Earlier this month, the Association for Research Libraries and the Coalition for Networked Information held a joint forum entitled “Reinventing Science Librarianship: Models for the Future,” in which I had the privilege of participating (I’m fairly certain I was the only graduate student in the room).
The schedule, proceedings, and a handout (PDF) containing a set of e-science talking points for library directors are all available from ARL, but just in case you want to know what happened at an extremely minute level of detail, I’m also posting my 30 pp of notes.
(I apologize for the formatting in advance: fixing all of Word’s spaghetti code is just more time and frustration than I’m willing to give…)
Here’s the first day’s portion:
October 16, 2008, 3:30-5 pm
Convener: Wendy Lougee
- “It’s not enough to be agile, you have to have a plan for collective action”
- Contemporary models for sci libraries aren’t aligned w/ where science is going
- Not just “paying the Elsevier bills”
- History – Dan Atkins, NSF Office est., Data Curation issues (“To Stand the Test of Time”), Task Force formed in 2007, changed to Working Group later
- Fundamental Q: How can we as libraries mobilize to deal with emerging trends in science/e-science?
- – With NITRD at NCO
- – NCO is in Office of the President, Office of Science and Tech Policy
- – NITRD draws on agencies like DARPA, the EPA, NASA, NIH, NIST, NOAA, etc.
- – E-science means diff things at diff times
- – To CG, E-science is (quoting straight from his slide):
- o Digital data driven
- o Distributed
- o Collaborative
- o Trans-disciplinary (not quote: requires more lifelong learning)
- o Fuses pillars of science:
- § Experiment
- § Theory
- § Modeling
- § Simulation
- § Observation/Correlations
- – Does the data deluge make the scientific method obsolete?
- o Seems like an overstatement to CG (statement made by Chris Anderson), but there’s something there
- – Come a long way in the last ½ century
- – ARPANet->NSFNet->Internet
- – 1st cell phone weighed 88 lbs; now they weigh 2 ounces and get connected at a rate of 30 per second (3.8 bn total so far)
- – Advances in computing have changed “not only what’s possible in computation, but what’s conceivable in computation”
- – How many computers do there need to be? (How many are there?)
- – How many libraries do there need to be?
- – Tech is now woven into the fabric of our society – important to banking, finance, education, entertainment….all sorts of things
- – Net result: a change that transforms technologies transforms society in the process – and vice-versa
- – Imagine a world where:
- o All the text in all the libraries in the world can fit in a storage device in your pocket
- o The network responds at the speed of light to a plain question with a perceptive answer
- o Your contact lenses merge your digital and physical worlds
- – “A world of higher dimensionality”
- – CI reduces the conventional barriers of place & time
- – Large scale info access is reshaping science in lots of ways
- – Role of U’s & Academic Libraries
- o Universities – involved with creation, collection, preservation, and dissemination of knowledge (historic and continuing role)
- § E-education – information driven, accessible, distributed, interactive, context-aware
- o Libraries – strategic plan for ARL: role is to preserve and provide access to digital artifacts
- § Reinventing the Library – Institutional Commitment, sustainable funding model, defining the library user community, legal and policy frameworks, library workforce skills, library as a computational center, sustainable technology framework
- – Moving data around all the time is not an efficient use – want to bring analysis to the data, not the reverse – makes library look less like a repository, more like a supercomputing center
- – LTER – Long Term Ecological Research Network
- – Long term research is needed to reveal slow processes or transients, episodic or infrequent events, decadal trends, multi-factor responses…
- – 26 sites in LTER network
- – We have a lot of the tools that are needed to support data management/preservation – Morpho, MetaCat, EcoGrid, Kepler
- – In about 2005, LTER got early adopters, and then LTER went out and “proselytized,” and usage shot up
- – Science will drive what LTER does – particularly climate change
- – Need breadth of knowledge at our disposal – from looking at things broadly to looking deeply at patterns and processes – in order to look at Eco science
- – Tech directions – CI enabling the science (not CI for CI’s sake), Whole-data life cycle, Domain agnostic solutions (since, given the financial situation, we won’t necessarily be able to develop systems for each individual domain)
- – Focus on CI enabling the science (end-to-end solutions) – discovery, use, and access of info; open access to holdings and tools
- – Support the data lifecycle – need reliable replicated storage infrastructure, interoperability solutions (WM looked at 7 data ctrs, found 8 diff metadata solutions – bad!)
- – Data interoperability – ontologies & semantic mediation
- – Domain Agnostic Solutions – tools that can work across humanities, social science, and sciences
- – Kilo Nalu Workflow – Kepler workflow – not incredibly user-friendly, but receiving wide use; allows you to create a data workflow, visualizations, document links between this and other elements (publication, etc)
- – Kepler has proven at least somewhat domain agnostic – has been used in a wide range of fields
- – Workflow Sharing Portal – allows scientists to actually replicate the scientific process
- – Sociocultural Directions
- o Experiential, career-long education and training – science doesn’t stand still – moves too fast to just get your degree and stop learning
- o …missed some…
- o Citizen Science – e.g. Cornell lab of ornithology
- o Building Global Science Communities of Practice via CI
- o …will include a wide range of partnering institutions – including not only academia, but also industry
- o Longevity of CI Enterprises
- § Broad, active, community engagement, educating students about best practices in science; extension of existing outreach programs
- § Transparent, participatory governance
- § (2 others, but slide changed too fast)
- – UNM is undergoing a reorg – specifically trying to reenvision CI (UNM “Academic CI Planning for the 21st Century”)
- – Characterization of the environment: “responding to a world of opportunities”
- o New methods of research in a networked world; rise of e-science
- o Data science & data scientists require new org environments
- o Social drivers: education reform and accountability; dawn of tech enabled social tools to connect prof & personally
- – Questions are turning to “where is the data? Where is the analysis? Where are the visualizations?” – answer – it ought to be in the library [RL has a big diagram showing areas where library is, and where it needs to be)
- – Data publishing/repositories
- o Authoring – describing the context of data collection for machines and humans
- o Curation – store, manage, preserve data for long term use
- § Metadata req’s
- § Software creating metadata integrated with instruments & sensors
- – Self-correcting databases
- o More responsibility for PI’s – publisher and curators (data publishing)
- o Bi-directional communication – database providers<->users
- o Role for librarians as middleware – facilitators of data-driven partnerships
- – Collaborative Webspaces
- o Shared Reading/Discovery
- § Virtual whiteboards
- § Discussion tools
- § Shared annotations
- § Searchable discussions & transcripts
- o Webspaces supporting:
- § Wikiomics/wikipathways
- § ….etc
- o Browse information using:
- § Maps of concepts & domains
- § Collaboration & citation networks
- § Taxonomies
- – Shared ideas, Shared presence, Shared creation
- o Can attack these problems at a small scale for small teams
- o Don’t need everything to be on the scale of the LHC – or a single individual collecting bugs in the Amazon – somewhere between
- – Challenges
- o Architectures that scale & interoperate
- o Making knowledge networks visible and traversable
- o Playing a different game in a different arena
- – 4-Layer Cake: Scientific domains – workflow – support – (?)
- – New Org Structures
- o Hybrid orgs – subject specialists -> intradisciplinary teams…
- o Fluid staffing loads
- o Dynamic and diverse teams
- – Need both informaticians (practitioners of informatics) and Informationists
- (embedded librarians providing research and knowledge management services to health professionals in context)
- – Differentiation: key success factors:
- o Char of successful orgs:
- § Do something others can’t
- § Do something well that others do poorly
- § Do something others have great difficulty doing well
- o What percent of research lib content and services are unique? What % of budget is going toward unique services?
- – Library collaboration (existing drivers)
- o Eliminate duplicative work
- o Geography (local & regional networks)
- o Common shared characteristics (spec coll)
- – Different dimensions for collaboration
- o Behavior norms
- o IP
- o Workflow & handoffs
- o Financial investments
- o People & staffing
- o Proj mgt & leadership
- o Sharing distributed infrastructure
- – Need to shift from focus on products to focus on process.
- – Collaboration to the next level
- o Might borrow idea from cloud computing – don’t have to think about how it works, but can use it – cloud collaborating?
- Wrap-up & synthesis of today’s speakers
Q: Seems like tech is not in the library, but outside of it – does RL think the tech should move into the libraries?
RL: Is defining libraries in a virtual sense – space where information systems & social interaction come together; fits somewhere in the workflow – not about using the technology in the actual building
Q: Domain Agnostic Solutions – Q from Stanford – some domains have a lot less development in terms of standards & metadata infrastructure than others – should the libraries be the ones to play catch-up?
WM: Good question… Answers through description of Kepler development – realized there needed to be changes, but didn’t want to go off on a separate tangent to incorporate them – took an integrative approach – brought people in from other domains and asked them what the system needed in order to work for them. Also need funding agencies & others to do more to make software user-friendly, do the documentation… Written commitment to shared aspirations, division of credit helps too.
CG: Time to rethink metadata – not only what it is, but what it means. Formal published literature is itself rich metadata – being able to connect publications to the data would be a great step forward in the metadata world. Works the other way too – data provides metadata for the publication. Libraries have a really important part in the solution to this problem, since formal literature is such a strong piece of that solution.
Q: Nobody said anything about campus IT folks – does the panel have a comment on this area, does it also need to be explored?
RL: If you think about it as a team, that would be a set of skills that would be needed on the team. Seems like a fuzzy space between libraries & IT. Ideally, librarian would be able to converse/be credible in that environment.
CG: Some librarians are CIO’s as well – increasing integration between the two. Academic campuses have a rich IT environment and lots of library resources – and a lot of faculty and students producing data – very special. Hopes the trend of interaction between librarians, schools of Info, schools of Computer Sci, and IT depts. continue to keep in touch.
WL: Not a monolithic culture to begin with.
WM: Has been thinking through what might be a valuable course for undergrads on data management. Hasn’t historically been much out there for undergrads to learn about that stuff – just stats – not sufficient in this day and age. Would facilitate their research, education, etc., for students from all depts.
Q: How did the librarians get involved in the team that you put together (to WM)?
WM: We brought them in. Identified needs within the project, and then went out and looked for that expertise wherever it was. Sometimes that didn’t make WM very popular. Had key need for librarians, went out and brought them in from all around.
Q: [From Office of Cyberinfrastructure person:] Seems to her that what took us from special purpose to general purpose computing was the identification of abstractions – into message layers, communication, visualization – all that depends on moving between layers. Seems like one role for librarians might be to figure out what those abstractions might be – help with solutions for preservation and access. [Not a Q, so no answers.]