Advanced Research Computing


Paid-for Project Prioritisation Criteria

ARC's Research Technology Professionals will always aim to take on paid-for projects that lie within our remit. However due to high demand we are unable to work with all groups that contact us. The following criteria are used to prioritise which projects we will take on. The short version is that we will prioritise projects where we can provide the most benefit, relative to other potential solutions. What is below expands on this principle and gives more explanation of what it entails.

Selected projects will likely result in:

  1. Mutually beneficial research collaborations - we prefer to work with researchers to push the boundaries of what is possible, rather than delivering products to a specification.
  2. Development of skills within the research group
    1. Transfer of software engineering knowledge to active computationally-focused PhD students and postdocs.
    2. Part of our mission is to promote research technology best practices across UCL. We therefore aim to partner with research groups by mentoring and training members of the group to be able to take on ownership of the project after our involvement. Groups that are committed to engaging in this process will be preferred.
    3. All else being equal, we will also prioritise groups with a low existing degree of computational expertise, where our involvement and training can have most impact.
  3. Use of the software or data analysis tools developed beyond the originating group, e.g. development of software components, tools, insight or methods which could be of benefit to other research projects. This may also include work to improve the usability of software by others, or improve installation procedures.
  4. Sustainable outcomes following the project, e.g. software that can be maintained and/or further developed by others. For instance projects where our involvement improves the structure and testing of existing software to facilitate modifications.
  5. ARC helping a wide spectrum of groups across UCL - in the interests of fairness, and in support of the aims above, we may prioritise groups we have not worked with before.

Normally we would expect projects to require at least six weeks of full-time-equivalent effort, to keep the administrative overhead small compared to the work done. We will take on shorter projects where there is a strategic benefit to doing so, for instance to develop a prototype where there are reasonable grounds to expect follow-on work, or to build relationships with new research groups.

We will also prioritise projects based on alignment with the focus of the department, namely the application of computational and data approaches to research and discovery, and our ability to deliver. The criteria below thus expand on what falls within our remit; where these are each at least partly met, the criteria above have more weight. These are written from the perspective of our Research Software Engineering profession; similar principles apply for our other professional groupings.

Projects are more likely to be selected which require:

  • Specialist developers with a background in research, with:
    • An ability to understand advanced scientific or mathematical topics in order properly to understand the project requirements
    • Sophisticated understanding of algorithms, methods, or tools typical of research software, e.g.: Parallel programming; Numerical methods and libraries; Research information management expertise; Semantic web; Information visualization; Statistical analysis
  • Specialist developers with an understanding of professional software engineering, with:
    • A commitment to engineering best practice, capable of managing complex interacting requirements to produce sustainable, clear, correct code
    • The freedom to focus on programming without distraction by publication pressure
  • Demonstrably correct, auditable code
  • Skill-sets not available elsewhere. For instance, projects are not likely to be selected which require:

    • Non-specialist programming or system administration skills - a standard IT contractor may be a better option, or support from other teams within UCL. We may be able to suggest suitable options - contact us at arc@ucl.ac.uk.
    • Open-ended research not involving the application of advanced computational and data-intensive research methods, such as purely developing new algorithms - these may be better suited to research collaborations with groups in computer science or numerical analysis, for instance. We may be able to advise on suitable contacts.
    • Websites based on a content management system (CMS) such as Drupal – talk to UCL Information Services Division for these.

Information about projects we have worked on may also be helpful guidance.