TRE Satellite Event notes - User experience in TREs#

Prompts#

  • How do we design TREs to be as seamless to use as possible?

  • Who are we designing TREs for?

  • Are there design principles we can generalise and open source?

Notes#

  • How do we design TREs to be as seamless to use as possible?

    • Or - what are the barriers to using TREs?

    • Unfamiliar - how do we get the tools we’re used to

      • and expectation: I’ve set up my own machine how I want it, and now I have to change everything?

    • Over-complication? If you just need Excel can you find it/use it without other things getting in the way?

    • Extra governance/ethics/etc - researchers have already gone through a lot of approvals; do TREs require extra work?

    • Education about approval extent - do researchers assume that once they’ve got a TRE they can do anything they want?

    • Users don’t know what they can do with a TRE - how to communicate that and engage new users

    • Different users want different things - from just Excel to CLI HPC systems

    • Even between TREs there are implementation differences; how do we manage the expectations of existing users as well as new?

    • Perhaps a catalogue - what different TREs are available, what datasets are available, how do you apply for/access/work with them?

    • Is there mileage in standardising TRE user experience? If you need to register for multiple TREs will you need to learn all of them?

    • Modularisation of TREs - one infrastructure but multiple options for UI, tooling, data access etc.?

      • Research in other areas suggests that giving users choice could improve productivity & happiness.

  • Who are we designing TREs for?

    • Is this just for researchers? What kind of researchers? What do they want?

      • If we’re building for NHS staff/clinicians they’re used to (old!) Windows machines

      • Will our TRE users be comfortable with the UX we provide?

    • Can we educate our users (researchers) in new ways to approach tasks?

  • Are there design principles we can generalise and open source?

Actions/next steps#

  • Explore modular components for building researcher-specified TREs - workspaces, environments, data access, tooling etc.

  • Training for both CLI and GUI tools is critical to encouraging users to feel confortable in any TRE.

  • Make sure that datasets (data catalogue), TRE options etc. are all properly published, advertised, quickly accessible and well-described.

  • central resource for TREs - (trustedresearch.ac.uk?) - publicity, manage a “trusted packages” repository of tools, catalogue all TREs, data catalogues etc?