1. science
2. medicine
3. medical research

eResearch Challenges
in New Zealand
Discussion
Document
eResearch
2020
Conversations
for change.
March 2015
eResearch
2020
TO VIEW ONLINE GO TO:
eRESEARCH2020.ORG.NZ
Overview
March 2015—eResearch2020
High-quality science and innovation
can have a transformational effect
on a nation. Research and development investment offers the power to
increase economic productivity and
competitiveness and improve health,
social and environmental outcomes
in unprecedented ways.
Indeed, New Zealand’s economic and
social wellbeing depends on the productivity and competitiveness of our
economy and the knowledge we have
C
to help make informed decisions as a
society. Science provides that knowledge and informs those decisions.
DR A FT NATIONA L STATEM ENT OF SCIENCE IN V EST M ENT:
2014–2024: M BIE, M AY 2014
Contents
Executive Summary
1
Introduction:
The future of research is digital 3
Research methods and expectations are
changing
3
We have been falling behind
4
Some fundamental messages include 6
About eResearch 2020
7
Four key observations of eResearch 2020
10
#1 Skills Lag: 11
The chasm in skills is widening 13
The challenge of change
15
#2 Research Communities: 16
#3 Align Incentives:
23
#4 Future Infrastructure:
27
Where should we be aiming?
31
A few expectations for research in 2020
34
Rethinking incentives for quality
36
Summary
39
Recommendations
39
Introduction
Executive Summary March 2015—eResearch2020
Executive Summary
eResearch 2020 is a 12 month qualitative study launched by
NeSI, REANNZ and NZGL in 2014. We spoke with researchers across a wide range of disciplines, as well as leaders of
ICT and management at our research institutions. In each
case we asked them to discuss the potential for change in
their work over the coming decade, and to highlight the
opportunities and challenges they see before us as a
national research system. The full interviews are published
at www.eresearch2020.org.nz.
What we found:
Our contributors made four major observations which we
think need greater focus in the context of national strategy:
11
Skills Lag: as a research system, we are under-investing
in the broad-based research skills and methodological training that underpins the international shift in research methods
to digital evidence and data intensive discovery. Consequently
we are putting at risk the quality of our research outputs and
our opportunity to grow high tech capabilities in the country,
today and into the future.
Research Communities: greater explicit investment support for research-discipline led collaboration is likely to lead
to greater cohesion, knowledge exchange, and collaborative
research outcomes across the system.
Align Incentives: governance and planning of contestable
funds is poorly aligned with the governance and planning of
institutional funds, leading to sub-optimal application of the
balance of government investment in research overall.
Future Infrastructure: provision should be made early
for investment in data, visualisation, and digital research
expertise.
Executive Summary March 2015—eResearch2020
Where to from here:
As a part of our conversation with the sector, we also sought
recommendations for change. These recommendations fall
into three broad areas:
1
Improve our professional research skills and standards:
·· Provide training and adjust incentives to address the growing skills gap in our research system.
·· Adopt comparative standards within and across disciplines for standardised data, common meta-data standards and ontologies for sharing information or making
decisions.
2
Adjust our approach to strategy and planning:
·· Promote sharing of risk and investment into the infrastructure and capability layer in our research system.
·· Better align governance, planning and funding cycles
for institutional and contestable funding in research.
·· Define our infrastructure requirements to address specific research community needs.
3
Focus our investments on future opportunities:
·· Emphasise strategies that leverage off-shore services
and infrastructure to preserve the on-shore human
capability and skills we require to be self-sufficient.
2
·· Research data infrastructure has been the missing
piece in our research system over the last decade.
“Visualisation” will be the key enabling technology in
innovation and discovery in the next decade.
Introduction
March 2015—eResearch2020
Introduction: the future of research is digital . . .
The distinction between “Research” and “eResearch” is disappearing. The standards, skills and expectations that are
currently the domain of a few, select, data, ICT, or eResearch
groups are diffusing into the wider research sector at a greater pace than ever. The speed of technological change means
rapidly adapting to new expectations must be a pervasive and
habitual component of a high performing research system.
Over the last few years, international research norms have
begun a shift to digitally driven research methods, and to new
standards in terms of evidence and publishing. The future
impact of New Zealand research and researchers will depend
on the ability of our research sector to adopt these new tools
and to excel against new expectations in a data informed and
digitally enabled research world.
Research methods and expectations are changing
The fourth paradigm of research — “data intensive discovery” — is expanding the tools and resources available to help
researchers understand the world at an accelerating pace.
We can now collect, share, and analyse data on a scale and
speed unprecedented in research history — and this holds
true across all fields of research and academia. In health, the
natural environment, ecology, agriculture, urban planning,
or in responding to security needs or natural hazards, we
3
see major growth in the use of telemetric sensors, genomic
sequencing, radio telescopes, social media, geospatial and
sources of real time information. To be meaningful, all of this
new information must be linked with meta-data, making it
searchable, storable, and reusable. In many cases, our power to collect, compute, and analyse this flood of information
outstrips our ability to visualise, interpret and communicate
the results.
Introduction
March 2015—eResearch2020
New technology is producing data that allows researchers
to move beyond theoretical models of the world around us,
and towards understanding and optimising systems —pastures, forests, factories, or energy grids—in real time. All the
while, international standards for the quantity and quality of
research evidence — data —are becoming more difficult for
New Zealand researchers to achieve.
M A R K GA H EGA N eR2020
We have been falling behind
Our abilities as individual researchers, and as a national
research system to collect, validate, analyse, visualise, store,
and curate research data are not keeping up with international expectations. With regard to training and skills, research
resources, and infrastructures needed, we are considerably
behind. New Zealand research institutions have historically
been late investors in new technologies as this minimises risk
“If we take a backseat in science
then we are really taking a backseat
in economic development and global
competitiveness.”
4
and reduces capital requirements; however the accelerating
rate of change still presents the challenge of increasing the
frequency and scope of our investments.
In investment, as well as in developing research skills and
shared research resources, our institutions appear to have
struggled to coordinate their strategies at a regional or national level. The culture in our institutions overly emphasises
the historical “information services” focus on their corporate
needs at the expense of prioritising needs that support their
research mission. Our major scientific endeavours appear to
Introduction
lack the effective, cooperative tools and support to be able to
live up to modern expectations for research methods.
Many of our New Zealand researchers are not sufficiently exposed to new information technology in research, nor are
they familiar or confident with 21st century methodologies
or requirements for reproducibility in science. Incentives for
researchers and institutions are focused firmly on producing
“more” research, rather than on “better” research — or on
improving quality in research methods. Instead, we observe
a planning and skills gap in the framework of our research
sector that leaves our universities, our Crown Research
Institutes, and our research community fragmented and often isolated in the research data space.
Evidence of this gap includes the often weak links
between the governance of research goals and long-term
planning in our research institutions. Additionally, we have
launched new initiatives, such as the National Science
Challenges, without a strong understanding of the digital
infrastructure, skills and governance implications of these
initiatives. We are only beginning to build collaboration in
data and research infrastructure into the foundations of our
planning (and not yet into our funding).
Only 5 of our universities make the list of the world’s
top 750 research institutions in terms of research impact
— the highest placed, Otago, sits at 249 on the current
rankings1 . One scenario suggests that a significant number
of New Zealand researchers may soon struggle to publish
5
their work in respected international journals. To ad“ Digital research infrastructure allows
dress this, we need a coordiresearchers i​n NZ ​to compete internated approach to upskilling
nationally at scale, to work on larger projects, and to ​make impactful​​ our researchers, resourcing
digital research from infrainsights without the complexities
of set up, operation or overhead​s​of​ structure and tools, through
to lifecycle management and
major ​equipment.”
1 http://www.leidenranking.com/ranking/2014 Global rankings for
Impact of papers (% in the top 10% of peer reviewed papers).3
unis didn’t rank in top 750 (#249 Otago 10.9%; #272 Auckland
10.5%; #420 Massey 8.7%; #514 VUW
7.5%; #522 Canterbury 7.4%)
TON Y LOUGH NZGL
March 2015—eResearch2020
Introduction
March 2015—eResearch2020
expertise. Most importantly, we need to alter the culture in
our research institutions towards quality engagement with
digital research methods and 21st century standards for impactful research.
Some fundamental messages include . . .
We can expect a splitting of corporate and research data
needs and consequently diverging investment strategies
for each. As service providers and “all of government agreements” mature, more and more of corporate data needs and
IT services can be provided effectively and at scale by third
party providers through cloud services and consumer technologies. Less institutional level investment will be required.
On the other hand, research IT, while technically advanced,
is likely always to be behind the corporate curve in terms of
mature commodity services — and therefore effective scale
— as the performance requirements and demand curves are
different. Therefore research IT investment strategies are
likely to focus on institutions cooperating for procurement
and delivery of research IT needs.
6
An increasing number of tools, both commercial and
open source, are coming online focused specifically on
digital research and the data lifecycle, but these are of quite
a different nature to corporate and consumer technologies as they require considerable training and a change in
researchers’ workflow. Furthermore, and right from the first
stage of data generation, research equipment and instruments (such as gene sequencers) are increasingly complex
and capital intensive acquisitions that often are not shared
nor well-connected to enable data processing and in some
cases remote operation.
Large scientific instruments are becoming more
complex and capital intensive, meaning that countries can
afford fewer than in the past. The expertise required to de-
Introduction
March 2015—eResearch2020
sign experiments and analyse results with these instruments
is increasingly hard to find, therefore the value of high speed,
high quality connections to overseas facilities is increasing
rapidly.
STEV E COTTER R E A N NZ
No single New Zealand research institution is likely
to effectively meet the specialised and various needs of
New Zealand research data at scale. Joint efforts, perhaps
facilitated with leadership from government or other neutral
infrastructure or policy agents, is likely to be required to ensure coordinated development across the research system.
About eResearch 2020
The eResearch 2020 outreach programme is a future oriented, national conversation with thought leaders within
the research sector that aims to assemble a comprehensive vision of researcher needs and essential skills over the
coming decade. eResearch 2020 is led by NeSI, REANNZ and
NZGL as co-patrons together taking a combined approach
to facilitating national dis“ We’re only doing things in half meacussions. eResearch 2020
sures, such as funding National Scibrings researchers across
ence Challenges without thinking
disciplines together to focus
about what support structures are
on particular themes, be it on
necessary. NZ needs a coordinated
research sector cloud strateeResearch ecosystem—it needs a
gies, skills gaps, institutional
comprehensive science data policy,
governance of research capaappropriate governance, incentives
7
bilities, or the infrastructure
for institutions to invest in data inneeds of the National Science
frastructure, a willingness to develop
Challenges and the Centres of
data-intensive research skills over a
Research Excellence.
long period of time, and a relentless
focus on quality.”
8
Introduction
9
March 2015—eResearch2020
Four key observations
March 2015—eResearch2020
Four key observations of eResearch 2020
1
2
10
3
4
As a science system, we are under-investing in
the broad-based research skills and methodological training
that underpins the international shift in research methods to
digital evidence and data intensive discovery. In addition, our
major capital and programme investments are disconnected, either geographically, or organisationally. Consequently,
we are putting at risk the quality of our research outputs —
our stock of knowledge and our opportunity to grow high tech
capabilities in the country, today and into the future.
RESEARCH COMMUNITIES: Impactful research collaboration
usually occurs at the inter-personal, research-discipline
level, not at project or institutional level. Greater explicit
investment support for research-discipline led collaboration
(rather than institutional collaboration), such as CoREs and
NSCs, and support for discipline-based national research
societies is likely to led to greater cohesion, knowledge
exchange, and collaborative research outcomes across the
system.
INCENTIVES DRIFT: Institutional funds, policies and practices
for ICT are applied to research methods with potentially ruinous consequences, distorting research design, subverting
resources, blocking collaboration, and leading to sub-optimal application of the balance of government investment in
science overall.
FUTURE INFRASTRUCTURE: Major innovation or discovery in
the coming decade is likely to stem from teams of researchers working across disciplines, institutions, and national
borders, based on data from a proliferation of sources —
including collaboration with industry — and with significant
reliance on compute, integrated systems, sensor networks,
and complex data analytics. Rather than lag behind comparator countries in science, provision should be made early for
investment in capacity and capability in data, visualisation,
SKILLS LAG:
#1: Skills Lag
March 2015—eResearch2020
and digital research expertise. New Zealand’s science system might then continue to contribute to economic productivity and competitiveness, and improve population health,
societal and environmental outcomes at first world levels.
SKILLS LAG:
Ensure researchers keep up with the pace of change
11
A critical problem within our science system and New Zealand as a whole, is a lag in the adoption of digitally driven
methods and eResearch skills, both in our research sector
and in industry. As more and more research activity moves
into the digital domain across all disciplines, we observe that
many New Zealand researchers are struggling to keep up
with changing standards within their international research
community and with the associated quality expectations
for evidence, data, and research methodology. Our abilities
both as individual researchers and as a national research
system to operate in a digital research world don’t appear to
be keeping up with international expectations.
This should not be unexpected. Many organisations
struggle to keep up with the pace of technological change,
and many research disciplines are particularly susceptible
to this struggle. Additionally, the competitive and commercial imperatives we instil in our research system encourage
a focus on shorter term outcomes than may be required
for long term skills acquisition. An informal survey quickly
reveals that over 75% of researchers do not have high speed
internet at their desks, usually in spite of their institution’s
membership of REANNZ, which suggests a lack of focus on
researcher needs within institutions.
Yet researcher needs are changing: until very recently,
almost all field research data collection in New Zealand was
#1: Skills Lag
March 2015—eResearch2020
done with forms and pencils. Not so long ago, researchers
sketched images of fauna or flora samples, recorded birdsong with sheet music, and noted broad based observations
of habitat rather than precise location data. Similarly, soil
samples were collected in jars, and geographical information
was limited to 10km2 plots. Digitisation of information, use of
connected mobile devices, high definition geo-spatial data —
these technologies and many more are changing the level of
scrutiny, the detail and complexity of information available
to our researchers.
NICK JON ES NeSI
“Our current approach (to investment)
means our research communities are
excluded from applying the most recent
advances in research methods aided by
the latest digital and information technologies, therefore excluding a growing
and highly innovative raft of methods
and approaches to research.”
12
New tools in computation and modelling permit not
only a deeper understanding of the world around us, but
also lift expectations for scientific discovery and research
quality. Many of our nationally significant databases are no
longer considered particularly large or complex data sources (though the information they contain still maintains its
relevance). What we can observe is a gap in our ability to
optimally leverage the information we are collecting in our
national databases, and a growing gap in the ongoing funding
mechanisms for supporting data sources in the long term.
#1: Skills Lag
March 2015—eResearch2020
The chasm in skills is widening
Key areas where New Zealand researchers and research institutions need help to lift their game include:
Reproducibility: much of our research output is not
sufficiently reproducible science, due to poor methods in
evidence, a lack of published method, or poorly interpreted
results. Greater rigour is required in designing and publishing
methods and data. In order to reproduce research findings in
digital research, we need to record and reproduce the provenance of research results, including the data semantics,
workflows, code etc. that produced the research output. No
matter the research field, features that drive reproducibility
need to be monitored and incentivised as keenly as we currently monitor and encourage the rate of academic publication. In some international research systems, data management planning and resourcing has become a first step in any
research proposal.
Long term sustainability of data/evidence: almost universally in New Zealand research, no support is available for
storing and managing data beyond the end of the research
project that generated it — this is especially the case for
public-funded research data. In several cases, New Zealand
research institutions have, or intend to, delete publicly funded research data that is no longer supported by research
project funding after the project has ended.
Standardised toolsets for research disciplines: we need
our graduates to emerge from university already equipped
13
with digital skills, tools and standards accepted and applied in the fields they are entering. We also need to adopt
comparative standards within and across disciplines for
standardised data, such as geo-spatial, genomic, or health
informatics data.
Research methodology and standards that flow into
economy and society: in leading digital research sectors,
such as genetics and human health, new technologies and
CRISTIN PRINT eR2020
#1: Skills Lag
March 2015—eResearch2020
digital data-driven methods are rapidly moving from research
to reality. In these fields in particular our researchers and
research institutions need to be able to lead the way in terms
of methods and practice. Unfortunately, even our leading
research institutions often still lack common meta-data
standards and ontologies for
“ At clinical scale the quantity of
sharing information or making
(genetic) data, and the processdecisions.
ing required to make sense of it,
Whether we expect the
will quickly eclipse the current
pace of technological change to
infrastructure capabilities in New
remain the same, or to accelerZealand.”
ate, the future implications of
this lag in digital research methods and evidence are concerning. Purely from a research sector perspective, we can
expect our lack of engagement with new methods to begin
to limit our ability to collaborate internationally, to publish
impactful research in leading journals, to access international research funding, and to generally lower the quality of New
Zealand research compared with first world nations.
There are without doubt exceptions to these consequences — we can quickly identify stars in our research
communities who are eResearch leaders on the world stage;
however the gap in skills between our research stars and our
general research population is worrisome. If we consider our
socio-economic wellbeing to be linked in part to our research
sector and our ability to understand the world around us, then
arguably this lag in our eResearch adoption may create limits
14
to our productivity, our social cohesiveness, or to our capacity
to monitor our environment, our borders, or our economy.
#1: Skills Lag
March 2015—eResearch2020
The challenge of change
As with much of society, change can be a challenge for researchers, sometimes individually, often as a community. To
a certain extent, negative engagement with new, technology
driven research methods is a feature of researcher culture:
in the pursuit of productivity researchers tend to stick with
the methods and tools they know. Researchers can be early
JA N E A LLISON eR2020
“A lot of the computer people in
research value their independence,
so are reluctant to cede control at root
level to ICT staff. We spend a lot of time
finding ways around corporate limits
to share data and back up.”
15
adopters — consumer tools (i.e. Dropbox, Twitter) are widely
used across all our research organisations. Researchers are
also highly independent, often preferring the self-sufficiency
and control that consumer applications allow them to the
institutional systems and attempts to homogenise that are
common in corporate ITS. This independence is often seen
to be at the expense of the scale, cost efficiency, and collaborative opportunity institutional IT services are expected to
achieve (though this needn’t always be the case).
In general, researchers are excited about the potential
new technology brings to their work; however they often believe themselves too busy with research and teaching to have
time to learn new tools or methods. In most cases, researchers will do whatever is required to get funding, suggesting a
strong motivation to upgrade capability if funding is attached
to a higher level of digital research skills.
#2: Research Communities
March 2015—eResearch2020
Many researchers will encourage their post-doctoral
and junior fellows to do the leg-work with digital methods and
analysis rather than learn these methods themselves. They
do this recognising an underlying change in the standards
of evidence and the expectations for impactful research.
There’s considerable scope for learning opportunities at the
institutional and community level to address this skills gap
if the mandate to provide the training were taken up. Furthermore, offering our researchers a better understanding of
the competitive research environment — both in New Zealand and internationally — and of the accelerating pace of
change, would contribute to expanding our researchers’ engagement with modern expectations for quality in research.
RESEARCH COMMUNITIES:
Design research infrastructure around users
16
Until now, our approach to infrastructure usually has been
centred on investment with research institutions, however
we observe that researchers and research are often better
connected within discipline-centric research communities
rather than at an institutional level. We also note that research communities are at very different levels of maturity
when it comes to digital research methodologies and data.
A clear concern we have seen in the research sector is
that the wide range of funding mechanisms are not explicitly
linked to infrastructure or long-term capability. In the case
of the National Science Challenges for example, the questions of collaborative infrastructure and data management
arguably needed to be dealt with in the upfront design of the
Challenges. Yet as infrastructure was specifically excluded
from being funded, it does not appear to have been treated
as a priority during the development of proposals.
A A RON MCGLINCH Y eR2020
#2: Research Communities
March 2015—eResearch2020
There appears to be an opportunity to gain greater
cohesion across the sector through enabling and promoting sharing of risk and investment into the infrastructure
and capability layer that underpins the National Science
Challenges.
We have found considerable demand for standards
and tools in the research sector that might help individual
researchers, research communities, and institutions align
on meta-data, quality and
“ Small groups of staff are forming
method. To a certain extent,
expert groups within Landcare, as
international standards such
pioneers to create unofficial groups
as ISO standards for research,
that are the first port of call for scior those standards imposed
entists with questions (rather than
by international research inigoing to NeSI).”
tiatives are beginning to bring
order; however the concern is
that as tools and data proliferate, it will become more difficult to share data for collaboration rather than easier.
In the drive to produce science of the highest quality
that is also linked to economic and societal outcomes, we’ve
already observed that our research sector has matured
unevenly in terms of skills, capacity, and impact. In general,
those elements or disciplines most closely linked to large
primary industry sectors are the most mature in terms of
the degree to which they collaborate, the high-tech skills
retained, access to equipment, and the ability to leverage
advanced methods. Conversely, disciplines linked to less
17
well-off sectors of the economy, or those research domains
that contribute more to social cohesiveness rather than
economic output, are in general less mature in their use of
advanced data tools or sophisticated digital and eResearch
methodologies.
If a knowledge-based economy is a policy goal, then
we need to develop tactics that produce excellent science
of the highest quality across all of our relevant research
JOCH EN SCH MIDT eR2020
#2: Research Communities
March 2015—eResearch2020
disciplines. This is not just for those with the highest
potential for short-term impact, but also those that focus
on sectors of future need or growth. A key criterion in an
increasingly data informed and digitally driven research sector is to ensure researchers maintain the capability and skills
to be able to innovate and to make serendipitous discovery
into the future. This requirement has implications not only for
investment strategies in infrastructure and skills, but also in
the potential for incentives to adopt technology and digital
methodology to be harmonised and embedded in government contestable funding.
There is arguably a gap in explicit funding for collaborative, inter-discipline endeavour. Possibly the most impactful links for collaboration and impact in science occur at the
“research-discipline community” level, within and across
groups of researchers who share a research domain or
discipline, rather than contained to an institution or a locale.
This suggests that mechanisms which promote connec“ We need off-shore input to advance
tivity, knowledge sharing and
science, and we make contributions
cohesiveness within research
back to off-shore science. Concerns
disciplines, yet on a nationof sovereignty are likely to have lital scale, will impact quality,
tle impact on this fundamental scicollaboration, and the pace of
ence paradigm.”
progress.
A fundamental tenet of eResearch, both in New Zealand and overseas, has been shared risk and investment
18
into capability and capital equipment. Shared investment is
most effective when needs and risk are aligned among the
participants, which is more often the case across research
disciplines that between research institutions. The National
Science Challenges are clear examples of a mechanism
which promote connectivity in our research system, however the collaborative underpinnings of shared risk and
investment in capability and equipment for these initiatives
SY DN EY SH EP eR2020
#2: Research Communities
19
March 2015—eResearch2020
is missing: instead the fundamentals of evidence and methodology appear to have been left to chance.
While international collaboration for education often occurs at institutional level, international collaboration
for research is almost exclusively based around research
communities or communities of practice. At an international
level, research data and infrastructure is consistently designed around the needs of a particular research community;
however this does not mean that it is limited in scale — CERN
data infrastructure is quite different to Elixir data infrastructure yet both are large scale, high performant, and significant
investments. While the scale may differ, the observation is
that research communities have specialised data needs, and
that we should avoid making “one size fits all” investments
across the entire research system as these too frequently
suffer from sub-optimal implementation and use.
The argument for designing research infrastructure,
and in particular research data infrastructure, around
particular research com“ Humanities in particular are still stuck
munities, follows from the
on the monograph as the output of
observation that evidence,
research endeavour. Even electronpractice, and collaboration
ic journal publishers are resistant to
in research differs from disDigital Humanities methods.”
cipline to discipline. Applied
engineers operate and think
differently to pure scientists; agricultural genomics has different needs and standards to clinical genetics; humanities
and demographics deal with different kinds of meta-data.
To be adopted and become embedded, research and data
infrastructure in New Zealand needs to fit both the digital
technology maturity of the research community concerned,
and the particular data and meta-data standards of that
research discipline.
The New Zealand Data Futures Forum notes that research data, whether publically funded or not, is likely to be
Introduction
20
March 2015—eResearch2020
Introduction
Small countries still have big
problems; we can’t rely on small
science to solve them. For example, to be at the leading edge
in dairy genomics is an economic necessity for New Zealand. A
small percentage difference in
growing grass, or dairy nutrition,
or milk production all have major
impact on our comparative advantage and economic fortunes
in the global economy.
For NZ, genomics is an
essential eResearch, compute
driven discipline. If we take a
backseat in the science, then
we are really taking a backseat
21
in economic development and
global competitiveness.
M A R K GA H EGA N eR2020
March 2015—eResearch2020
A RIA N DE WIT eR2020
#2: Research Communities
March 2015—eResearch2020
treated differently in terms of security, confidentiality, and
meta-data integrity, depending on the type of data and the
research discipline involved. To take the National Science
Challenges as an example, we’ve learned that not all of the
NSC teams have the same demands of infrastructure, and
that most NSC needs can be addressed with consumer
technologies. What’s missing for the NSCs is the organising
principle and leadership to help them make coordinated,
well-informed choices. Without well thought out policy driving key behaviours and incentivising decision-making at both
a researcher and an institutional level, we will struggle to
achieve the desired outcomes from our national investment.
Our overarching comment is that, in making investment
decisions in infrastructure, particularly research data infrastructure, the needs and maturity of the research community in question should be considered early in the process.
“ With both collaboration and data,
This does not mean that our
personal relationships and networks
research institutions do not
are still very important, especially at
play an important role —
the start of a project, as these build
only the institutions have the
trust and uncover data and expertise
organising structure and lonthe project needs to succeed.”
gevity to invest in and manage major investment programmes and infrastructure. In
the coming decade, New Zealand is likely to have one shot
at developing digital infrastructure at a national scale in our
research sector. We need to begin that development process
22
with cross-sector research communities, as well as with
individual institutions, if we aim to design the best outcomes.
To accommodate such a wide range of needs, we can identify
some central principles:
Stop defining infrastructure needs at an institutional
level, instead digital infrastructure should be designed for
particular research communities
#3: Align Incentives
March 2015—eResearch2020
Encourage institutions to make connected investments
right from RfP stage —build sector infrastructure planning
into institutional strategy making processes where possible
Define 8–10 research communities and then have them
specify their digital infrastructure needs to be met — and
don’t invest in digital infrastructure until we have these
research community needs defined
Recognise digital research communities have wide
variety of maturity, so embed tactics that lift skills within
immature communities, and across institutions, without
slowing down advanced communities.
ALIGN INCENTIVES:
Connect planning and research governance
23
We observe that “institutional governance” and “research
governance” work to different funding timeframes, operate
independently when it comes to planning, and are not jointly
engaged in the rapidly changing research environment. Our
research institutions are medium to large scale New Zealand enterprises that plan strategic activity against corporate timeframes over 5 to 15 years; however our research
committees and research funders operate research strategy against 2 or 3-year funding horizons. The outcome we
observe is research projects that are designed with limited
goals to avoid needing major infrastructure resources, but
that have budgets larger than necessary in order to purchase small scale capital items that do not lend themselves
to shared use — or long term management. This misalignment between research governance and institutional planning arguably limits funding for major capital items, as well
A LLISON JA N E eR2020
#3: Align Incentives
March 2015—eResearch2020
as limiting the resources our researchers can access and
share.
To the extent that the government employs both institutional and contestable funding across the science system,
we also can see misaligned incentives and disconnected governance of these different types of funding at an institutional
level. Indeed, many researchers complain that senior institutional management do not understand or wish to engage with
the changing paradigm of research methods and data. Investments in line of business applications necessarily dominate
institutional budgets, yet there
“ I think the barrier is existing funding
may be room for incentivising
models and individual institutions begreater cross institution proing limited by both their ownership and
curement of shared corporate
internal funding models. For example,
systems such as HR, finance,
putting funding for computing hardlibrary, learning management,
ware and/or time on a grant often isn’t
grant management, and other
possible. We can now put NeSI as a
corporate services platforms.
consumable on a grant, but if we were
While a level of shared instisticking to the rules, we’d still have
tutional information services
a hard time accessing the compute
investments might be worth
equipment we need for our students.
investigation, devising new
We can’t guarantee we’ll always have
incentives in the institutiona grant, therefore we like to keep a lital management system that
tle hardware around and available just
release investment to drive
in case.”
shared capability in research
and education, especially for
24
our best researchers and educators, are even more desirable.
It would be fair to observe that our institutions are
littered with minor capital equipment acquired through project funding, often poorly maintained or supported, and that
few others in the research sector or even the same institution are aware of —or use. The culture is to keep such assets
in the “department” and out of sight of wider institution. If
these assets do become shared (i.e. commoditised use) the
#3: Align Incentives
March 2015—eResearch2020
fully-costed use of the asset is typically revealed to be much
higher than expected; thus sharing is discouraged. This
represents a misalignment between research governance
and institutional planning, and arguably limits funding for
major capital items as well as limiting the resources our
researchers can access and share. The National Science
Challenges and the National Statement of Science InvestNIC M AIR eR2020
“A challenge for us is that we don’t have
the organisational structure in place to
harness the infrastructure and services
that are available to us.”
25
ment both suggest an attempt to improve whole of system
alignment, however it may be worthwhile for government to
investigate aligning the funding periods of the institutional
and contestable funding mechanisms — at least for minor
capital expenditures.
Our Crown Research Institutes, created to operate
broadly around commercial research models, are extremely
well connected to each other at the administrator and executive level. The CRIs are relatively small enterprises which
have recognised that at an infrastructure level they cannot
afford to act in isolation. Coordinated tactics and responses
at the CRI level are becoming more common. Unfortunately
this is not the norm in New Zealnd research —the chief architects of long-term planning and strategy appear to be the
leading researchers in our institutions, not the administrators or executives. Researchers, ingenious and experimental
by nature, devote much of their ingenuity in New Zealand
to acquiring project funding. Once again, this leads to the
situation where our research institutions feel continually
#3: Align Incentives
26
March 2015—eResearch2020
“cash-strapped” and our research infrastructure is unable
to scale or is poorly suited to shared use.
Some of our institutions already recognise this issue,
yet still struggle to coordinate their activities. We observe
that no two universities in New Zealand are taking the same
approach to crafting a long-term investment strategy that
can also support mandated short to medium term research
goals. NeSI, NZGL and REANNZ represent the few coordinated, significant long-term investments our research institutions have made in a joint fashion, each one with government
leadership, and of these only REANNZ has comprehensive
representation in its membership / customer base. Half of
our universities have formally or semi-formally instituted
committees dedicated to progressing digital and eResearch
methods; half do not. At a sector level, our system is highly fragmented — in many cases, our institutions appear to
struggle with collaboration, despite deliberate mechanisms
for coordination within their respective industry bodies.
In general, it is not clear where institutions should draw
“pre-competitive” lines so they can collaborate on national
infrastructure (for example) while creating healthy competition in research and academic. What we can observe is that
collaboration occurs more easily in the realm of contestable
funding than in institutional funding. In general, this is arguably due to the close, collegial nature of individual research
discipline communities. For this reason, government initiatives that fund collaborative research based around research
disciplines, appear more likely to produce impact than those
focused on institutional collaboration.
Furthermore, improving the cohesiveness, knowledge
exchange, and skills within research discipline based communities appears to produce greater quality scientific outputs than cross-discipline investments at an institutional
level.
#4: Future Infrastructure
March 2015—eResearch2020
FUTURE INFRASTRUCTURE:
Visualise data for innovation and discovery
27
Our capacity to innovate and make serendipitous discovery is
linked to our ability to understand and leverage information.
A major part of human interaction with information is visual
— the ability to see what is occurring is a crucial aspect of
understanding and of cooperative endeavour. Good visualisation is a powerful force in innovation e.g. the (ever larger
screen) smartphone is now a key tool for productivity and
society, and is fundamentally altering almost every facet of
daily life.
Visualisation of data, the ability to bring the power of
vision to computational modelling, data analysis and collaboration, is a future critical capability across all research
fields (and ultimately in major public services areas such as
health, law enforcement, and managing natural hazards).
In many instances current New Zealand research practice
is linear; data-sets inform batch compute processes which
produce models or answers that can then be assessed as an
end-product. As more and more research activity becomes
digitally driven, researchers (most likely working in teams in
multiple locations and even time zones) will need to visualise
this linear process as it occurs, so that they might intervene
to adjust, to experiment, and to essentially make discovery.
Critical gaps in our research capability for the coming
decade are in the tools, standards, and infrastructure to support advanced data, analytics and visualisation. In the very
near future, our ability to interact with and visualise data will
determine our capacity to innovate, to make serendipitous
discovery, to manage our economy and institutions, and to
respond to crises. In these terms, visualisation in research
will become less concerned with merely displaying the outputs of science, and instead will be the interface through
#4: Future Infrastructure
March 2015—eResearch2020
RICH A R D TEM PLER eR2020
which researcher understanding and discovery occur. This
will be even more important in research collaboration, where
researchers must work from the same observed points of
reference, or with datasets of enormous size and complexity.
“There’s an assumption of benevolence
on behalf of large corporations that is
not necessarily justified, so a practice of
not putting too many eggs in one basket
could be worthwhile.”
28
The NZ Data Futures Forum foresees a future where “data
will be abundant and ubiquitous; it will be connected and
linked to people, things and places; it will be used and reused, processed and reconfigured”. In this future, our capacity to innovate and make serendipitous discovery is closely
linked to our ability to understand and leverage information
not only in the research sector but across the wider economy.
Sensor proliferation is increasing the volume of data
that we have access to, not just research sensors but also
the “internet of things” and the ability to monitor our health,
economy, and environment. In the short term this leads to
spectrum crunch — the limited ability of our communication
technologies to convey all of this data without interference
or delay — which is a concern of its own. In the long term we
expect massive growth in data generation and transmission
— continuously generated, real time data, flowing towards
analytical systems that enable decisions to be made, and
data to be stored and later retrieved for action. In such an
environment, our researchers move beyond constructing
theoretical models of systems and can begin to analyse and
STUA RT CH A RTERS_eR2020
#4: Future Infrastructure
March 2015—eResearch2020
optimise systems in real time. This capability, to understand
a system in near real time, will present an innovation in research methods that may offer considerable advantage to
those researchers who can access it.
Data intensive discovery is the analysis of data to discover trends or new knowledge; however it presupposes
that the data is discoverable. The library system and now
online published journals have been the key research tools
in finding secondary information, yet significant data exists
in New Zealand institutions that is opaque to our researchers. Even something as ubiquitous as water monitoring becomes complex in New Zealand as the water management
data produced and held by our Regional Councils is heterogeneous, and not particularly integrated — we’ve little
standardisation or transparen“ Ultimately, the visualisation cy in this area. As a result, our
becomes the key interaction ability to understand, respond
and the tool for operation, to and manage flooding hazards
debugging and understand- and weather events is unnecessarily constrained. Unless they
ing the model.”
have interpersonal relationships across their community, researchers in New Zealand
have very little opportunity to discover the data that resides
in institutions other than their own.
Without a doubt, a degree of reliance on off-shore infrastructure providers is inevitable and welcome — Dropbox,
for example, is the number one shared tool in our research
29
community today. Indeed, there are certain things we really
want the market to run, those services where price and reliability matter should be provided by a market — provided
the market is willing and able to provide them. Research infrastructure, including networking, data, and compute, are
clear outliers here — the market is seldom able to provide
the reliability, flexibility and peak performance required for
impactful research at an acceptable price. There’s a culture
#4: Future Infrastructure
30
March 2015—eResearch2020
change required to come up to speed with the technological
change that’s occurring, as previously niche or specialised
services become mature and commoditised. New services,
such as cloud computing permit easy access to hardware as
a service, while potentially releasing resources to focus on
ever higher performant computing, more advanced networking, or greater specialisation in data technologies.
In our small, first world economy, the researcher consensus is that we should not aim to be isolationist, that
the speed of technology advance is such that we can only
keep up through extensive use of off-shore providers and
services. We should also endeavour to be self-sufficient in
those strategic circumstances that require it, and we should
actively invest to ensure the underlying human capability to
understand new technologies is never lost to us. Researchers consistently identify the key capability for New Zealand
as the ability to “tightly couple” data and computational
analysis to the analytical tools and visualisation that enable
decision support. This “tight coupling” ability implicitly includes elements of technical capacity, model manipulation,
and human skill — i.e. it is represented in both the skills of
the people we have access to, and the infrastructure and
technology we are able to leverage.
A challenge in New Zealand institutions is that the
skills required to provide and operate this level of technical capacity are not research skills, nor are they general ICT
skills, but require specialists in high performance computing
and networking, genomic sequencing and informatics. One
factor in the ability to attract and retain people with such
skills is the access to the infrastructure and equipment they
are trained to work with. If the ability to visualise data, compute and analyse in timeframes sufficient to allow decision
support and crisis response needs to become the governing
principle of infrastructure and technology investment, then
consequently the development and retention of talent pools
Aiming
March 2015—eResearch2020
that support this capability are equally vital. Our observation is that placing requirements such as “real time analysis”,
“computational steering” and “collaborative visualisation” at
the forefront of on-shore investment decisions will ultimately
contribute to New Zealand’s research and support key national capability in crisis response, decision support, and
goals in health, energy, the environment and the economy.
JOH N R AIN E eR2020
Where should we be aiming?
The government’s draft National Statement of Science Investment is clear about the links between science investment and the economic and social development of New
Zealand. We need our research sector to have the skills that
will eventually translate into
“ It’s vital we understand that eRescience making a greater
search isn’t something separate;
contribution to our economic
rather it’s how all research (including
competitiveness and social
primary sector research) will be done
well-being. A contribution
in the future—so eResearch needs to
also to improving our prebe built into all our plans, our projects
paredness to respond to criand our institutions into the future.”
ses, be resilient to hazards,
influence our economy, and
make informed decisions as a society. A clear focus of current policy has been increasing R&D as a percentage of GDP,
largely from business investment. We observe that New Zea31
land has a gap to close to ensure our researchers become
sufficiently familiar with digital and eResearch methodologies, and with the skills and capabilities expected of world
leading researchers, to be able to contribute to the policy
goals for the science system.
To achieve these national goals, our research sector
needs to function as a best-in-class small country “research system”, leveraging larger resources off-shore, but
Aiming
March 2015—eResearch2020
maintaining key skills and core capabilities in New Zealand.
To increase cohesiveness at a national level while maintaining competitive tensions in the system we need to adjust to
an eco-system view of digital research skills and infrastructure. This means taking an end-to-end approach to digital
research resources — not only in technology investments
(such as data or networking) but also across incentives and
skills training, policy, funding and governance. The more consistent and aligned our policies, strategies and incentives
are, the more efficiently we can implement digital methodologies and enhance skill pools with our limited resources
to achieve strategic outcomes for New Zealand. Some of the
core capabilities and skills we could consider strategic outcomes include but are not limited to:
The ability to understand and optimise at a system
level (e.g. a pasture; a water catchment; a hospital) in real
time, as data streams in.
The ability to plan for and collaboratively respond in
real time to an emerging situation (e.g. a volcanic eruption,
an earthquake, a disease outbreak) across multiple research
institutions and government departments.
The ability to visualise tightly coupled, complex data
and compute models, to allow interactions with data that
enable innovation and serendipitous discovery.
32
The ability to collaboratively deliver on research goals
across institutions, inter-disciplinary communities, with industry, and with local, regional, and central government.
The ability to track, store, secure, manage, and share
private, public, and proprietary data across social networks,
public services, and research institutions.
The changing pace of tools and technologies, and the
explosion in the use of models and data for decision making in government, industry and society make our ability to
Aiming
March 2015—eResearch2020
work with advanced computational and data analysis techniques in the research field even more vitally important.
New Zealand research is strong on foundational capabilities
such as open source software platforms and data lifecycle
leadership in specific domains (e.g. climate, bioengineering,
geonet), however these eResearch capabilities are yet to be
well-engaged by governance, strategy and policy in the re-
CRISTIN PRINT eR2020
“At clinical scale the quantity of
(genetic) data, and the processing
required to make sense of it, will quickly
eclipse the current infrastructure
capabilities in New Zealand.”
33
search system, nor are they linked within and across our research institutions. There is need for clear policy positions to
be developed for all publicly funded research which requires
and resources long term access to data and analysis models
(the evidence) as well as the research output. This may be
most easily achieved by aligning data requirements across
all the different research funding mechanisms, and ensuring
funding is available for ongoing data management (i.e. after
the original funded project has completed).
Technology is changing the future of research just as
rapidly as other areas of our society, yet by taking a silo-ed
approach we actively run the risk of not keeping up, or not
being able to participate. The infrastructure, skills and resources needed for research and data intensive discovery
differ from corporate and consumer needs, and access to
those skills and resources are a necessity in contemporary
global research. It is these research methods and data tools
Expectations
March 2015—eResearch2020
devised in our universities and research institutes that ultimately diffuse out to our society to then drive innovation,
improve health care, increase our social and economic development, and ultimately ensure New Zealand’s status as
a first world country.
A few expectations for research in 2020
Research Quality
By 2020, evidence and methodology in research will be completely digital. Research quality is assessed through the visibility and repeatability of research methods and evidence
in the public domain — in a digital world this places new
requirements on researchers. Open access to publications
and data will put researchers’ methods on display and play
an increasing role in research impact assessment, leading
to enhanced researcher skills and infrastructure services.
Digital Research Communities
By 2020, research, compute and data infrastructure will be
designed and deployed for particular research communities, rather than for individual research institutions. Data
communities will have a wide variety of maturity — some
communities will need significant development while other
research communities will demand advanced analytics and
data management infrastructure. Very advanced research
34
communities may move beyond national infrastructure and
become self-supporting in a global context.
Integrated Research Sector Planning
By 2020, sustained engagement and linkages will be in place
between the governance of research activity and the governance of research institutions. Integrated planning for
research strategy, institutional development and national
Expectations
March 2015—eResearch2020
infrastructure strategy will operate against aligned funding
horizons.
STUA RT CH A RTERS eR2020
National Capability
Beyond 2020, the ability to innovate and make serendipitous
discovery will depend on data, compute and networking that
is tightly coupled with visualisation. Visualisation is one of
the key ways in which you turn data into usable information.
It aids the discovery process and enhances capabilities —
computation is key to creating visualisation, and networking
35
“Ultimately, the visualisation
becomes the key interaction and the
tool for operation, debugging and
understanding the model.”
is key to interacting collaboratively with visualisation. Technology investment decisions for on-shore infrastructure will
need to lend themselves to this tightly coupled functionality
in order to support key national capability needs (including
crisis response and decision support), economic development, and goals in health, energy, and the environment.
More and more, data and compute that is tightly coupled to and interacted with through data visualisation will
play a role in how we understand real-time systems such as
urban systems, pastures and farming systems, demographic
systems and other myriad macro and micro socio-economic
system-level drivers.
Rethinking incentives
March 2015—eResearch2020
Rethinking incentives for quality
36
Our project funding and resourcing models for research need
to be applied to a different scale and timeframes if we are
to understand, manage, and leverage data for scientific insight, social development, or economic growth. For many
of our leading thinkers and major research institutions, the
quantity, frequency, and detail of data we can now gather
simply overwhelms the scope of funding and resource available to work with it. A number of factors are contributing to
this situation:
We typically invest late in infrastructure and big
ticket items, and many of our investments are designed to
catch up to international standards rather than to advance
with them. This strategy reduces risk in the investment and
ensures we take advantage of opportunities to collaborate;
however it limits our researchers’ access to new technologies, capabilities and techniques, and to the ability to make
serendipitous discovery. This lag in access appears to follow
through to a corresponding lag in adoption and ultimately in
research impact.
Our research sector is both small and highly fragmented — we haven’t a strategy for reaching a critical mass
of researchers and infrastructure. In many cases, any particular New Zealand research institution may only have one
or two researchers with skills in data intensive research or
computational analysis. Little cross-fertilisation of knowledge and skills occurs at a research institution level except
when researchers change institution (this is not to say that
researchers don’t share information at a community level,
below the institutional engagement layer).
As different research disciplines engage with digital
methods at different speeds, expectations of research output and evidence have become uneven across disciplines.
GAVIN M A RTIN eR2020
Introduction
Rethinking incentives
37
March
March 2015—eResearch2020
2015—eResearch2020
“ NZ likes to do per capita calculations to show
how good we are or how we manage our
investments, but a comparative per capita
measure is the wrong approach in this area
for such a small country — it is leading us to
under-invest.
Small nations, particularly those in
Scandinavia (and elsewhere too) are investing substantially more — and more diversely, that is in basic sciences as opposed to a
focus on applied science. It’s not because
they want to, or they love scientists and
engineers, but they see they have to in order
to build the science based research ecosystems necessary to underpin any successful
economy in the future.
To make the point, a comparative per
capita approach to making infrastructure
investments simply would not be affordable for China — and would inevitably lead to
over-investment. China has to use a different measure to be able invest at appropriate
levels for their economy — and so must we.”
Rethinking incentives
38
March 2015—eResearch2020
To a certain extent, our funding system for research
rewards quantity of published journal articles but does not
incentivise continuous improvement in research quality or
adoption of new methods.
Our fully-funded science system means funding for
generating data is captured within projects, which arguably
limits cross-project visibility of data and promotes duplication in data collection. A clear fault occurs when project
funding ends and project generated data is subsequently
lost.
We need to incentivise our researchers to upskill / improve their research methods and expectations by ensuring
funding and resources are tied to quality as well as quantity
— where “quality” is measured beyond the simple check box
of “published”. Different publications are of differing standards, and have differing expectations for evidence, reproducibility, and impact.
Government and institutions have already taken significant but partial steps, including brokering or providing
affordable access to infrastructure such as high performance computing. A further step will be the implementation of training and guidance resources, with accompanying expectations of performance that allow researchers to
upskill themselves and gauge their improved ability. Finally,
to overcome inertia and incentivise research engagement,
we might consider reserving premium research funding for
researchers who’ve demonstrated engagement and adoption of world-class methods and tools that have been made
available to them, and can therefore certify the quality of
their research. In the same vein, we would need to ensure
funding also reached promising young researchers who are
native to digital research and whose skills will spill-over to
lift the abilities of those they work with.
Summary
March 2015—eResearch2020
Summary
The aspiration is for the New Zealand research system to be
functioning as a best-in-class small country sector in 2020.
One facet of this will be a determined effort to augment the
capability of New Zealand researchers with stronger skills
in reproducibility, and the evidential integrity of our digital
research output. Only a small fraction of New Zealand researchers are really engaging with the methodologies, skills
and practices that likely will be required of all researchers
within the coming decade. Another aspect of best-in-class
sector design for small country systems will be comprehensive integration with major off-shore research resources, as
well as smart, coordinated use of market and commercial
providers for efficiency and scale.
Finally, we will need to maintain within New Zealand
the core digital skills and capabilities that are integral to
functioning as a country, and that will be increasingly central in supporting our well-being as a society. These include
the ability for essential government services to function, our
ability to respond to hazards, and the capacity to innovate
and design new products and services.
Recommendations
1
Provide training and adjust incentives to address the
growing skills gap in our research system when it comes to
digital methodologies and research quality. As funding ap39
pears to be the prime motivator of researcher behaviour in
our system, we should consider reserving a portion of current
funding as a “premium” or top-up for researchers — both
experienced leaders and promising younger generation researchers. In parallel, we need our research system to better
support those seeking to improve their skills in new research
areas and meet new expectations in terms of quality. In particular, we need our graduates to emerge from university al-
Summary
March 2015—eResearch2020
ready equipped with digital skills, accepted standards and
applied tools in the fields they are entering.
2 Too much of our thinking and strategy is focused around
the silos of our institutions, rather than across a broad community of researchers. We should take the opportunity offered by the National Science Challenges to enable and promote sharing of risk and investment into the infrastructure
and capability layer in our research system. This includes
promoting cohesiveness and knowledge exchange within national research communities. As well it recognises our digital
research communities have wide variety of maturity, thus implementing tactics that lift skills within immature communities without slowing down advanced communities (and their
associated industry R&D capabilities). To achieve this we
should try to stop defining infrastructure at an institutional
level, and instead design digital infrastructure for particular
research communities. We should resist investing in digital
infrastructure until we have the target research community’s
needs properly defined. We need research communities that
guide and coordinate infrastructure investments across our
research institutions, just as we need our research institutions to be central to the operation and provision of infrastructure services if our infrastructure investments are to
be effective.
40
3
The rapid speed of technological change may actually
offer us the opportunity to move away from a fragmented
approach and take a national eco-system view of digital research capability and how we meet changing expectations
in research quality. Eco-systems don’t develop overnight,
however taking a national view of digital research capability
as an end-to-end system begins to suggest some early policy and strategy opportunities. In practical terms we need to
adopt comparative standards within and across disciplines
for standardised data, common meta-data standards and
ontologies for sharing information or making decisions. We
Summary
March 2015—eResearch2020
JA M ES SMITHIES_eR2020
also need to front-end data management planning and resourcing as a first step in research and public project planning so that ongoing support is available for storing and managing digital resources beyond the life of the projects that
generate it.
At a more generic level, we can improve the efficacy
of allocating our limited resources by better aligning governance, planning and fund“ New Zealand needs to remember,
ing cycles for institutional
though, that we’re living in an increasand contestable funding in
ingly globalized world. That offers exresearch. We want to incencellent opportunities —and I for one
tivise both of these towards
am optimistic about the opportunities
quality as well as quantity
digital technologies offer to engage
of outcomes, and avoid the
with the wider world —but we can’t
sub-optimal allocation of
rely on the likes of Google, or even the
capital towards project reInternet Archive or the United Nations,
sources that don’t support
to preserve and develop our culture.”
shared use. Particularly desirable would be any newly
devised incentives in the institutional management system
that drive capability in digital research capabilities and infrastructure for our best up-and-coming researchers and
educators.
41
4
As a small, first world economy, the researcher consensus is that we should not aim to be isolationist, that the
speed of technology advance is such that we can only keep
up through extensive use of off-shore providers and services. At the same time we should endeavour to be self-sufficient in those strategic circumstances that require it, and
we should actively invest to ensure the underlying human
capability to understand new technologies is never lost to
us. Researchers consistently identify the key capability for
New Zealand as the ability to “tightly couple” incoming data
and computation analysis to the analytical tools and visualisation that enable collaboration and decision support – in
Summary
March 2015—eResearch2020
PETER H U NTER eR2020
fact, “visualisation” is likely to be the key enabling technology
in innovation and discovery in the coming decade. As such, it
could be the best foundation skill and infrastructure driver
for strategic investment decisions.
While it will take significant expertise in the field to
define investments at a technical level, our observation is
“Inevitably, our society’s
problems in the future are going
to have a data and computation
aspect to them.”
that placing requirements such as “real time analysis”, “computational steering” and “collaborative visualisation” at the
forefront of New Zealand’s on-shore investment decisions
and research capabilities will ultimately contribute best to
New Zealand’s research goals. It will also best support key
national capability in crisis response, decision support, and
goals in health, energy, the environment and the economy.
42
Acknowledgements
March 2015—eResearch2020
Acknowledgements
eResearch 2020 would like to acknowledge the input of our
expert advisory group and thank them for their generous
support and efforts in guiding this project.
•
Professor Penny Carnaby Lincoln University
•
Professor John Hine Victoria University of Wellington
•
Professor Keith Gordon University of Otago
•
Associate Professor Blair Blakie University of Otago
•
Professor Shaun Hendy University of Auckland
•
Associate Professor Cristin Print University of Auckland
•
Tim Chaffe Enterprise Architect University of Auckland
•
John McMaster CIO Plant & Food Research
•
Lynley Smith IS Manager (Acting) GNS Science
•
Steve Cotter CEO Research and Education Advanced Network
New Zealand (REANNZ)
43
•
Tony Lough CEO New Zealand Genomics Limited (NZGL)
•
Nick Jones Director New Zealand eScience Infrastructure (NeSI)
To view the document online go to: eresearch2020.org.nz
2020
eResearch
If we consider our socio-economic wellbeing to be linked in part to our research
sector and our ability to understand the
world around us, then arguably a lag in
eResearch adoption may create limits to our
productivity, our social cohesiveness, or
to our capacity to monitor our environment,
our borders, or our economy.