---
tags: unconference-2022
title: Nordic RSE 2022 unconference
---

# Nordic RSE 2022 unconference

:::info
This is the public shared document. Everybody visiting this page can edit.

- unconference page: https://nordic-rse.org/events/2022-online-unconference/
- Link to this page: https://hackmd.io/@nordic-rse/unconference22
- Zoom room: https://uit.zoom.us/j/69645916718 (if you found this info without registering, please do [register](TODO) also, it takes 2 minutes)
- [Event page](https://nordic-rse.org/events/2022-online-unconference/)
- [Code of conduct](https://nordic-rse.org/events/2021-online-unconference/code-of-conduct/)
- You can create a separate HackMD for your session but please reference it from this document and please make it editable by everyone
- [How we do an unconference](https://hackmd.io/ue-yci-sSMKx458ChZab0w?view)
- If you have difficulties moving to a breakout room, please send message to Zoom chat ("can you please move me to room N?")
:::

## Day 1: 2022-10-18

(all times in CEST, [convert to your timezone](https://arewemeetingyet.com/Stockholm/2022-10-18/13:00))

- 13:00 : Welcome and Intro to the unconference format (HackMD, proposing sessions, scheduling)
- 13:15: Introduction to Nordic-RSE
- 13:30 : Invited talks, reminder that its recorded until 14.50 (chair Matteo Tomasini)
- 13:30 : How do skills learned during a PhD translate to industrial and professional work? (Richard Darst)
- 13:50 : BREAK
- 14:20 : Research software development in an open science landscape: on reform, co-creation and opportunities for professional establishment (Sanna Isabel Ulfsparre)
- 14:40 : Break
- 14:50 : Session 1
- 16:00 : Summary of day 1, social time starts
- 16:10 : Breakout rooms open again for continued discussion
- 18:00 : Close breakout rooms and Zoom call

## Day 2: 2021-06-30

(all times in CEST, [convert to your timezone](https://arewemeetingyet.com/Stockholm/2022-10-19/13:00))

- 13:00 : Introduction to the day and unconference scheduling
- 13:10 : Session 2
- 14:20 : Break
- 14:40 : Session 3
- 18:00 : Zoom closes

---

## Proposed sessions for day 1

:::info
Break for 10 minutes, please view and vote on these proposals. Breakout rooms will be announced after the break.
:::

### Digital humanities as a technical career path

### My proposed topic
fljdfhalgkha

### what do we want from nordic-rse
proposal: I heard this idea of "what do we want from nordic-rse", we have just introduced each other, what if we do that soon?

### Discussion/Workshop: Walk through the Nordic results of the 2022 RSE survey
- Need volunteer to chair / take notes
- in http://nordic-rse.org/international-survey-2022/

### Proposed for day 2

Please vote for a session if you would like to attend it. We will avoid overlapping sessions that have a lot of votes. So we won't use the votes to find out whether something happens or not but to avoid scheduling very popular contributions at the same time.

Vote like this:
- suggestion
Votes: oooooooo
- another suggestion:
Votes: ooo

Anyone may add a session ideas here. If possible, please indicate the duration (up to 70 min = full Session). Please describe it in a sentence or two like this:

### Day2 Session2 (13:10 CEST)
- AI and law (Enrico Glerean): is anyone here working with AI/ML? Is AI one of the first cases of research software being regulated by an European Act? https://artificialintelligenceact.eu/ Is it ethical to work on AI development if what you invent can be mis-used? Ideally someone who is actually working with AI can tell others, Enrico is happy to facilitate the conversation :)
- Vote: o
- Want to listen but am not an expert: ooo

- "10 simple things to do to translate your PhD skills to (software) industry": https://hackmd.io/@nordic-rse/10-things-for-industry-career [rkdarst chair?]
- Vote: oooo

- Software papers and their role in academia (Luca Ferranti) [hackmd link](https://hackmd.io/35ELVloGTm2weYbxVySAOA)
- Vote: oooooooo

- planning something around [Advent of Code](https://adventofcode.com/) (events, sessions, what languages we will use, how to share solutions)
- Vote: :christmas_tree: :christmas_tree: :christmas_tree: :christmas_tree:
- Should this maybe be left for the social chat? I feel that many people would like to be in this group... [matteo]
- Good idea! ooo

### Session 3
- Make model validation sexy again (Sunniva Indrehus): Code written to simulate real physical systems typically needs to provide an ensemble of input parameters to create a valid simulation. Writing logic for model validation is a repetitive and tedious task that needs to be handled with great care when performed manually. This talk will show how the modern software stack can simplify model validation of a traditional finite element method. With the combination of pydantic, docker, and Fast-API we can even make Fortran77 sexy again.

## How do a PhD's skills translate to industry?
Richard's list of translatable skills:
1) Creating new things
2) Find and interpret new information from primary and secundary sources
3) Technical writing (real papers with peer review, not just reports for educational purposes), both geared inside the field and outside the field
4) You can prove or disprove a task: you can take on new tasks / methods and know when they can yield a result or when to stop
5) Working independently
6) Learning to let go of impossible / too cumbersome tasks

Comments/questions:

- does open source activity make RSEs more attractive to industry?

- [Do we need data scientists?](https://2022.pycon.de/program/BKSMFA/) Data Scientist: funny talk I saw in the PyCon Berlin 22
- Funny point in this talk: managers thinks that data scientists do domething else then the data scientist !
- also how about blog post output and twitter/ social media visibility

- Question (when Richard mentioned "spending too much time on the same problem" -> lack of project management): would training in (software) project management be a skill that an RSE could learn in academia? where to start? [enrico]
- +1

- Question for the audience: Sometimes Academia becomes the only option due to language barriers. In Finland many (but not majority) of "data" companies hire foreigners, but at a certain level language skills are necessary for career advancement. How do you feel is the situation in Sweden and Norway or anywhere else you are from? Do we (immigrants) need to overachieve to show our values to industry compared to the average native applicant? [Enrico]
- In my institute (Norway) we are many internationals (40%ish). Managers can be foreigners, but I think it is true that it is hard if you do not talk Norwegian
- In Norway/UiT for higher management/leadership levels/roles it is a disadvantage to not be fluent
- How about outside academia? Do you have foreigners friends at sotware companies?
- Yes, I live in Oslo. Here there are usually a high percentage of non-norwegians in the software industry. (Oda, Cognite, Finn, etc.). In Norway I think it is very difficult to get a job if you do not have a network in Norway.
- Is is at least possible in Sweden. Anecdotal, but I have an American aquaintance working at a games company who haven't needed to learn Swedish in like 10 years since he found work where he could function fully in English. It took him some time to find work, though.

- Ability to "see" the big picture and communicate well (you wrote a thesis!)
- yes! problem solving
- The communication skill is really important in the industry. I have not seen that this is a common skill for "very technical" people +1
- great point! communicating complex things concisely

- Is there any list "10 simple things to do to translate your PhD skills to (software) industry"? If not we should write one :)
- We should make one! Can have some that you start early, some you do later on once you really start looking.
- Sounds like this would be a perfect workshop for tomorrow ;)
- We've actually begun talking about how to keep in mind that many Phd students don not continue in academia when we do doctoral courses. Maybe in general Phd education need to take this more into account.
- I remember seeing an estimate somewhere that less than 5% actually stay in academia, I will try to find the reference. I think many doctoral researchers do not know that when they start their career.
- I propse we start drafting here: https://hackmd.io/@nordic-rse/10-things-for-industry-career

- Does focusing on software engineering during PhD improve or limit later career opportunities outside of research?
- Good point: if one truly wants to become a professor, right now these "extra" activities might take time to what is considered priority in that type of career.
- Or if you do it early enough, does it make your research work more efficient in the medium term, so you can get to the higher-level stuff long-term?

- Postdocs in Companies (finland): https://www.podoco.fi/
- From that page: "PoDoCo is a matchmaking program supporting long term competitiveness and strategic renewal of companies and employment of young doctors in the private sector." ... so apparently others realize that PhDs to industry is important, too.

- Please share the link to the paper/blog that was mentioned in voice
- https://wissenschaftsrat.ch/images/stories/pdf/fr/SWR_2022_Can_we_do_better_Appropriate_size_and_org_of_ERI_system.pdf

- There was a great comment (I think by Sunniva?) about the "network" when looking for jobs outside academia. Could Nordic-rse promote some sort of network or at least increase visibility? I am unsure how... linkedin group to facilitate companies searching for individuals with software skills (and academic skills) in the nordics? (Enrico)
- what about job listings? like companies could post jobs in the nordic RSE website
- That could also be a way to get some extra money for the association. Let a company to a pub on the RSE-website for a little fee :)
- [Sunniva] Great idea to help each other with job postings. I think linkedin is really important if you want to contact the industry. Maybe a dedicated stream in The CodeRefinery Zulip could be interesting?

## Research software development in an open science landscape: on reform, co-creation and opportunities for professional establishment
Comments / questions below.

- seems to me that there is a lot of overlap between what Richard talked about earlier, about skills, and the kind of skills that will be useful for FAIR work. Should we spend some time to talk about how FAIR is useful to both academia and industry? [Matteo]

- how can we better motivate master and PhD students that following and learning about FAIR principles are good for them?

- it would be VERY important to take into account the "buzz word" effect for RSE!!! We should do as much as we can to avoid it...

- I think that until the FAIR principles (for data, software, etc) become part of the career advancement requirements for academics, only the ethical motivation of individual researchers is pushing them to be open, transparent, and FAIR. Do you know if in your organization (or at national level) such metrics for "openness" are being discussed? Of course with every metric, comes a way to manipulate it... (enrico)
- I'm not sure how it works in Finland, but in Sweden grants agencies put a lot of pressure (at least in words) about open source - e.g. by requesting data management plans, as Olivia said before. So, the pressure (also from EU) is there, and it will no doubt extend to software. The problem, as always, is how much academia wants to be proactive in such matters.

- [name=rkdarst] Not exactly on-topic, but "supporting open science" is one of the things I emphasize in trying to get support for RSEs here at Aalto. In fact, a large part of our projects somehow relate to the last "FAIRing it" phase. So yes, is very important!

- FAIR is expensive (in the broadest sense): would it make more sense to (start to) focus on some large projects and make them as FAIR as possible or everything should be FAIR and all the "expenses" are left to the researchers?
- But I think that FAIR RS is in itself valuable. If you write FAIR, you write it following best practices. It's expensive in the same way as best practices for good software are expensive.
- one thing that I [name=rkdarst] haven't seen enough is is "how far to FAIR does each (sub-)task need to go?" - we should put more effort to where it matters most, not everything can be perfect.

- top-down vs bottom-up open science/RSE? I [name=rkdarst] somehow feel that RSEs are good at bottom-up support of research (but our observation bias: we support computational research!), while I see open-science as more top-down in our departments. In reality, they are rather similar. Does anyone else feel this? (note: I realize this is a RSE conference, so we probably have a biased perspective, too!)
- I believe this might be due to how the RSE niche evolved in time. We kinda cut it ourselves (bottom-up) in places where there was a will to have us, while what Sanna Isabel suggested is that our position should be better integrated into the existing support infrastructures (e.g. libraries - this would be a top-down approach)

### EOSC Session content about FAIR research software
To give a taste of what kind of issues might be discussed on a larger open science scale:

Title: Integration of Research software into the EOSC infrastructure: Lessons learned from Computer science

Monday 14 November 2022
15:00 - 16:30
Room:
Main organiser: Umeå University, Sweden
Presenter information:
Mohammad Reza Saleh Sedghpour, Department of Computing Science, Umeå University
Sanna Isabel Ulfsparre, Department of Scholarly communications, Umeå University Library
EOSC future support: Pihla Kauranen, CSC ([e-mail])

Short Description: FAIR research software is essential to the quality assurance and reusability of research. As EOSC evolves, it is crucial to integrate infrastructures to share, collaborate, evaluate, reproduce, and preserve research software for use in the academic landscape and beyond. With the publication of the FAIR Principles for Research Software (FAIR4RS Principles), we expect that there will be an increase in demand for FAIR software, and related platforms, in all fields of research.

During this session, we will explore how computer science practices can inform further EOSC infrastructure development. Such practices can advise the development of similar methods and platforms in a broad range of other academic domains. In research fields with similar characteristics, they should also be directly transferable. Besides practical implementations, the session may also inspire policy development for open science research software practices.

The session will include examples of:
Closed crowdsourcing as a review process – ACM (Association for Computing Machinery) artifact evaluation and badge quality marking
Open crowdsourcing as a review process
Open crowdsourcing as citizen science and strategic advantages during software development.

Tweet: As EOSC evolves, it is crucial to integrate infrastructures to share, collaborate, evaluate, reproduce, and preserve research software for use in the academic landscape and beyond. This session will explore how computer science practices can inform the development of practices and policies for open science research software.

Relevance: FAIR research software is essential to research quality assurance and reusability. As EOSC evolves, it is crucial to integrate infrastructures to share, collaborate, evaluate, reproduce, and preserve research software for use in the academic landscape and beyond. During this session, we will explore how computer science practices can inform further EOSC infrastructure development.

## Unconference session 1 (18.10.2022)
### RSEngineering, FAIR principles and the future of the profession
Questions:
1) What would be a minimum viable skillset for a RSE professional?
2) What would you see in a professional profile for a RSE professional?
3) What would you like policy makers, strategists at HEI:s and companies to know about your profession?
4) What would be your role at an 1) researcher support center 2) RI 3) faculty/department 4) National competence center 5) company?
5) What learning paths would you like there to be for people who want to become an RSE professional?
6) What further training opportunities would be helpful in order to develop your profession?

- I'm interested to hear what the others in the meeting think - those who haven't commented lately.

- [name=rkdarst] I have trouble putting this into words, but how much of our problems are caused by research being too solitary - metrics don't value helping others enough? And thus everyone has to learn everything - which is too much.
- Metrics could help support collaborations as well, they are just not used like that now

- Do we define research as promoting new knowledge or getting citations for yourself? Most of us probably use the first definition, other academics use the second. (we are probably the happier ones!)

- Who here feels that research supporters / "research engineers" are important?
- [name=rkdarst] at least at Aalto, I have see a big transition and it seems easy to justify our value now. Getting the actual funding still takes work, of course, but it's more practical difficulties than lack of value these days.

- How much of the current "only care about publications" is because existing research management simply doesn't know how to do anything else - why would they promote the vauel of something they can't do? If so, more knowledge of RSE skills would help.

- I also thinks areanas that recognize software quality are imporntant. Like EU that tries with their TRLs.

- Which journals do you peer review software for?
- JOSS
- elife
- plos (various)
- Nature Scientific Data
- PNAS
- NeuroImage

- [name=sunniva] [Reproducebility checklist](https://www.cs.mcgill.ca/~jpineau/ReproducibilityChecklist.pdf) I think this is a nice guideline/starting point for forcing quality for the software
- +1

Check out this organisation and its system for software quality control and review https://reproducibility.acm.org/ they have a badge system etc.
- Cool! Thanks :)

- yeah, like discussing now: "what can *we* do now, other than talk?" 1-2 goals for the next year for the nordics?
- See section below "After Day 1 social"

- How did you learn to write good software?
- By myself, reading books/blogs and observing other non-academic open-source projects. No real academic teaching or examples.
- Collaboration
- by trying and failing
- .

### Discourse theory

- sorry for distraction but speaking of discourse theory, do you have a good resource I could look at to get an overview/starting point?
- I like Mouffe and LaClau discourse theory, but there are several strands.
- thanks!
- This is a good book about the basics if you read Swedish:
Diskursanalys som teori och metod
av Winther Jörgensen, Marianne, Phillips, Louise Förlag:Studentlitteratur AB
Format: Häftad
Språk: Svenska
Utgiven: 2000-05-26
ISBN: 9789144013022

## After day 1 social discssion

* Checklists (for other RSEs but also for researchers who wants to learn (everyone can be RSE))?
* Would it benefit to have a section "for institutions" on nordic-rse website, for the top-down approach of convincing other institutions in creating RSE positions (e.g. by showing current organizations with RSEs, national policy on open research methods e.g. the Finnish one, other EOSC recommendations, other examples not from the nordics)?
* "hackaton": publish a snapshot of your code to zenodo (or other repo with DOI) and make a Zenodo community for nordic RSE to list all contributions? The hackathon could be over a month e.g. #nordicRSEvember
* more blog posts
* 10 things to consider to transfer skills from phd to academia https://hackmd.io/RNl32Ka0ToO9HQ6RJvY9pQ?both

## Feedback after day 1

* one thing that went well

* one thing that could be improved in the future
- better communication with speakers and better role coordination but we know that now. the planning has been a bit chaotic. sorry for that.

## Questions day 2

- question to all: is the hackmd weird/jumpy? I saw a comment appear and disappear and worry that we need to do something about the document (in another event earlier today we had problems)
- works well: ooo
- is weird:
- someone erased my comment / it's buggy :)
- if it continues I can create a hedgedoc document in our sigma2 system or should we move text somewhere else to make it shorter?
- seems to work well now
- blame the user :)

### Lessons learnt from developing a multi-year RSE project (Ghislain Vaillant)
Comments below.

- really good point to design for readability (and testability?)

- Reference: The Programmer's brain, by Felienne Hermans, ISBN 9781617298677

- I like the technical dept control. It will ALWAYs come back to you if you ignore long term planning on something
- funny warrior cartoon. I feel this comes often, Finally, someone can work on these things that all of us know should ne improved
- yes it always feels like such a heroic effort to go into legacy code/ history

- All the suggestions for readability also have the secondary advantage that they help to minimize "enbugging" (that is, minimize mistakes when you write code). They should be followed always, even in non-collaborative projects!

- Note about Damian Conway: if you have the chance of seeing a talk by him, go. He's a great speaker and talks about super useful stuff.
- thanks for recommendation
- some resources: http://damian.conway.org/Resources/

- Reminds me what I said once: hardly ever the problems in software projects are technical, hardly ever your problem will be "I really cannot come up with the right algorithm / data structure / library", most times problems are social (misunderstanding of expectations/goals, technical debt for lack of documentation)
- yes! In my technical work my main obstacle and effort goes into scheduling a meeting, getting people on the same table, getting a common understanding of what has been done and what we want to do in future. The technical part seems like such a small part compared to the rest.

- link to results of the last RSE survey for those interested, with some insights into how RSEs split their work time: http://nordic-rse.org/international-survey-2022/

- [name=rkdarst] the analogy to physical labs is useful and we should consider it more. you wouldn't buy M€-worth equipment without people to keep it working. Same for lots of small equipment.

### Make model validation sexy ~~again~~ (Sunniva Indrehus)
Comments below:

- [name=mtomasini] Genuinely impressed that you already wrote tests and error messages during your PhD :D
- I only wrote errors :D
- Same, same
- I created puzzles
- [name=sunniva] it was not very impressive tests. More like checking that this input gave me this output each time. But the data model we cheked quite properly. This was to not waste computational time on a simulation that is not valid from a physical point of view. My PhD was also in computational physics.

- this is really nice! I have been using dataclasses in Python which can define types but this can do more since the available types are often not enough to restrict valid data
- I agree :D

- what if you need to call a library written without pydantic from a code using pydantic? can one send in the derived instances/objects and it still works? Or what is your experience with such situation? Does one need to "translate" before calling a legacy library?
- relies on using type hints (mypy)
- [name=sunniva] if you have used type hints in you class definitions (looks like it is a part of python since [3.5](https://docs.python.org/3/library/typing.html) ?) I think it would be pretty easy to translate. But you would have to create a pydantic class to do validation and get the error message for "free".

- Attrs: https://www.attrs.org/en/stable/

- json-schema: https://json-schema.org/
- https://pydantic-docs.helpmanual.io/usage/schema/

- [name=sunniva] Link to [demo](https://github.com/sunnivin/demo-make-model-validation-sexy-again) repo and [slides](https://github.com/sunnivin/make-model-validation-sexy-again/blob/main/model-validation-sexy-again.pdf), the documentation is excellent, but contains typically web-services examples. I find that seeing an example which is relevant in your field (like a turbine model) give you a lot of value.
- thanks!

### Software papers and their role in academia:

Discussion on: https://hackmd.io/35ELVloGTm2weYbxVySAOA

- Devil's advocate question: why not publishing a paper that uses the software and use that as citation / rank / whatever reference? (I have had supervisors telling me this when I wanted to write a software paper, (enrico))
- [name=matteo] That's something quite common in my field (evo biology). But I think the point is to have a software published without having to go through the hassle of a full on publication where you need to show novelty beyond your method.
- one problem is that people who will later join the software project will not be on that paper. but same problem if the software publication never gets updated.

- are jufo1 publications included in university stats at all? (or even, are they bad, since they lower the percentage of jufo2-3?) If they are in university stats somehow, it provides at least a small motivation for university-level support of RSE work...
- jufo 1? yes, definitely.
- Then jufo1 would definitely help me to get RSE funding
- I've also had experience of "ad personam" treatment for journals. A professor important enough and involved enough in the "upper floors" published in a Jufo 0 forum and gave special grant to publish there (and later got it changed to jufo 1)
- oh wait, so there is preferential treatment in academia? [surprised pikachu face] (sorry, I couldn't resist)

- That is a good point about reaching the right audience (specific discipline), maybe a software tool for a specific audience is better published in a niche journal as a paper with results and the software as a side product.
- What is important for us, I think, is that we have a "general" paper on which we can get software reviewed and published, even if the method finds a more suitable place in a niche journal. But I think this could be outside of the JUFO system. It should be something that has visibility among RSEs & co.
- +1 especially the lighter process for publication (no need to write a full paper)

- Google scholar of journals with word "software" in the name ranked by their h-index
https://scholar.google.fi/citations?hl=en&view_op=search_venues&vq=software&btnG=
(JOSS is the second)
- +1!

---
*always write at the bottom*