Nectar Feedback 5 November

Posted under: From the Blog 4th of November, 2010

The Nectar consultation paper is at http://nectar.unimelb.edu.au/docs/NeCTAR_Consultation_Paper_October_2010_FINAL.pdf

What follows is a personal response to the paper. I have not gone to the detail, rather wanted to make three key points: build on what is working, focus on research leadership, manage projects as programs.

Grow the Babies – improve existing services

Invest in ARCS v2 – user stats reflect take-up, despite ARCS clumsy interfaces, several volume services seem to be on a strong improvement trajectory. Don’t turn the promising services off just when they seem to be on the up: 2,500 evo users.  600 data fabric users, and counting. Extend their growth v2 period on condition of progressive sustainability.

The same goes for the AAF and any other services that can show a vibrant researcher user base and show progress in establishing sustainable funding. Build on and improve services that have existing volume. Address known weaknesses in improved versions.  Nectar should fund project capital for targeted service improvement, where commitment for operating exists or can be made.

Invest in overlooked but critically needed fundamental service usability, especially ease of discovery and first use payback. Take a lesson from the success of the iPhone apps store.  Services that have been adopted and seem to be well used reflect not just the effort of their creation but the massive effort in adoption by those researchers and their groups that have learnt to use them. Reduce that effort and provide earlier return as encouragement.

University IT groups are mostly out of denial about their need to support research, some have been for a few years.  While their historical focus has been on survival issues like admin systems and basic ICT infrastructure, a handful of unis now have well organized research support groups. CAUDIT can probably help bring this leadership to bear across the sector. Bridging is needed here to help the IT groups and their nascent engagement with the research community. Deeper research/IT relationships can expose the services IT has and that can be extended to help research. Again build on what is there – fund bridging projects through CAUDIT or anyone else well organised enough to get a result.

Can we also get out of denial about how much use is made and will be made by researchers, and everyone else, of cloud services? ARCS is starting to build articulation into cloud services, most unis use the cloud for student email. Individuals vote cloud with their feet every day – Dropbox or Mobileme anyone? Facebook, Youtube, etc , etc.  These are mainstream services that can be exploited by researchers individually and in groups. Issues like data protection and IP are readily handled with effective risk management. Why are we building clumsy research analogues – this is expensive entertainment when we might instead do more to use what the cloud is offering and innovating, with cheaper and easier to use services. Accessing the cloud should be  a much bigger part of all programs.

In summary, build bridges to existing services, improve their usability and serviceability, and integrate these existing and easily added research sector services, University IT Services, commodity and Cloud Services. Bring them together for the researcher.

Help the Family – foster research leadership to use process to get value from technology

Be careful not to try to fix a technology adoption problem with just more technology. Researcher busyness is a big barrier to adoption. We need to understand and exploit the social dynamics of research groups and their processes for adopting technology to improve their research process, e g by targeted development of HDR student and early career researchers as early adopters and research process innovators. Basic business analysis and consulting techniques for key research team members who innovate process and can become next generation research leaders.

Consistent service framework is required across projects and service providers. And a sensible approach to reuse, redevelopment of existing services. Again think apps store. Accessing 6 research services should not mean 6 different provisioning processes and another 6 passwords, 6 more user interfaces. Any reduction in unnecessary diversity is worth having. This is not about one size fits all but about reducing unnecessary diversity down to requisite variety – how many research workflow systems do we need – more than 1 and less than one different one for every researcher. Perhaps an ideal is one per discipline and one framework across them? Research community leaders will be fundamental to this increased standardisation as discipline and research community-led standards tend to stick as long as they don’t run counter to, or too far ahead of market standards. The reality of consumer technology and market standards needs to sink in too – researcher innovation should focus on discipline and immediate process. What is good for the research process?

Discipline orientation is fundamental to ICT in research process adoption curves and is generally global. Particle physics, astronomy, genomics, and other disciplines progressively embed ICT in their research processes, as technology reaches new price performance points for the next stage of exploitation in research tools and process. We need to ensure we are building into the global trends, using what is being done internationally as foundation, building on this prior and prospective effort and becoming part of global coalitions around discipline embrace of ICT. Nascent researcher leadership into these global roles should be fostered. Capability should be grown with personal tools and resources for researchers who can lead aggregation and re-composition of existing and new services into new targeted workflows for their groups. Australian researcher leaders are well placed to broker international research coalitions, resolving standards and process transition issues.

Nectar should fund capital projects with this broad perspective of what is being built – capability, process and data are infrastructure too in the world of in silica research. Look after the family using the technology.

Look After the Body – beware project reductionism weakening the whole.

Virtual Labs are a particular metaphor for research process orientated re-composition of lower level and existing tools and services.  Large separate projects should demonstrate use of existing lower level services and consistency at least at discipline level. The temptation to build it all anew must be avoided. Project leaders should expect to demonstrate that they are in touch with external developments and minimising duplication.

The Lead Agent model has an inherent danger due to the understandable desire of DIISR and the lead agents to delegate risk as quickly and with as little non-project overhead as possible. Rapid project fan-out, and down, reduces opportunity for tactical coherence across projects however. Programmatic management will be necessary to achieve more than a basic project management consistency of standard reporting and so on. Basic project management is fine of course if projects have little interdependence, and the odd one can fail, but that is probably not the case with Nectar, or for that matter the data or network programs. Ultimately these things all come together around the researcher, where coherence will be essential. Fragmentation has been one of the biggest problems with efforts so far and more effort must go into bringing it all together.

So the bodies of projects will need to be managed quite actively as programs, and the programs will need to be well joined  as an endeavour to improve researcher process and hence productivity. Light weight, passive program and endeavour management will not be adequate.

Many interdependencies are known but many will emerge, e g multiple projects relying on one stretched vendor. All will require active management across project boundaries.

2 comments. Would you like to comment?

Add your comment