diff --git a/paper.tex b/paper.tex index bded3f0..251113f 100644 --- a/paper.tex +++ b/paper.tex @@ -65,15 +65,15 @@ \section{Introduction} Research software has been written and used for decades in an increasing range of disciplines. It has been established that most research requires research software for its results~\autocite{Hannay2009, Hettrick2015}. To solve pressing research challenges, better software is crucial~\autocite{Goble2014}. -During the past decade, research software gained ever-growing attention and is becoming accepted as a research result on its own. +During the past decade, research software gained ever-growing attention and is becoming accepted as a research result in its own right. -The number of people developing software in academia is constantly rising~\autocite{Hannay2009, Hettrick2015}. -Research Software Engineering encompasses creating, adapting or maintaining Research Software. +The number of people developing software in academia is continuously rising~\autocite{Hannay2009, Hettrick2015}. +Research software engineering encompasses creating, adapting or maintaining research software. It also includes consultation and training with the aim to enable researchers to some of these tasks themselves~\autocite{goth_foundational_competencies_2024}. These actions are very diverse and so are the environments they are performed in. This position paper focuses on groups of research software engineers that provide their services for an entire research organisation or at least a substantial part of it. -We advocate the establishment and support of dedicated, central RSE groups in German research organisations, with clearly defined tasks, contact points, and, in particular, sustained funding, for the benefit of all researchers in their organisation. +We advocate the establishment and support of dedicated, central RSE groups in German research organisations, with clearly defined tasks, points of contact, and, in particular, sustained funding, for the benefit of all researchers in their organisation. Similar initiatives have been formed in other national contexts (see~\autocite{barkerkatz2024,martin2023} and references therein). We provide an overview of the various tasks these teams have and discuss potential realisation strategies, learning from already existing RSE units. @@ -115,7 +115,7 @@ \section{Motivation for central RSE units} \subsection{Pooling: a necessary ingredient} The main focus of this paper lies on central RSE teams since the benefits of RSEs to research in general are described elsewhere, \eg{} in~\autocite{Anzt2021}. The main advantages of central RSE units all stem from the pooling of resources. -There are at least three aspects to RSE pooling that research institutions can benefit from: funding, diverse knowledge, and support contacts. +There are at least three aspects to RSE pooling that research institutions can benefit from: funding, diverse knowledge, and a central point of contact. First, pooling of \textbf{funding} allows organisations to invest in building up institutional knowledge by supporting RSEs to become experts. A central RSE team on long-term contracts will act as a knowledge hub due to their accumulated experience in and support of several disciplines as well as established contacts within the organisation. This is comparable to commercial/industry R\&D departments or so-called inhouse consulting~\autocite{Grima_2011}, where key software architects and developers establish a knowledge hub that can be consulted by project teams as necessary. @@ -138,7 +138,7 @@ \subsection{Pooling: a necessary ingredient} Furthermore, by centralising RSE expertise, the unit benefits from economies of scale, making it feasible to handle also less frequently needed tasks across multiple groups. This synergy broadens the scope of possible software-based solutions while distributing costs and resources more efficiently. -The third aspect to pooling RSEs is visible most of all from a users perspective: a \textbf{single, central contact point} for digital challenges is valuable to researchers, whose first problem often is not knowing whom to contact, partially because while they know what they want, they might not know what they need. +The third aspect to pooling RSEs is visible most of all from a users perspective: a \textbf{single, central point of contact} for digital challenges is valuable to researchers, whose first problem often is not knowing whom to contact, partially because while they know what they want, they might not know what they need. A central RSE team can, due to its proximity to research, much better listen to the wishes expressed by researchers and then help formulate needs and act as a channel to either fulfil them themselves or reformulate and redirect the request. The results are increased research speed and quality and, with that, a higher reputation of the entire research organisation. @@ -219,13 +219,14 @@ \section{Vision}% For our purposes, the term “institutions” is used broadly to include universities, other colleges, associations like Max-Planck, Helmholtz, Fraunhofer, or Leibniz, as well as other research-performing organisations. They show a wide variety in organisational structure as well as internal scientific diversity. Thus, there can be no single optimal blueprint for such an RSE unit for all research institutions in Germany. -We instead describe modular components that can be mixed and matched based on the respective local environment. +We instead describe modular components that can be combined for the respective local environment. We have identified nine modules that can make up an RSE unit. -In practice most RSE units will only support some of these modules. -Different RSE units will focus on different sets of modules. -Thus, it is likely that no two RSE units will be, or should be, alike. -However, these nine modules together with assumed weights are part of a simple model of an RSE group which provides both a quick overview of an individual group as well as a way to compare groups. +In practice, most RSE units will only support some of these modules, +and different RSE units will focus on different sets of modules. +Thus, it is likely that no two RSE units will be, or need to be, alike. +However, these nine modules, along with their assigned weights (indicating relative importance), form a simple model of an RSE group. +This provides both a quick overview of an individual group as well as a way to compare groups. The nine modules are described below. \subsection{Module 1: Foster a local Network of RSEs}% @@ -235,7 +236,7 @@ \subsection{Module 1: Foster a local Network of RSEs}% Research software is an integral part of modern research. There are, therefore, researchers in every institution that at least sometimes perform the tasks of an RSE\@. We call these RSEs embedded RSEs. -These RSEs typically work isolated from similar RSEs in different groups, within the same institution. +These RSEs typically work isolated from similar RSEs in other groups within the same institution. The central RSE unit together with the embedded RSEs form a hub and spokes network. Connecting embedded RSEs has multiple positive effects, both for them as well as for the organisation: @@ -247,13 +248,13 @@ \subsection{Module 1: Foster a local Network of RSEs}% Information also flows from the embedded RSEs to the central RSE unit allowing it to discover common requirements and opportunities to centralise some services. How an RSE unit realises this task will depend heavily on its environment and resources. -We only mention a few examples here to provide inspiration, with the explicit claim of incompleteness: +We only mention a few examples here for inspiration, with the explicit claim of incompleteness: talks, seminars, workshops, hackathons, as well as informal meet-ups all facilitate establishing a local network of RSEs. As a foundation, a central RSE unit employs experienced RSEs, mostly at the post-doctoral level, who are not only expert software engineers and scientists, but also good communicators with the ability to work in interdisciplinary teams. At least a core of a central RSE unit's employees need to have permanent contracts to be able to offer that deep expertise that requires years of experience. Moreover, an onboarding process can serve as an entry point for new RSEs, whether in the central RSE unit or as an embedded RSE, into an institution's network. This gives an opportunity to gauge how the new colleague can benefit from the RSE unit's teaching services and whom they might want to network with based on their planned work. -Similarly an off-boarding process can help to make sure that all acquired knowledge that is relevant to the institution is passed on to someone who stays, even if possibly not within the specific research group. +Similarly an off-boarding process can help to make sure that all acquired knowledge that is relevant to the institution is passed on to someone else, even if possibly not within the specific research group. \subsection{Module 2: Consultation Services}% \label{sec:consultation} @@ -318,7 +319,7 @@ \subsection{Module 3: Development Services}% Without a central RSE unit to anchor such specialised expertise, it would be difficult to establish such a service. Also, having this expertise centralised allows for synergies in the deployment procedure for mobile apps: The RSE unit can create institutional accounts with the app stores and manage the time-consuming deployment process including hard-to-setup procedures like code signing. -Besides the technical benefits of this central deployment procedure, the institution will also benefit from the increased visibility and potentially be able to build a brand with its technological output. +Besides the technical benefits of this central deployment procedure, the institution benefits from the increased visibility and the ability to build a brand with its technological output. \end{mdframed} \subsection{Module 4: Teaching Services}% @@ -364,7 +365,7 @@ \subsection{Module 5: Create a Network of Institutional Partners}% Thus, a close collaboration between RSE and RDM groups helps everyone: both RSE and RDM groups by being able to offer a more comprehensive service than when working alone, as well as the researcher, who benefits from receiving this single coordinated service, instead of dealing with two independent entities. The question whether RSE and RDM should be located in two separate groups or should be combined in one common group is intentionally left open, as the answer depends on local, pre-existing circumstances. -Some research institutions might host a dedicated \textbf{HPC group} which may or may not be part of the central IT unit. +Some research institutions host a dedicated \textbf{HPC group} which may or may not be part of the central IT unit. HPC is an RSE-related field, so HPC groups might already provide training, consulting and funding opportunities in this area. At the same time, HPC by nature focuses on highly efficient, many-core, if possible parallel computations. The challenges of an average researchers often start a long way before reaching that level, and they might not ever need to consider HPC\@. @@ -406,12 +407,12 @@ \subsection{Module 8: Software Maintenance Service}% For research software, dedicated archiving solutions such as Software Heritage~\autocite{DiCosmo2020,DiCosmo2023} or Zenodo's GitHub integration~\autocite{GitHubZenodo} exist. In contrast to research data, however, the long-term availability and usability of research software requires more than an adequate archiving method: Software maintenance is an ongoing change process of software after its release. -It includes both fixing bugs that are discovered in the software and adapting the software to changes in the execution environment such as hardware, operating system, toolchain and software dependencies. +It includes both fixing bugs that are discovered in the software and adapting the software to changes in the execution environment, such as hardware, operating system, toolchain, and software dependencies. In the scientific community there is a demand for long-term maintenance of research software, but academic labour turnover and missing funding schemes make research software maintenance often rely on the (potentially unpaid) efforts of individuals. -An RSE hub with long-term core staff can partially solve this problem by taking over maintenance tasks. -In order for this to be feasible two criteria need to be met: +An RSE hub with long-term core staff can partially address this problem by taking over maintenance tasks. +In order for this to be feasible, two criteria need to be met: \begin{itemize} \item The software needs to be developed according to software engineering best practices with a strong emphasis on testing and continuous integration. \item The RSE hub needs to be involved during the development period either through development or consultation services in order to ensure that best practices are followed and the required knowledge is transferred to the hub. @@ -454,7 +455,7 @@ \section{Existing Implementations}% In order to understand the service portfolio of these existing RSE units, we conducted a small survey that received a total of twelve responses from Germany, the UK and the US\@. We asked RSE units for the composition of their service portfolio --- the results are shown in Figure~\ref{fig:survey}. -From the gathered data and the additional free text information of the participants we conclude that the service components that we have identified in Section~\ref{sec:vision} are indeed relevant for existing RSE units. +From the gathered data and the additional free text information of the participants, we conclude that the service components that we have identified in Section~\ref{sec:vision} are indeed relevant for existing RSE units. Additionally, we see a large diversity in the weighting of these components, which is to be expected given the different environments of the RSE units. Within this diverse data set, we identified two rather different archetypes of RSE units: those that offer development services and those that do not. The RSE units offering development services would typically invest a lot of their resources into this component, whereas others put a much larger emphasis on teaching and consultation services. @@ -473,17 +474,17 @@ \section{Realisation Strategy}% Following that, we describe a potential transition pathway, starting from existing structures that have grown in research alliances such as, \eg{} DFG-funded Collaborative Research Centres or also in research departments of an institution. This is complemented by discussions of the possibility of outsourcing RSE services and of the challenging task of identifying and hiring suitable RSE candidates. -\subsection{Funding Possibilities}% +\subsection{Funding Sources}% \label{sec:funding} We see four basic options for financing RSE positions, which we will briefly explain below: \begin{enumerate} \item ordinary budget positions, -\item the overhead of externally funded projects, +\item overhead of externally funded projects, \item explicitly requested person-months in externally funded projects, and \item dedicated RSE calls. \end{enumerate} -While each option stands for itself, in reality, an institutional RSE unit will most certainly finance its staff by an appropriate mixture of possibly all four options. +While each option stands for itself, in reality, an institutional RSE unit will likely need to finance its staff by an appropriate mixture of possibly all four options. The mixture at a particular institution depends heavily on the local conditions. At research institutions, it is important to resolve the conflict between time-limited research funding and the need for permanent positions, the latter being required to remain competitive with industry when hiring highly-qualified RSEs. @@ -497,7 +498,7 @@ \subsection{Funding Possibilities}% \item Funding organisations additionally allocate a portion of the direct project funds as overhead, which is typically divided between the institution and the applicant. We propose using a small percentage of the overhead agglomerated at the institution to permanently finance central RSE positions. Assuming a third-party funding income of €50 million annually and a 20\% overhead, €100,000 in permanent funding requirements for one person-year would only account for 1\% of this overhead. -\item In project applications involving the development of research software, corresponding person-months should be applied for to finance RSE tasks. +\item In project applications involving the development of research software, corresponding person-months should be requested to finance RSE tasks. In this way, an applicant can book a fixed number of working hours from the RSE pool and pay for the costs accordingly. This model has been successfully implemented at several UK universities. In order to scale, it needs to be supported by an institutional policy. @@ -515,7 +516,7 @@ \subsection{Funding Possibilities}% \subsection{Transition Pathway} -We present a possible transition pathway from RSEs distributed over an institution and associated purely with research working groups and corresponding projects towards an institutional RSE unit. +We propose a potential transition path, beginning with purely embedded RSEs---distributed across the institution and linked to individual research groups and their projects---and leading toward a centralised institutional RSE unit. After starting with initial measures not necessarily requiring dedicated funding, we discuss developing a business plan and how the RSE unit can be established. We conclude with measures for promoting its growth. @@ -526,7 +527,7 @@ \subsubsection{Initial Measures} \paragraph{Network of RSEs} Forming a network of RSEs localised at an institution can be initiated by any existing RSE individual or group that is preferably already in contact with other RSEs at the institution. -An institutional dedicated mailing list, chat group and possibly other communication platforms can be created and a request for participation can be circulated via institutional channels such as an employee newsletter. +An institutional dedicated mailing list, chat group, and possibly other communication platforms can be created and a request for participation can be circulated via institutional channels such as an employee newsletter. First common events such as social gatherings or RSE-related seminar talks can be organised and announced via the communication platform. If the initiative is geographically local, this process can be accompanied, facilitated and strengthened by founding a local de-RSE chapter\footnote{A list of existing chapters: \url{https://de-rse.org/chapter/}.}. Such network-building has been successfully initiated and implemented at several German research institutions such as the German Aerospace Center and the Forchungszentrum Jülich. @@ -542,7 +543,7 @@ \subsubsection{Developing a Business Plan} Decision processes at institutions usually require a business plan on which to base the decision on whether to establish an RSE unit. The business plan should include an outline of the RSE unit, its responsibilities and offerings as a subset of the nine modules presented in \autoref{sec:vision}, the funding, and the resulting benefits for the institution and its researchers. -A rather difficult and crucial question can be the positioning of the RSE unit within the organisational structure of the institution. +A crucial and sometimes difficult question is how to position the RSE unit within the institution’s organisational structure. A canonical place would be a new subunit of an existing unit close to software, training services and computing such as the local or central IT unit or the library. Since most institutions already have an RDM unit, it seems natural to add the RSE unit as a parallel structure. @@ -568,11 +569,12 @@ \subsubsection{Establishing the Unit} The first position is an RSE coordinator who is the contact person for all embedded RSEs. Among other things, they are also responsible for organising meetings, developing training programs and reporting to stakeholders. The second position is a central RSE, responsible for providing selected services and infrastructure. -These central positions complement the network of RSEs as described in~\autoref{sec:network}. +These core positions complement the network of RSEs as described in~\autoref{sec:network}. -Drawing from the business plan and considering the actual initial staff situation, a first task of the centrally-funded structure is to define a basic service portfolio according to the modules described in \autoref{sec:vision}. -In addition to the already mentioned networking and teaching, see~\autoref{sec:network} and \autoref{sec:teaching}, it seems natural to start with consultation, see~\autoref{sec:consultation}, +Drawing from the business plan and considering the initial staff situation, a first task of the centrally-funded structure is to define a basic service portfolio according to the modules described in \autoref{sec:vision}. +In addition to the aforementioned networking and teaching activities, see~\autoref{sec:network} and \autoref{sec:teaching}, it seems natural to start with consultation, see~\autoref{sec:consultation}, as this allows to evaluate the potential necessities for other services such as development, infrastructure provisioning, and maintenance. + An extension of the initial service portfolio for a larger target audience requires the acquisition of funding for further positions, see below. The best RSE unit can easily become useless if other departments as well as RSEs outside the central hub do not know about it. @@ -634,11 +636,11 @@ \subsection{Staff Acquisition/People} These RSEs will bring a diverse set of skills centred around the topics of research, digital tools, and team-based work and hence can easily offer the consulting services mentioned in the previous section and guide people to their implementation in their workgroups. To fill gaps, the RSE unit can also maintain a roster of freelance workers. -In order to retain RSEs it should be possible for them to become experts in a field and hence this should make this job more attractive to budding RSEs, +In order to retain RSEs, it should be possible for them to become experts in a field and hence this should make this job more attractive to budding RSEs, in order to mitigate the problem that some will only see this job as a one-year stint after their PhD and then move on to something else. To facilitate the retention of skilled people, industry has long identified education as an effective tool. For RSEs, this should be helped by yet to be formed academic facilities that enable them to keep on learning skills after their first professional qualification, supported by the respective certification programs. -In the longer run, Research Software Engineering should be integrated into the existing study programmes. +In the longer run, research software engineering should be integrated into the existing study programmes. One option here is the creation of an RSE master as a specialisation for computer science or application-domain bachelor programmes. This should be complemented by adding a minor in these application-domain study programs such as biology, music, engineering etc.\ to facilitate the communication between the corresponding two groups of RSEs. There are already some master's programs available (\eg{} in Berlin, Munich, and Stuttgart), that develop this specialisation on top of a domain bachelor.