The Law Library of Congress began harvesting legal blawgs in 2007. The collection has grown to more than one hundred items covering a broad cross section of legal topics.

Questions:

• What are the criteria for inclusion here? (It’s apparently curated, not an opt-in service like I suggested previously. Note that the two strategies are not incompatible.)
• The archived pages are openly accessible (just like pages in the Internet Archive). This is contrary to concerns which some had raised about copyright in either the blog or third-party content (like blog designs). Is there a special exemption for LOC? Or are they relying on the same argument as the Internet Archive (or Google’s cache, for that matter)? (P.S. Is the Internet Archive’s argument fair use or a specific statutory exemption? What about Google?)
• Did LOC ask the authors of these blogs for permission before archiving them, or do they consider it part of their general duty to collect and preserve? (For that matter, do the authors of these blogs even know they’re being archived by LOC?) Does it matter, either as a matter of etiquette or of law?
• The archiving frequency for all the blogs seems to be monthly. Should it be more often, less often, dependent on the blog (some are updated more frequently than others) …?
• Are other branches of LOC harvesting blogs in those subject areas, too? Are other national libraries doing this?
• What’s the history of the program? What’s the case they make for doing this? What’s it cost?

I don’t do much with preservation, so I plead ignorance. Is anybody preserving scholarly blogs (or other “grey” online scholarly resources), other than general-purpose national archiving and the Internet Archive? (Is it en vogue to do Web archiving for certain domains/types of contents vs. general purpose?)

Also, a suggestion: What about an opt-in archiving service for scholarly blogs? (Not opt-in for an individual page, as with WebCite, but for an individual blog/site, to be harvested on a recurring basis.) I don’t suggest it as the end of a solution, but as the beginning of one. An interested library or archive (or consortium) could, I assume, provide such a service fairly cheaply (e.g., someone could build the service in their basement over a weekend). The opt-in could be a simple Web form, asking for the URL of the site and some metadata, maybe also getting a license to provide open access to the preserved copies if the original goes dark. (P.S. Preservation/copyright experts: Is such a license needed?) This might provide a higher level of service than general-purpose Web archiving, e.g. the ability to categorize sites by scientific domain or topic, more frequent/robust archiving than is accorded to pictures of cats with captions, etc.

I imagine I’m glossing over an entire body of literature on the topic. Hopefully someone will let me know where I’m wrong!

National Science Foundation
Edward Seidel

Data-driven collaborations for complex problems: Affecting every field of science
Science and society being transformed by cyberinfrastructure and data-driven approaches (e.g. Pampers simulating diapers) — cf. Wired, “The End of Science”

NSF vision: national, integrated system to enable new paradigms of science
1. virtual organizations for distributed communities
2. high performance computing (what do we do with the data?)
3. data visualization / interaction
4. education & workforce

any cogent plan must address the phenomenal growth of data in all directions

goal: catalyze development of a system of collections that’s open, extensible, and evolvable; support development of new tools and services
national digital data framework, integrate with national CI

issues:
how to build data-driven science? methodologies, culture, education, working cross-disciplines
technological and economic issues: how do we do it? how do we pay for it?
open access: data, software, publications — we need to change the paradigm

rough plan: fill out holes in current CI; workforce development; bring in computation science; new problems: new groups to attach complex problems

Lucy Nowell

current NSF policy on data management is vague — no real consequences. goal: provide for clear, effective, transparent implementation.
concern: if we require sharing, we could damage the science. not everyone within NSF agrees on open data.

data sharing required for replicability and thus integrity of science
we need to accelerate the pace of science to deal with 21st century challenges — “we can’t afford cottage science anymore”

we’re now producing data faster than we can score it
related projects: long-lived repositories project by Bernard Reilly of Center for Research Libraries; Blue Ribbon Task Force on Sustainable Digital Preservation and Access
DataNet: achieve long-term preservation and access with systems that are economically and technologically sustainable

new international task force on data preservation and access? potential first meeting at OR meeting in May 2009

Sylvia Spangler

critical role of BRDI: act as conduit for encouraging National Academies to actively address these issues

Q: What can this group say when NSF program officers tell us that our science will die if we have to share data?
A: Nowell: When the community owns this, then the program officers who stand in the way will find themselves overwhelmed.

Q: Mike Carroll: Is the choice completely open vs. completely closed? Can we have deposit requirements with limited or delayed access?
A: Nowell: There’s a lot of issues — from data on human subjects data, to classified data, issues of data curation.

National Institutes of Health
Elliot Siegel of National Library of Medicine

[P.S. room is overcapacity -- nice that there's interest here]

Data Sharing policy: if you get more than $0.5m, your proposal must include a plan for making data available to other researchers or explain why you can’t
also Public Access Policy, GWAS Policy (must submit de-identified data to dbGaP), Clinical Trials Registration and Results Reporting

unique aspects of biomedical data: wide range of uses (clinical and research) and users, privacy

topics of interest:
how to recognize and reward data sharing by researchers
how to embed informatics/data sharing in training
development of a research agenda for managing scientific data and information
promote effective approaches for data sharing

Q: Cathy Wu: Consideration: collaboration with different sectors. Policy issues with EHRs.
A: Collaboration is being promoted by HHS — will be interesting to see how this plays out.

Q: Robert Chen: When government tells external researchers to share, intramural researchers may not always do so. We discovered a group at NIEHS has lots of useful data that they haven’t gotten around to sharing. How to lead by example?
A: I’ve been involved with the tech transfer officers for each institute, they’re responsible for working for intramural researchers to help them do that.

Q: Lesk: What can we do to help you?
A: Once upon a time, had a problem communicating grant requirements to investigators — less of a problem today — but there’s an educational role to be played. Public access is the will of Congress and the will of the people — at some point you’ve got to stop debating and focus on getting it done.
Q: NIH doesn’t take as much credit as you could — NIH is really doing a lot to support data availability.

Q: Attempt to explore universities taking stewardship for research data — is that a growing trend?
A: Gregory Farber, National Center for Research Resources: I think it is a growing trend. The idea that data’s going to sit in centralized repositories will become the exception rather than the rule.
A: NIH is looking at a number of different ways to promote this.

Defense Technical Information Center
Phil Casey

Our mission fits well with what the board is trying to do
We primarily work for DOD, but we make it publicly available when possible
Primarily medium of exchange is text
About 50% of documents are unclassified, unlimited

Data: taking a few exploratory steps

Potential actions for BRDI:
draft federal policy/responsibilities

National Institute of Standards and Technology
George Arnold

Standard Reference Data Act — allows NIST works to be copyrighted, unlike most gov. works
NIST databases: some free, sells some, subscriptions for others
how do we fund data preservation? if we can charge for access, it’s funding for preservation [This is a worrisome line of thought...]

Q: Robert Chen: NIST – CODATA
A: We’ve been an active participant in CODATA.

Q: Michael Lesk: How do you collaborate with federal agencies which share similar data needs, e.g. data exchange?
A: We haven’t look at this at a strategic level.

Library of Congress
Peter Young

Library’s science collections must address the digital transformation of scientific research

LOC digital initiatives:

• American Memory
• National Digital Information Infrastructure & Preservation Program
• World Digital Library
• E-Deposit of Electronic Journals

Potential model for data?
2005 LOC report: LOC should collaborate, but necessarily as primary curator
eScience Team est. 2009: to develop position papers, collection policies, policy recommendations

Q:
A: Board can help LOC understand what our role should be, and help communicate it to Congress.

Q: If a federal agency had responsible for digital data, I’d think it’d be LOC. Would LOC want this?
A: LOC’s primary audience is Congress — we do what Congress tells us to do.

Institute for Museum and Library Services
Joyce Ray

Funding for repositories since 2002; but not the primary funder
Funded CDL, FCLA – DAITSS, Alabama Digital Preservation Network, U. of Denver – Collaborative Digitization Program, New Jersey Digital Highway, JHU – National Virtual Observatory, UCLA – Cuneiform Digital Library, MIT – preserving CAD architecture designs

Purdue: developing services for researchers who want to share data

Notes from the first open session, Briefings from Federal Interagency Data and Information Groups.

[missed part of the beginning]

Interagency Working Group on Digital Data

Recommendations:

• working group should become National Science and Technology Council subcmte
• departments and agencies should develop their own policies but make their policies publicly available
• agencies should promote a data management planning process for projects that generate preservation data

Q: What about Obama’s CTO?
A: Chris Greer: This is a good sign for us.

Q: Working with libraries?
A: Greer:

Q: Universities may have counter-sharing cultures, e.g. concerns about tenure.
A: Greer: NSF has some experience here. Proposal should describe impact not just on your area of science, but on society as a whole. Initially we got 15 page proposals with 2 sentences on broader impact. Later, we got better responses. Researchers got serious about it. We could change the culture in a similar way.

Q: What incentives might we use?
A: Greer: We find carrots easier to use than sticks. Ex.: Physical research organizations: Limit access to equipment. Virtual organizations: Your prestige depends on how *many* people are using your equipment. Similar for data. We need ways to track use of data by others, important for promotion, prestige, rewards. Sticks: If you have a management plan, you can review performance under that plan and deny further increments or renewal if they don’t follow through.
Cita Furlani: We don’t have any sticks — it’s all carrots.

Q: Michael Lesk: What needs do you have that this group might help address?
A: Greer: We hoped for a broader look at the role of sectors outside the federal government. One of the most important is the academic sector. What’s the perspective from the university in terms of preserving their faculty’s digital products? What’s their commitment, biz model, what are their needs & expectations from us? Same questions apply to commercial sector: How will we interact with the commercial sector for long-term preservation and access? Ex.: Google Sky built on top of federally-funded telescope. We need to think about how to match needs with business model.

Q: Human consent: Not much standardization in consent agreements. How can we retain datasets from human studies, letting others use them?
A: Greer: There’s significant effort across government in area of EHRs. Point we make in the report: No one size fits all. Different communities of practice, different kinds of data have different kinds of constraints. Many agencies have this as a central issue: HHS, Census. Bottom line: It’s an unsolved problem. But we have expertise and mission-specific interest in these areas.

Q: Have tried to convince wealthier universities to spend more money in this area. But people in the libraries think that someone is supposed to do the work for them before they do anything — originally done by publishers. This is a new idea. Harvard isn’t used to thinking about preserving the output of Harvard. [Is this really true?] What’s the precedent for thinking this way?
A: Greer: Universities have mission for creation & dissemination of knowledge. Role of individual university re: its faculty is up to that university. There are a number of universities conducting experiments in this area: Big 10 repository effort, National Virtual Observatory, California Digital Library, variety of others, Georgia Tech. There’s a lot of space to explore here. Digital preservation doesn’t have to be physically centralized — can be very distributed. This is space worth exploring.
Q: If there were funds made available to individuals or universities — requirement and funding (authority and resources) — to do preservation, then they’d take it more seriously. Universities now think it’d be nice if we could. Nat’l Virtual Observatory: People don’t trust it, think it’s a one thing. Even with Worldwide Telescope, people won’t use it, won’t trust anyone besides the government because they don’t see permanence. If we pair mandate with funding for long-term archiving, then we might get results.
A: Greer: There are people exploring this — there’s not one solution, there’s probably a range of them.

CENDI
Ellen Herbst of the National Technical Information Service, on behalf of CENDI.
CENDI is voluntary inter-agency group of federal STI managers. Member agencies represent 97% of federal R&D budget.

Mission & goals: coordination & leadership, improvement of STI systems, promote understanding.
CENDI shortly will publish list of STI issues for policymakers to consider

Q: Goals, priorities?
A: Herbst: We’re career employees of the federal government — hesitant to advocate on behalf of policy. We come from a broad range of agencies, with our own issues — to rise to the level of CENDI, it has to apply to everybody.
A: Elliot Siegel: Each agency publishes its own report.
Q: Would it be a problem to identify common issues?
A: Elliot Siegel: May not be agreement — agencies may prefer to speak for themselves.
A: Herbst: We also work with non-member agencies across the government.

Q: What would you like this board to do?
A: Herbst: We’ve talked about outreach. December workshop: We got real scientists to speak about what matters to them — a bit novel. One goal is to commit to follow through on common areas of interest, exchanging information.
A: Siegel: We try to point out the value of STI, of collecting & disseminating it. We’d appreciate assistance in articulating this. For instance, we need to articulate the value of libraries — the Internet doesn’t just solve all problems.

Q: Data on value proposition, ROI: Anecdotes are useful.
A: We were going to do a white paper for the new administration, but instead made it a briefing paper. In preparing, we gathered some.
A: Bonnie Carroll: It’s hard to do that — we have vignettes, but it’s hard to quantify.

Q: Paul Uhlir: We did a workshop with OECD on assessment methodologies for economic & social benefits of access to PSI online. We’d like to follow up with a focus on public scientific information. We’ll discuss tomorrow.

U.S. Group on Earth Observations
Greg Withee of U.S. Geological Survey

Discussed GEO, USGEO, and GEOSS
Top international priority of USGEO: Full and open data access

Q: Michael Lesk: Different countries have very different policies re: availability of data. U.S. vs. UK Ordnance Survey vs. Germany weather data. What are the relative merits of these policies?
A: Withee: Wasn’t always like that — in the ’80s and ’90s, Europe moved more into commercial ventures — also France. Many articles on failures of those approaches — they haven’t produced the income to support the ventures, and cut off later value added. [cf. Bayh-Dole?] Data policy for GEO — companion paper might be useful to point these out, but politically it can be challenging.
Q: Paul Uhlir: Paper with OECD compared U.S. and European models, access and re-use, esp. geodata and meteorological data.