would be approached by multiple groups 12:01.360 --> 12:07.240 The NCSA that wanted to collaborate with them and they didn't want to do that. They just wanted to collaborate with one person and do their research 12:08.440 --> 12:13.020 Internal competition on staff when we wanted to hire somebody who would hire them and how that work 12:13.920 --> 12:20.820 Huge discrepancies on titles and salaries because different groups were running under different systems and so that led to unhappy employees and 12:20.820 --> 12:26.260 In equity and people that couldn't necessarily move from one group to another one project to another easily and 12:27.180 --> 12:35.580 New clear no clear entry point for new collaborations as I said and finally no clear path for advancement except if you were a PI and you want to grant 12:35.580 --> 12:41.940 That was a good thing and that would help you advance. Oh, and no job security as well because if you didn't win then that was it 12:43.140 --> 12:45.140 Okay, so that was bad 12:45.140 --> 12:47.140 But then things got a little bit better 12:47.540 --> 12:49.140 Well, more than a little bit better 12:49.460 --> 12:52.500 So one of the NCSA groups at that time was called 12:52.820 --> 13:01.020 Image and spatial data analysis or I SDA and this was kind of an isolated group that was working in this particular area of image and data processing 13:02.660 --> 13:07.220 And a person named Kenton McKennery who is one of the co-authors of this talk took over that group and 13:08.100 --> 13:13.900 And he was frustrated with the situation and he wrote a manifesto that basically said how things should work in his opinion 13:13.900 --> 13:17.980 And that included things that said basically the PI role as sacred 13:18.220 --> 13:22.180 All right the PI's are the people that should be making decisions not some other level of management 13:23.220 --> 13:25.220 There should be an emphasis on job security 13:26.500 --> 13:33.420 Governance and decision making should be as democratic as possible among the PI's at one level and among the staff at another level 13:34.420 --> 13:36.260 There should be an agreed upon hierarchy 13:36.260 --> 13:41.020 They should be agreement between the PI's about when to hire and when to expand and who to hire 13:41.220 --> 13:46.900 It shouldn't be each person doing it just for their project and there should be agreement on which projects to pursue 13:47.820 --> 13:52.740 And that trust was the key for efficiency and that if the PI is could trust each other and could work together 13:52.900 --> 13:57.180 Then they could actually collectively support all of the people and grow the organization 13:58.900 --> 14:01.500 And so some new principles then came out of this and 14:02.260 --> 14:07.180 The principles that we ended up with were to try to align projects and diversify funding 14:07.380 --> 14:12.180 So the idea is that if you have a big project you use it to start small projects that do some of the pieces 14:12.180 --> 14:16.220 And if you have small projects you can bring them together to turn them into bigger projects 14:16.220 --> 14:21.580 And the bigger projects get bigger funding that's more sustainable the small projects have shorter term funding 14:21.580 --> 14:25.180 But there ways of exploring ideas. So thinking of these all together as a system 14:25.940 --> 14:28.380 I'm trying to reuse this software as much as possible 14:29.780 --> 14:36.140 Trying to actually be a group as opposed to a set of projects. So the PI and staff have a larger voice 14:36.140 --> 14:38.140 A louder voice together 14:38.140 --> 14:42.900 To have a management committee that beats in addition to team meetings to bring everybody together again 14:43.620 --> 14:48.620 To look for opportunities for synergies and ways of leveraging the work that's happening in one place and another place 14:49.060 --> 14:54.020 And to provide stability for staff. So if one project ends, there are other projects that can hire the staff 14:54.020 --> 14:58.260 Because they were involved in bringing them in initially and they know who they are and they know what they do 14:59.780 --> 15:03.060 To think of ourselves as collaborators and not contractors 15:03.060 --> 15:06.980 And I think this is actually a big deal that happens for a lot of research software engineers 15:06.980 --> 15:09.140 But it was really I think a big change for us 15:09.940 --> 15:12.660 So if we're working with somebody that's a researcher 15:13.220 --> 15:19.700 We want to be seen as a as a copi. We want to be involved in the process of developing the ideas of doing the work 15:20.580 --> 15:27.380 To be co-authors on the paper that comes out things like that not just to be somebody that you go to and say can you write some code for me that does this 15:29.460 --> 15:34.580 And then finally and this is a little bit repetitive but a little different way of looking at it is 15:34.740 --> 15:40.980 To think of staff members as working for the institution not for the project and they're hired by the institution 15:40.980 --> 15:42.980 They're not hired by the project 15:42.980 --> 15:49.860 That there's one broad shared reusable continuous job ad that is reused by any project that needs to hire people 15:49.860 --> 15:53.300 We don't have to have each project right their own job ad and have their own competitions 15:54.020 --> 15:56.020 That there's a rotating hiring committee 15:56.020 --> 15:58.020 that helps 15:58.020 --> 16:04.500 Spread the load across people that we have a formal career path and we use senior staff to do training in a formal way 16:04.580 --> 16:05.860 as well 16:05.860 --> 16:07.140 so 16:07.140 --> 16:09.380 Vision and goals for NCS and software 16:10.100 --> 16:14.900 So one of them is that we support scientific discovery through software and this is really research discovery 16:14.900 --> 16:19.860 Sorry, we always have issues with using science versus research, but it's it's really scholarly research 16:21.140 --> 16:25.940 We amplify the efforts we support again. We're not just writing code. We're actually making the project better 16:27.060 --> 16:31.620 We're known for our staff and our software and we're a leader in the research software engineering movement 16:32.180 --> 16:36.900 We strive to lead in the areas that we undertake and that we embody trust and responsibility 16:37.140 --> 16:43.940 So again, these are principles that we try to work with with different researchers. We don't always do all of these as well as we should but this is what we try to do 16:45.780 --> 16:49.780 So if we look then back at the the RSE movement and how we're fitting into it 16:50.420 --> 16:55.940 So I joined NCS a in 2016 about 10 years ago and I was active in the RSE movement 16:56.500 --> 17:01.460 Through the UK at that point and so so I brought this idea and I said you know 17:01.460 --> 17:05.140 I think this is what we really are right. We are RSEs. We're not just 17:07.220 --> 17:10.100 Research developers. I think was the title that was being used before that 17:11.220 --> 17:17.540 And so this gave us a sense that we at NCS a weren't alone. There were other people that had the same problems and we could learn from them and work with them 17:19.460 --> 17:24.900 In 2018 I co-founded the US RSE organization and I was on the steering committee for the first four years 17:25.460 --> 17:31.140 And then Kenton McKenry from again from NCS a was on for the next three years and now we have a pinging shen 17:31.620 --> 17:36.420 Who is also on this so we're trying to make sure that we have a continual role in shaping the organization 17:37.220 --> 17:39.220 At least as the members vote for us 17:40.260 --> 17:46.660 And we see ourselves as a US exemplar of an RSE group and so I just want to point out one one paper that we wrote in 17:48.420 --> 17:54.420 2019 that was looking at the model that we had the model that was taking place at the University of Notre Dame and 17:54.500 --> 18:01.140 The model that was taking place at the University of Manchester in the UK and so we compared these three different models for how do these different groups work? 18:01.140 --> 18:08.500 How do they hire people? How do they find projects? How do they support themselves and and we think that this is I don't know three out of maybe 18:09.220 --> 18:15.140 I don't know five eight different models that maybe makes sense maybe there's more but but we think people that want to start a new group 18:15.140 --> 18:17.460 Can look at these models and see what other people have done 18:18.900 --> 18:22.020 We also have again a formal career path. I know you can't read this 18:22.420 --> 18:27.380 Sorry, let me just jump back for this second and say the slides are at this DOI 18:27.380 --> 18:32.740 They're also linked from the fussed-end page about this talk so you can you can look at any of these at any point 18:33.380 --> 18:36.180 But we do have a right a formal structure where we are 18:37.060 --> 18:38.100 Things like 18:38.100 --> 18:44.340 Responsibilities that change over five levels from associate research programmer all the way up to principal research programmer 18:44.420 --> 18:50.580 Decision making that comes in a different point super vision that comes in a different points 18:51.220 --> 18:57.700 Kind of requirements about education and training and experience and then kind of preferred experience and so we've tried to again formalize this 18:57.700 --> 18:59.700 So that when people want a promotion 18:59.940 --> 19:06.900 We know what they need to be doing at the next level and we can be we can see are they able to do that and they can see what do they need to learn as well 19:08.020 --> 19:11.860 And and this gives people also of course the the chance to move up over time 19:12.820 --> 19:16.020 So I'm getting close to the end just to mention 19:16.740 --> 19:20.260 We kind of have three different formal project roles right now 19:20.500 --> 19:25.700 We have the the PI or copi or the Stuart the person that's responsible for this excessive of a project 19:26.100 --> 19:28.100 Their job is 19:28.100 --> 19:35.940 Primarily oversight and guidance and organization and and helping staffing and and being the person that's responsible for the deliverables 19:36.500 --> 19:39.860 This is generally some of that's working like about 8% of the time 19:40.020 --> 19:43.220 So this is kind of like one month a year spread out over a year 19:44.580 --> 19:47.220 And these people can work on multiple projects at once 19:47.620 --> 19:52.020 And we have technical leads who are the person that's really responsible for overseeing development 19:52.420 --> 19:54.420 Coordinating the developers 19:54.420 --> 19:59.860 Probably doing some of the development themselves but also training the developers the newer people the younger people that are coming in 20:00.900 --> 20:06.420 And this is often about a 10% position and again they can be doing multiple of these at the same time 20:06.820 --> 20:11.380 And then we have the core developers who are actually doing the the real main development work 20:13.380 --> 20:17.140 Perhaps they're involved in publications and reports as well as they get somewhat more senior 20:18.420 --> 20:22.020 Ideally, this is 50% time they're working on two of these at a time 20:23.140 --> 20:31.460 In practice we don't always get 50% of people from a project and so sometimes they can be working on maybe up to four for more senior things 20:31.460 --> 20:35.060 So maybe four projects at a time we try to keep junior people more at two 20:35.060 --> 20:38.260 But again, as people go up maybe it's a little more. I'm I will say 20:39.140 --> 20:43.860 50% time my deal the reason we say that is that if we get full-time support for somebody 20:44.180 --> 20:46.020 We hire two people at half-time 20:46.020 --> 20:48.020 All right, we don't hire one person 20:48.020 --> 20:53.700 And the reason for that is so that there's redundancy so that there's backup so if somebody leaves somebody else can come in 20:53.780 --> 20:58.580 But also so two people can talk to each other about the project and it's not somebody just working on this in isolation 20:59.220 --> 21:01.700 So that's a kind of a conscious choice that we've made over time 21:02.660 --> 21:10.180 Okay, so next the last slide is just an archite of where we are so there is a few people that are kind of up at the top leading 21:10.180 --> 21:13.220 We've got a bunch of different groups that are focused in different areas and 21:14.740 --> 21:21.060 Yeah, and I think things are are generally going well when we are able to build different skills in different areas that help with 21:21.700 --> 21:23.460 What our collaborators are interested in? 21:24.820 --> 21:26.820 So to wrap up 21:26.820 --> 21:29.380 We've had a bunch of successes. I think at this point 21:30.020 --> 21:38.500 This software director is about a fifth of the organization in terms of both staff and funding and I think that's probably a reasonably healthy healthy size 21:39.780 --> 21:45.860 We haven't had anybody that had to leave because of lack of funding in at least 10 years and I'm not sure if anything happened before that 21:45.860 --> 21:49.540 Because that's the time that I've been there. I mean obviously people have left for a variety of reasons 21:49.540 --> 21:53.860 But but nobody's had to leave because we didn't have the funding for them and and I think that's 21:54.580 --> 21:57.620 quite important in in telling people that this really is an 21:57.780 --> 22:02.100 An institution that they're joining and not a not a three-year project. They're joining for example 22:03.780 --> 22:10.580 We have I think quite a happy staff and in a good culture and people see opportunities for personal and technical growth 22:10.980 --> 22:16.420 And I think we're making an impact on research at Illinois in the country and on the RSE movement 22:17.620 --> 22:19.620 We do have a couple of challenges though 22:19.860 --> 22:21.860 So one of them is funding 22:22.020 --> 22:28.500 The current external environment is a quite a large challenge. I don't know if you've heard but there's some stuff going on in the U.S 22:28.500 --> 22:30.500 And it's made this a little bit difficult 22:31.860 --> 22:36.980 So we're not exactly sure for a variety of reasons how much funding we're gonna be getting in from different government sources 22:37.620 --> 22:42.740 and philanthropic sources and industry as well and so there's lots of different effects that are leading to challenges 22:43.460 --> 22:48.660 We also have almost no institutional support from the university as a whole in terms of funding 22:49.140 --> 22:54.660 And there's other RSE groups that are very heavily supported by their institutions. All right, so we're supported by projects 22:54.660 --> 22:59.140 We just have a lot of projects and we can kind of build up critical mass enough that this works 22:59.860 --> 23:04.660 But if it stops working then we don't have the fallback of the institution. So we'll see what happens with that 23:05.380 --> 23:12.100 And then the other thing is that hiring and AI are both making our existing processes difficult 23:12.580 --> 23:17.460 So hiring is becoming more challenging recently and AI is 23:19.460 --> 23:24.500 Making everything more difficult. I guess is a short way of saying it as we've heard in previous talks 23:25.140 --> 23:27.140 So just as an example 23:27.140 --> 23:31.860 We used to have a pre-screening for hiring or we'd ask people to do a programming exercise 23:32.740 --> 23:36.980 Which doesn't make any sense anymore because they just ask an AI agent what the answer is 23:37.620 --> 23:43.220 And so we don't know how to evaluate people in a good way until we actually know them until we meet them and talk with them 23:44.980 --> 23:45.540 And 23:45.540 --> 23:48.180 AI as a whole also leads to questions about 23:48.820 --> 23:50.820 With our collaborators 23:50.820 --> 23:55.860 If they think that they can just use AI to write software themselves and they're going to write good software that's sustainable 23:55.860 --> 24:01.460 Then do they need us and and how do we convince them that they actually still do and that there's something that we're doing that's different 24:01.700 --> 24:05.140 So so we do have some challenges, but that's where all these things and 24:05.460 --> 24:07.460 Happy to answer any questions. Thank you 24:07.780 --> 24:09.460 Thank you 24:14.980 --> 24:16.980 Couple minutes 24:17.780 --> 24:19.780 Yeah, go ahead 24:37.460 --> 24:43.060 If there's a sort of adjunct of the computer size department, that's kind of worked or a research as a cheaper 24:44.260 --> 24:46.260 But they also 24:54.740 --> 25:03.620 Yeah, so the so the question is kind of basically where where do RSEs fit within the organization and there's kind of pros and cons and different 25:04.180 --> 25:08.180 aspects and so I think there's a few different models as you mentioned 25:09.300 --> 25:12.180 In some cases they report to the the IT organization 25:13.060 --> 25:18.980 In some cases they report to the the the research organization when there is a central research organization 25:18.980 --> 25:24.420 Which is the case at Illinois at some cases they're more related to computer science as you said 25:26.100 --> 25:32.100 It seems that there is not an answer for this at all and it really depends almost entirely on the institution and 25:32.420 --> 25:35.620 And even when it happens when something happens in an institution 25:35.620 --> 25:42.820 It's not necessarily the thing that you might say from the outside is best, but for some reason the people in the institution have thought it's the way that it's going to work 25:43.140 --> 25:44.660 So it's 25:44.660 --> 25:50.180 There is a lot of discussion about this and it I think it's another thing where there's a small number of models and they have pros and cons 25:50.500 --> 25:55.300 But it's but it almost more depends locally on what makes sense rather than these kind of global questions 25:57.940 --> 25:59.940 One more 25:59.940 --> 26:01.940 question. Yeah, good 26:03.060 --> 26:10.820 All down. Sorry. Yeah, sorry. Sorry. It's all you first. Yeah, yeah, but hiring would record open source 26:11.220 --> 26:13.220 Commission away 26:13.220 --> 26:15.220 And follow up 26:15.220 --> 26:17.700 Yeah, because open source was mentioned when they took 26:18.260 --> 26:24.260 Yeah, yeah, a property of your project. Yeah, yeah, yeah, yeah, okay, so 26:24.660 --> 26:26.660 Let's see so 26:26.660 --> 26:32.180 Is open source work a something that kind of can be brought in in the hiring process and can be 26:33.380 --> 26:36.500 Evidence of somebody's qualifications and the answer is completely yes and 26:37.060 --> 26:41.780 Ideally that happens and we're very happy when that happens, but sometimes we have people that are coming from industry 26:41.780 --> 26:44.980 And they can't actually share any code or or something else. So it's 26:45.540 --> 26:53.780 We like to do that when we can but it isn't always an answer and then and then a comment that's completely true that I didn't really mention open source 26:53.780 --> 26:55.780 Which is kind of a strange thing here 26:56.740 --> 27:04.180 But but I will say that from from my point of view, I think there's there's research software kind of as one space and there's open source is another space and they 27:04.820 --> 27:06.820 They overlap a lot, but not completely 27:07.700 --> 27:10.820 And and I think we would fall into that in which 27:11.860 --> 27:14.820 Probably 90% I would say of what we do is open source 27:15.700 --> 27:21.460 Not everything there there's some research software that we work on that's not but most of it is and and that's what we try to do and 27:23.380 --> 27:25.780 And once somebody is add in CSA that is something that we 27:26.740 --> 27:28.740 That we see as 27:30.100 --> 27:34.660 Something to strive for is is working openly and and when we have projects that are open 27:34.660 --> 27:41.780 We start we try to start them as open from the very beginning and so we we try to follow the best open source practices we can 27:42.900 --> 27:49.860 With the recognition again that some projects that's not the right answer for even though almost all the time it is there's there's a couple where it's not 27:52.420 --> 27:54.420 Did you want to 27:55.780 --> 27:59.060 I assume that in your wallet the RSC doesn't have control over 27:59.540 --> 28:08.020 The architecture they have any and what about do so yeah, so question about the RSEs and control over their own research 28:09.060 --> 28:15.460 So the this is actually a tricky question because we have another track the besides the RSE track that's called a research scientist 28:15.860 --> 28:18.820 And those are people that do have complete control over their own research 28:19.700 --> 28:24.820 But have a much different model and they get evaluated based on papers more often rather than software 28:24.900 --> 28:28.820 Although it can be both where the RSEs are getting evaluated almost entirely on software 28:31.380 --> 28:39.620 We don't have a great answer for people that are in the middle other than as people get to be more senior and they can bring in their own projects or create their own projects 28:40.020 --> 28:45.460 Then that big they have the flexibility to do what they want with those projects. So in some ways 28:46.180 --> 28:53.540 If you can bring in funding to support something then you can do that as long as it doesn't seem like it's tremendously out of line with everything else 28:54.100 --> 28:56.100 As long as it lines up with the institutions 28:56.740 --> 28:58.100 Objectives 28:58.100 --> 29:02.340 So if somebody wants to do some research that's related to software and they can get funding for it 29:02.340 --> 29:03.940 They will have time to do that 29:03.940 --> 29:10.820 But we're it's it's hard to start that up because we don't really have kind of seed funding or a way to to do that 29:10.820 --> 29:16.260 So it almost has to to come out of a collaboration with somebody else or it has to be something that's such a great idea 29:16.740 --> 29:18.980 That you can write a proposal without having done much work 29:20.020 --> 29:22.020 And it happens 29:22.580 --> 29:28.980 But it typically is people that are more senior that have had some experiences that they can build on and and turn that then into a proposal for an idea 29:31.540 --> 29:33.860 Okay one one more. I'm not sure where we are in time 29:35.140 --> 29:37.140 Okay, so I go ahead. Sorry 29:52.020 --> 29:54.020 From this 30:18.660 --> 30:20.660 Question is basically kind of how 30:22.180 --> 30:25.460 There's an appearance that we're not interacting with the larger open source community 30:26.820 --> 30:30.180 And they open let's say open data community. I'm not sure if that's exactly fair 30:30.180 --> 30:32.180 But let's let me try that one for the minute 30:32.900 --> 30:36.820 So we I think we do interact with them, but we don't necessarily 30:37.380 --> 30:42.020 Talk about it very much. We kind of do it more as individuals as opposed to as an organization 30:42.500 --> 30:47.140 So a lot of the people that are working on these projects are using other open source projects as part of them 30:48.100 --> 30:54.740 And they're contributing back to those often as individual contributors and and that the time that they're using to do that 30:55.140 --> 30:57.540 is time that the project effectively is paying them for 30:58.100 --> 31:02.340 It's not kind of volunteer effort that they're doing on the side. So so I feel like we do 31:02.820 --> 31:09.700 Participate in a number of projects in some cases some of the things that we do we try to turn them into larger or more general open source projects 31:11.140 --> 31:16.820 We're often not as successful with that as we might like and and that could be from lack of 31:17.140 --> 31:19.140 experience 31:19.140 --> 31:22.420 It could be because things that we think are general are still somewhat specialized 31:24.260 --> 31:31.380 I don't have a good answer, but we I would say that we do try to do this, but whether or not we're as successful as we could be probably as a question 31:31.940 --> 31:33.940 Yeah, no, I 31:33.940 --> 31:39.940 I mean I I appreciate the the comment and again I think as individuals we do this well as an organization 31:39.940 --> 31:43.700 We probably don't prioritize it very much. Sure. We don't talk about it at least as enough 31:43.700 --> 31:46.660 Yeah, okay. Thank you