Workshop for e-Infrastructure trainers - break out group conclusions

Hartree.jpgTaking up most of the day at the 2013 Workshop for e-Infrastructure trainers were the break out discussions, which produced a wide range of interesting observations and conclusions. 

Here follow the results of the two main groups and the questions they considered.

Group One (11.00-12.00)

What do we need to create a community of e-Infrastructure trainers?

  1. Within e-infrastructure there are well established silos, which we will call vertical communities. There is need for a cross cutting e-infrastructure community, which we will call a horizontal community. This horizontal community will be a platform for sharing basic information and training.
  2. To establish this is horizontal community, we need a bottom-up approach. Where individuals organise cross-disciplinary information sharing. This will enable the vertical communities to understand that the applications they use are not unique and realise the effectiveness of this method, and initiate a top down approach.
  3. There is a lost skill set of basic programming skills that has been overlooked, in favour of advanced training.
  4. There is not enough core materials available on HPC, i.e. textbooks. Training should convey conceptual ideas and not only recipes.
  5. Need community support for the people who provide IT services for research.
  6. Identify a new community that is service orientated when supporting research.

Improving the understanding and status of training in research

  1. What is the problem - getting the right people on the right course
  2. Incentives to train and setting the expectations of the level/content of the course
  3. Supply and demand - feedback to scaling out - physically and online
  4. Reuse and collaboration (licensing)
  5. Funding for training to ensure the return on investment

Improving the understanding and status of training in research

  1. Disconnect between skills needed by PGs and those that UGs have.
  2. Problems with recognition (hours, CV, "career progression") for trainers within institutions, research communities and funding bodies.
  3. Senior academics / supervisors always prefer domain specific training to computing training.
  4. Professional bodies should be pressured to take computing seriously when accrediting courses and degrees, and consider accrediting general computing courses as well.
  5. Accreditation of computational training - equivalent of a PGCert?

Improving the understanding and status of training in research (3)

  1. Make attendance mandatory.
  2. Make training as engaging as research.
  3. There is no one-size-fits-all approach to students.
  4. Provide transferable values in training.
  5. Provide training for academic management.
  6. Make it mandatory but not useless.
  7. Establish branding.
  8. We need to set up a seed group that provides accreditation.

What statistics do we need to make the case for training and how do we collect them?

  1. Statistics are easy to collect, but are only useful for the funder, not the user.
  2. Understanding what are we trying to measure and how to measure it? What should be numerical and what not?
  3. Follow up with the aim of providing a story and raising awareness of further training
  4. What are the additional benefits of instructor led training face to face training?
  5. Collaboration

Working with technical writers to produce quality training materials

  1. Difficulty of communicating your own product, because you know it too well
  2. Resource problems - Need to share materials as resources are tight?
  3. Eight to one ratio - eight hours to create one hour of training material
  4. How to do quality control and ensure good material
  5. Web analytics and understanding where people have problems

Group Two (14.00-15.00)

The future of funding for training

  1. Need for a gap analysis on e-infrastructure training across disciplines.
  2. Leverage existing cross-university groups
  3. How to access direct/indirect funding
  4. How to access alternative funding streams - build shared knowledge
  5. Need to lobby strategically and speak with one voice

Providing incentives to produce good training materials

  1. Efficiency. Providing training material promotes efficiency and avoids redundancy.
  2. Face to face training provides a unique environment to understand what would make better software.
  3. Analysis of online metrics might provide an insight as to where people are finding difficulties with software.
  4. Kudos, recognition – requires both a technical social infrastructure for supporting recognition.
  5. Microaccredatation, attribution, provenance, licensing (cc-by, cc-0). Mechanism towards anti-siloing of training materials. Promote re-use.
  6. Standardised training material. Is it accredited? Is it internationally / nationally recognized. Who gives it accreditation? Will a measure of quality control provide an incentive.

How do we best support the community of people we train at scale?

  1. Support network post-training is important for medium- long-term success. Follow-ups maintain relationship, and can provide extra evidence.
  2. Different mechanisms for support: Doctor's surgery, research group experts, Skype, forums (but caveat that do not always succeed), social media (contentious!)
  3. Blended support combining face-to-face training and supplemented by online interactions post-training.
  4. How to handle success? - i.e. fast audience growth rate. Recognising when you hit critical mass, avoiding excessive overhead beforehand. Build the network - train the trainers and outsourcing are options.
  5. Naturally growing the community. Alumni-style networks possible, but must be useful.

What role should distance learning play in training?

  1. Massively Open Online Courses: Started with Stanford’s introductory courses. Now many other universities do them. Good way of delivering university courses at a distance. Also, distance learning Masters, Undergrad degrees and subscription based training. Includes Q&As, screencasts. Finally, open, community training eg. on YouTube. This must be moderated – up to date and vetted. It’s good to base this on something people already use, not on a new platform.
  2. Funding – who pays for it? If it gives you an accredited qualification people will pay – but otherwise, trainees will not want to pay. Technology is already here for delivering distance learning – that bit doesn’t need investment.
  3. The content of generic training is more suitable for distance learning – though it might not be necessary! Specialist training is hard to deliver online, but in many ways is more appropriate!
  4. Going to a training course means that the training gets your whole attention. It also delivers added value, for example networking and real working relationships.
  5. Upfront of investment in online training is lower, so it’s easier for trainees to switch to a more appropriate level.