Tuesday, December 19, 2017

2017-12-19: CNI Fall 2017 Trip Report

The Coalition for Networked Information (CNI) Fall 2017 Membership Meeting was held in Washington, DC on December 11-12, 2017. University Librarian George Fowler and I represented ODU, which was recognized as a new member this year.

CNI runs several parallel sessions of project briefings, so I will focus on those sessions that I was able to attend. The attendees were active on Twitter, using the hashtag #cni17f, and I'll embed some of the tweets below.  CNI has the full schedule (pdf) available and will have some of the talks on the CNI YouTube channel. (I'll note if any sessions I attended were scheduled to be recorded and add the link when published.) The project briefings page has additional information on each briefing and links to presentations that have been submitted.

Dale Askey (McMaster University) has published his CNI Fall 2017 Membership Meeting notes, which covers several of the sessions that I was unable to attend.

DAY 1 - December 11

Plenary "Resilience and Engagement in an Era of Uncertainty" - video

CNI Executive Director (and newly-named ACM Fellow) Clifford Lynch opened the Fall meeting with a plenary talk.

Cliff gave a wide-ranging talk that touched on several timely issues including the DataRefuge movement, net neutrality, generative adversarial networks, provenance, Memento, the Digital Preservation Statement of Shared Values, annotation, and blockchain.

Our recent work investigating the challenges of timestamping archived webpages (available as a tech report at arXiv) is relevant here, given Cliff's comments about DataRefuge, provenance, Memento, and blockchain.

Archival Collections, Open Linked Data, and Multi-modal Storytelling
Andrew White (Rensselaer Polytechnic Institute)

The focus was on taking campus historical archives and telling a story, with links between students, faculty, buildings, and other historical relationships on campus. They developed a system using the Unity game engine to power visualizations and the interactive environment. The system is currently displayed on 3 side-by-side monitors:
  1. Google map of the campus with building nodes overlaid
  2. Location / Character / Event timeline
  3. Images from the archives for the selected node
The goal was to take the photos and relationships from their archives and build a narrative that could be explored in this interactive environment.

Always Already Computational: Collections as Data - slides
Thomas Padilla (UNLV), Hannah Frost (Stanford), Laurie Allen (Univ of Pennsylvania)

Always Already Computational is an IMLS-funded project with the following goals:
  1. creation of a collections as data framework to support collection transformation
  2. development of computationally amenable collection use cases and personas
  3. functional requirements that support development of technological solutions
Much of their current work is focused on talking with libraries and researchers to determine what the needs are and how data can be distributed to researchers. The bottom line is how to make the university collections more useful. There was a lot of interest and interaction with the audience about how to use library collections and make them available for researchers.

Web Archiving Systems APIs (WASAPI) for Systems Interoperability and Collaborative Technical Development - slides
Jefferson Bailey (Internet Archive), Nicholas Taylor (Stanford)
Jefferson and Nicholas reported on WASAPI, an IMLS-funded project to facilitate the transfer of web archive data (WARCs) or derivative data from WARCs.

One of the motivations for the work was a survey finding that local web archive preservation is still uncommon. Only about 20% of institutions surveyed downloading their web archive data for preservation locally.

WASAPI's goal is to help foster and facilitate greater local data preservation and data transfer. There's currently an  Archive-It Data Transfer API that allows Archive-It partners to download WARCs and derivative data (WAT, CDX, etc.) from their Archive-It collections.

Creating Topical Collections: Web Archives vs. the Live Web
Martin Klein (Los Alamos National Laboratory)

Martin and colleagues looked at comparing creating topical collections from live web resources (URIs, twitter hashtags, etc) and creating topical collections from web archives. The work was inspired by Gossen et al.'s "Extracting Event-Centric Document Collections from Large-Scale Web Archives" (published in TPDL 2017, preprint available at arXiv) and uses WS-DL's Carbondate tool to help with extracting datetimes from webpages.

Through this investigation, they found:
  • Collections about recent events benefit more from the live web resources
  • Collections about events from the distant past benefit more archived resources
  • Collections about less recent events can still benefit from the live web and from the archived web 

Creating Topical Collections: Web Archives vs. Live Web from Martin Klein

DAY 2 - December 12

From First Seeds to Now: Researching, Building, and Piloting a Harvesting Tool
Ann Connolly, bepress

bepress has developed a harvesting tool for faculty publications in their Expert Gallery Suite and ran a pilot study to gain feedback from potential users. The tool harvests data from MS Academic, which has been shown to have journal coverage on par with Web of Science and Scopus. In addition MS Academic pulls in working papers, conference proceedings, patents, books, and book chapters. The harvesting tool allows university libraries to harvest metadata from published works of their faculty, including works published while the faculty member was at another institution.

Being unfamiliar with bepress, I didn't realize at first that this was essentially a product pitch. But I learned that this is the company behind Digital Commons, which powers ODU's Digital Commons, so I was at least a little familiar with the technology that was being discussed. 

bepress was recently acquired by Elseiver, and this was the topic of much discussion during CNI. The acquisition was addressed at a briefing "bepress and Elsevier: Let’s Go There", given by Jean-Gabriel Bankier, the Managing Director of bepress on Day 1.

Value of Preserving and Disseminating Student Research Through Institutional Repositories - slides
Adriana Popescu and Radu Popescu (Cal Poly)

This study investigated the impact of hosting student research in an institutional repository (IR) on faculty research impact (citations). They looked at faculty publications indexed in the Web of Science from six departments at Cal Poly and undergraduate senior projects from those same departments deposited in the university's Digital Commons. For their dataset, they found that the citation impact increased as the student project downloads increased. One surprising finding was that the correlation between faculty repository activity and research impact was weaker than the correlation between student repository activity and research impact. The work will be published in Evidence-Based Library and Information Practice.

Annotation and Publishing Standards Work at the W3C - recorded
Timothy Cole (Illinois - Urbana-Champaign)

Tim presented an overview of the W3C Recommendations for Web Annotation and highlighted a few implementations:
Tim also talked about web publications and the challenges in how they can be accommodated on the web.  "A web publication needs to operate on the web as a single resource, even as its components are also web resources."

Tim also gave a pitch for those interested to join a W3C Community Group and noted that membership in W3C is not required for participation there.

Beprexit: Rethinking Repository Services in a Changing Scholarly Communication Landscape - slides
Sarah WippermanLaurie Allen, Kenny Whitebloom (UPenn Libraries)

Since I had learned a bit about bepress earlier in the day, I decided to attend this session to hear thoughts from those using Digital Commons and other bepress tools.

The University of Pennsylvania has been using bepress since 2004, but with its acquisition by Elsevier, they are now exploring open source options for hosting Penn's IR, ScholarlyCommons.  Penn released a public statement on their decision to leave bepress.

The presenters gave an overview of researcher services provided by the library and an outline of how they are carefully considering their role and future options.  As they said, Penn is "leaving, but not rushing." They are documenting their exploration of open repository systems at https://beprexit.wordpress.com/.

There was much interest from those representing other university libraries in the audience regarding joining Penn in this effort.

Paul Evan Peters Award & Lecture  - video

Scholarly Communication: Deconstruct and Decentralize?
Herbert Van de Sompel, Los Alamos National Laboratory

The final talk at the Fall 2017 CNI Meeting was the Paul Evans Peters Award Lecture.  This year's honoree was our friend and colleague, Herbert Van de Sompel. Herbert's slides and the video of the talk are embedded below.

Herbert discussed applying the principles of the decentralized web to scholarly communication. He proposed a Personal Scholarly Web Observatory that would automatically track the researcher's web activities, including created artifacts, in a variety of portals.

Herbert referenced several interesting projects that have inspired his thinking:
  • MIT's Solid Architecture - proposed set of conventions and tools for building decentralized social applications based on Linked Data principles
  • Sarven Capadisli's dokie.li - a decentralised article authoring, annotation, and social notification tool
  • Amy Guy's "Personal Web Observatory" - tracks daily activities, categorized and arranged visually with icons
These ideas could be used to develop a "Researcher Pod", which could combine an artifact tracker, an Event Store, and a communication platform that could be run on an institutional hosting platform along with an institutional archiving process.  These pods could be mobile and persistent so that researchers moving from one institution to another could take their pods with them.

Paul Evan Peters Lecture from Herbert Van de Sompel

Final Thoughts 

I greatly enjoyed attending my first CNI membership meeting. The talks were all high-quality, and I learned a great deal about some of the issues facing libraries and other institutional repositories.  Once the videos are posted, I encourage everyone to watch Cliff Lynch's plenary and Herbert Van de Sompel's closing talk. Both were excellent.

Because of the parallel sessions, I wasn't able to attend all of the briefings that I was interested in. After seeing some of the discussion on Twitter, I was particularly disappointed to have missed "Facing Slavery, Memory, and Reconciliation: The Research Library’s Role and Georgetown University’s Experience" presented by K. Matthew Dames (Georgetown) and Melissa Levine (Michigan).

Finally, I want to thank and acknowledge our funders, NEH, IMLS, and the Mellon Foundation.  Program officers from these organizations gave talks at CNI:

2017-12-22 edit: Embedded and added link to Cliff's plenary talk.
2018-01-03 edit: Embedded and added link to Herbert's award lecture.

No comments:

Post a Comment