ASIST Annual Meeting 2006 report
incorporating the DASER (Digital
Archives for Science & Engineering Resources) Summit III and Tri Society
Symposium
Information organization, preservation, and selection
Information retrieval
Information use
Knowledge sharing in online communities
Grants
Web Quality Evaluation Tool
Claire McInerny and Nora Bird at
Rutgers. An outdated link is
http://www.scils.rutgers.edu/~clairemc/foodsec/webqualitytool.html--work has
been ongoing since 2000. More detail on CD-ROM. A wide-ranging set of scales
from grammar to coverage to link logicality, it produces the following ratings:
• 71 - 82 = Excellent Site,
trust it
• 64 - 70 = Very Good, bookmark it
• 57 - 63 = Good, but proceed with caution
• 50 - 56 = May offer something, but don't trust information without
investigation
• Below 50 = Use the site information with a grain of salt
(pronounced “ferber,” as in Edna)
A metadata scheme to link works, editions, translations,
summaries, etc., instead of just creating separate main entries and hoping users
will find most of such related records. [How close is this to a semantic Web,
at least Berners-Lee’s idea of one?] MARC records can be “FRBR-ized” by
algorithms using authority files. See
PowerPoint presentation by Ed O’Neill, one of its major figures.
What does this look like to an end user? Examples:
FictionFinder (search for, say,
Edna Ferber; or try the tag cloud—tag
clouds explained in Wikipedia); the
North Carolina State University
library catalog so famous for its advanced Voyager features, which is
FRBRized—what the exact relationship between FRBR and the facets in the left
column of the results screen is yet to be elucidated. The interface presented
is Endeca, under consideration by
the UT Arlington Library Webvoyáge Advisory Committee, mentioned on the
UTA Library Catalog Design blog.
Scientific data records management
a.k.a.
data curation (see, for example,
UIUC’s Data
Curation Education Program (DCEP): an increasing need as data proliferate,
records are sought out for re-examination years later, scholars change
institutions, etc. Reliability, accuracy, and authenticity must be maintained.
This is somewhat related to database administration. One organization devoted
to the issue: International Research
on Permanent Authentic Records in Electronic Systems (the
InterPARES Project)
based at the School of
Library, Archival and Information Studies at the
University of British
Columbia. It is funded by U.S. NSF and Canadian
research councils. Data curation (organization, preservation) is [becoming?] a
huge need, which librarians are missing. UNC has a program in it.
Access to scientific data: social and technical challenges and strategies
IP rights: economics of database
protection: Coase’s theorem (cost of producing a database, and government’s
cost effectiveness in supporting database-producing institutions).
The assumption that no incentive leads to no production,
has no proof, in the case of databases.
Communities: One
researcher described best practices that support the needs of a designated user
community for a data collection, especially when there are multiple contributors
or data are collected by community-based observation. “Involving dataset
managers in the collection process and encouraging scientists to define and then
support the needs of their community has led to enhanced access to data.” In
other words, make data collection a conscious part of community effort. (These
communities may be groups of scientists or agencies, not necessarily natives
taking readings for scientists.) This supports format security,
interoperability, extensibility, transparency, compactness.
Unanticipated users: beyond
“designated” users. Is such broad use appropriate? [Can it be prevented? What
steps should be taken in collection and archiving to ensure future value and to
prevent inappropriate use (extensibility)?]
Federated searching
From a study of user perception of MetaLib:
- Users frequently did not understand federated searching, or selecting
certain databases to include in the search. When asked to describe (or draw
a sketch) how it worked, they could not answer accurately.
- Used for preliminary searching, not after all else failed.
- “Useful, but hard to use.”
- In post-interviews, only 27% of users mentioned the SFX button when
asked what the system displayed in their searches.
- Interest in help displays: 27% of students, 5% of librarians.
Implementing in libraries
- Users have Google expectations (always get results), but fail—they do
not distinguish between finding articles and searching databases. They make
assumptions we fail to match. They seldom ask for help.
- University of
Rochester has an anthropologist on
staff, for ethnographic research. Some of her activities:
- “Recount your steps in writing a term paper.”
- Midnight dorm visits with
video cameras
- Gave out disposable cameras and had students take pictures of their
study environments.
- Screen shots
- “Mark, on a campus map, everywhere you go.”
- Users do not want to have to choose (databases) before they search.
They want relevance ranking.
- SFX pop-up confused students. Now the display includes title, citation,
abstract, “more from (database name)”
- Some users develop their own
folksonomies (there is also a
Wikipedia entry) with
their peers. They want general user reviews à la Amazon, as well as a “peer
reviewed” indicator (which is possible only in individual databases’ native
interfaces. Most have never heard of ILL. Many use recommendations from
their friends as the primary source for choosing databases [not faculty?
Here they seem to stampede after anything mentioned in a classroom.] An
excellent source of findings is
Fear and loathing of implementation: Examining the instructional issues
surrounding the rise of federated searching, a PowerPoint presentation
by Lynn Lampert, Cal. State Northridge, citing MySpace.com and the
socialization of the Web: Implications for libraries discussed by Dr.
Noshir Contractor, University of Illinois Urbana Champaign (but no longer
available on the Web, apparently).
Context in information retrieval (plenary address by
Susan Dumais of Microsoft)
- Relevance feedback is even more important and desired
now, as purchase suggestions become more helpful. (This carries over into
online catalogs, as exemplified by the UTA Library Catalog Design blog
entry, “Libraries
playing catch-up to commercial vendors .” (See previous paragraph about
desire for user reviews.)
- Methodology of dealing with the context issue:
- Assay user differences w/r/t performance
- Isolate sources of variation
- Accommodate: do something about it.
- Domain of context work: ranking determinations,
presentation
- Search macros
set subdomain
- Memory landmarks (not actual date, but relative
mnemonic such as “when I was doing x”
- Especially useful with desktop search engines
- Capture what was typed just before—query dynamics
- Capture time of year, e.g., “turkey” links to
Thanksgiving and Christmas
-
Microsoft Community Bar: side bar for discussion of page displayed
- Personal relevance—rich model (based on desktop
search, browser history)
- User interface can be enhanced [remember this person’s
employer]: tabs, fisheye (larger font size => higher relevance), slider,
metadata such as geographic location
Miscellaneous notes from other libraries: if we build it, will they come?
Austin's closure of its undergraduate library
- General public was OK with it; librarians were outraged. William Wulf,
president of the National Academy of Engineering, tells the story of the
button makers at the dawn of mass production: they felt that their
high-quality, customized, hand-made product would guarantee their continued
employment, because the new factory-produced buttons would soon be seen as
inferior by the public.
- Undergraduate library was repurposed to include more study and
conference space, and a writing center.
- Austin's aspirational peers: Cornell, Dartmouth, Georgia Tech,
Michigan, Emory, Indiana, Stanford, . . .
- Their latest idea is an open building with modular walls.
Interacting with information
From Danielle Plumer
(pronounced plumber), a medievalist technologist at Austin's i-school:
- Predicted by the
Institute for
the Future and IEEE Spectrum and
the Pew Internet & American Life
Project (the"Pew Internet Study"):
- further progress towards a common language (which may be English--if
not, then non-Roman characters in domain names)
- privatization of libraries
- universal translator, OCR of handwriting, speech recognition within
11-20 years.
- There will be more interaction with information: modeling,
manipulation.
- Social communication technologies to create social presence in libraries
- Library 2.0: user-centered, multimedia, inclusive of users' presence,
folksonomies, more use of software in interaction (see p. 4 of Lankes's
presentation linked below)
- Collections are moving increasingly online, but services are not.
- Are users getting anything out of library blogs and wikis? Awareness of
others can lead to better communication with them. [Theoretically.]
- Characteristics of the library of the future to look forward to:
- Interpersonal information transfer (a librarian can be the other
person)
- Communities of discovery
- Experimental
- Technologically attractive
- Serving two levels of users: the minimum
I-just-need-two-articles-for-my-assignment and richly interactive (an
optional additional layer, not confused with basic content)
- Crucial points made in Dave Lankes's presentation,
Participatory networks: The library as conversation (pdf)
- (Knowledge is created through conversation. Books, videos, Web
pages are merely artifacts of those conversations.)
Scientists' information behavior
Institutional repositories
From proceedings: “[D]igital repositories are unlike
digital libraries such as Perseus,
in that repositories provide a submission mechanism whereby scholars can
deposit an electronic copy of their work at the time or soon after creation.
Digital repositories are also called open access archives because of the lack of
tolls, fees, or other legal and economic restrictions to access the content they
make available.”
There can be a reluctance to contribute because the
repositories are institutional, not disciplinary. [This is another need for
harvesters such as
OAIster.]
DRM is coming up in institutional repositories. :-{((} An
ALCTS conference session addressed this.
The networked book
- Sam Hastings: “I envision a different type of working environment where
the scholarly and research endeavor is much more closely tied to cyber
communities and personal digital libraries,” i.e., a move away from paper
and towards dynamic online discussions.
- Postprint discussions captured (then selected by an editor and printed)
- Author becomes editor
- It is “peer review on steroids” [an idea similar to the filtering
mechanism of arXiv.org and the like]—but which reviewers qualify?
- Which version does the library catalog?
- Intellectual property? Copyright?
Information behavior of chemists
- Does electronic availability enhance use?
- Chemists are less likely to embrace OA
- Most accessed: references, experimental sections (ACS supporting
archive)
- Highly accessed ≠ highly cited—often they just look up a method.
- PubChem—repository for
assay methods and structure data
- (as opposed to papers or journal articles)
- Use and
discovery processes (especially of chemists)
- (how much they use literature in coming up with new ideas)
- NCSU departments each have two library liaisons—one faculty and one PhD
student.
- Methods:
- critical incident technique (asked scientists for paper they
published recently, seminal paper they found significant (cf.
Brenda Dervin’s questionnaire)
- interviews, work flow process analysis using cards (thinking,
discussion, experimenting, literature search, . . . )
- Novelty is not everything—also linking back to older literature which
may provide a new explanation
- Scientists’ questions when perusing the literature:
- How to find other people with whom to consult or collaborate?
- Knowledge claim or discovery?
- [What, then, are engineers’ questions?]
- Janet Arth of Library Enterprise Operations at
U. of
Minnesota studied chemists’ citation
behavior:
- their use of JCR and LJUR in local libraries
- Self-citation affects impact factor hugely.
- LJUR is useful, but also, check publication year to decide on
backfiles (which are very important).
- It is good to compute a local impact factor.
- It is good to analyze download frequency data.
Highlights:
Do people have a bias against sponsored links retrieved in
a search? (Study was conducted by someone employed in the sponsored link
business.) Trust in sponsored links was ranked lower than regular retrievals;
but some users don’t care.
Chat reference affect factors: It is easier to hold back
(not bother to express emotion). Users wanted to know librarian’s name—wish for
personal connection. Users also wished for more emphatically emotional
communication (i.e., expressive tools available to them as well as to the
librarian, to express dissatisfaction, disagreement, correction to discourse).
Flow (Csikszentmihalyi)
in experiencing online searching: roughly speaking, the ultimate in searching
satisfaction, in which the user feels enjoyment, concentration, and control.
Architecture of
complexity (Barabási): Bursty activity is caused by a complex of heuristics
humans us to prioritize their tasks.
Most research on knowledge-sharing communication has been
confined to sharing within an organization.
There are various ways to
type communities of practice, such as sharing behaviors, software used, and
formality of organization.
New ways of
sharing include [of course] blogging.
The question is:
is a blog a tool of collaboration? For example, is a course blog interactive?
[The Internet study,
Bloggers: A portrait of the Internet’s new storytellers [pdf] suggests
not—“The main reasons for keeping a blog are creative expression and sharing
personal experiences. The majority of bloggers cite an interest in sharing
stories and expressing creativity. Just half say they are trying to influence
the way other people think.”]
Social informatics: one major expert is
Elisabeth Davenport of Napier
University.
The salient finding is that online communities of practice,
often outside professional communities, do facilitate sharing and accomplish
problem solving. [Shades of
The wisdom of
crowds?] For example, one researcher studied airline frequent flyers
participating in FlyerTalk, an online
forum: “In some cases, airlines are effectively paying them to fly. These
communities are adept at exploiting the unintended interactions of travel reward
programs and airline reservation systems, turning to complexity of these
products back on their corporate creators.”
The ways in which communities come together and translate
their problem or goal into an interactive, sharing mechanism (by agreeing on
standard representations—“a common language”) include
sense-making
(Brenda Dervin was very much present at the conference—her
IMLS project goal
is “to develop boundary bridging concepts that enable more effective application
and collaboration in both system design and research.”)
“The dark side of Wikipedia”: saboteurs, those who game
the system, and trolls. Research on trolls is very scarce. Even the definition
of “troll” is not firmly established. Are they simply people with different
points of view? Perhaps better would be to ask whether their behavior is
constructive or not.
Pnina
Shachaf is conducting research but it doesn’t seem to have produced any
publications yet.
A community of practice is built by surveying, focus group,
and building an e-home with blogs and wikis.
Tools: social network analysis, counting which profile
fields filled in at services such as blogger.com (can detect cultural biases).
Findings: in personal blogs, Chinese share more knowledge,
U.S. share more
personal information. In corporate blogs, however, there is no difference.
Social ties are growing weaker while virtual ties (between people who have no
other connection) are growing stronger (sharing, affirmation).
Examples: Ryze,
tribe,
LinkedIn, Friendster,
CustomerVision BizWiki.
Characteristics of online communities:
- Information seeking leads to information sharing.
- News [within such communities?] is social, contrary to professionals’
imaginations.
- In ICT (information communication and technology)’s triad of
corporate-military-private, each group denies the other.
- There is a need for “IT volunteers” in disaster response. See the Word
Summit on the Information Society. See also the
World Summit on the
Information Society and
comment in Wikipedia.
The topic of user studies and user-oriented research was
used to study convergences, divergences, and potentials for dialog. In other
words, can practitioners of different disciplines talk to each other? The
activity of “bridging gaps” or spanning boundaries was studied using focus
groups and journaling.
Interdisciplinary research is viewed favorably, but
considered difficult because of entrenched differences, competition over
funding, different terminologies and mental models. The solution is to build
teams, not try to design everyone into a monolithic system.
Grant
availability—a presentation of IMLS
National
Leadership Grant program
- Modeled on NSF
- Types of projects funded:
Collections Management, Community Engagement, Conservation, Demonstration
(to improve practices, programs), Digital Collections/tools (building),
Formal Education, Informal Learning, Partnerships, Public Programs, Research