Lovely view from the conference room
After a long hiatus while on research leave, I’m back out on the road getting in touch with the profession and taking notes. Access is both a favourite and an excellent reentry path given its warm and friendly environment. Within about an hour of being at Access, I had already worked up two proposals to take back to my organization to take up some new work, so my confidence in its ability to refresh and invigorate my thinking was rewarded.
One small note about these jottings: some of these papers had co-authors or co-creators who were not at Access. When I take notes at conferences, I drop any collaborators not actually speaking, not to slight them, but because part of the reason I keep these notes is to help me connect names to faces and names to specific projects. Having a visual memory is a great way for me not to lose track of all the threads in my head.
The Trouble with Access
Kim Christen – Washington State U
The part of this talk that made a strong impression on me was when she walked through items in the Plateau Peoples’ Web Portal and showed the differences between standard museum or archival description and rich description provided via community curation by the people whose culture created the objects. The difference is stark; rather than being superficially descriptive, they are narrative in nature, telling what an object signifies and speculating on various possible elements of its creation. Most importantly, the objects have a traditional knowledge license applied to them.
Mukurtu requires one to name all of the people involved, the who. Cultural protocols, the how, are equally important as they provide “granular levels of access.” One also needs categories.
Talked extensively about TK labels (traditional knowledge). Likened them to a “user agreement” for a Website, a request that users abide by certain rules when using the materials.
Visualizing Province-wide Public Library Transaction Data with Elastic Stack
Dale Storie – Scott Murray, Saskatchewan Information & Library Services Consortium
Talked about using Elastic Stack to visualize their log usage. One thing that popped for me was that they’ve created Slack alerts that let them know if the response time for a property is too long so that they can investigate. Layers on top of other monitoring they have.
Advocating for digital privacy: the centrality of public libraries in a uniquely 21st century struggle
Jonathon Hodge – Toronto Public Library
Asked how many public library people were in the audience. Healthy number of hands.
Asked who is going to stand up to those who violate privacy. Noted that tax systems are complicated, yet people navigate it. How? With the help of intermediaries such as accountants and tax preparers. Similarly, privacy is very complicated, and we need to step in as an intermediary in that environment.
TPL has taken steps to realize this role. They’ve enabled privacy protecting software on their public computers, e.g. Tor browser. Other methods: public education classes, events, podcasts, partnerships.
Don’t Be So Sensitive – A Data Security Journey
Hannah Rainey – North Carolina State U
NC State sets four levels of information, assigns them colours and titles, ranging from ultra-sensitive to not sensitive. Noted that privacy and security are not the same thing and that we need to understand the difference. Their project pursued two questions: what data do we (the libraries) have and where is it? Perhaps not surprisingly, she found a dearth of literature on libraries and their data security practices.
Applied a data audit framework from the U of Glasgow (HATII) with four main steps: planning audit, identifying/classifying assets, assessing management of data assets, and reporting and recommendations. She assumed the role of auditor, and applied their campus framework to library data.
Conducted 22 hour-long interviews to identify data and discover practice with it. Discovered that people wanted overarching, consistent guidelines. Ultimately she found that their data security practices aligned with campus policy, but also that core library data was not yet even included in those campus guidelines (circulation transactions, eresource statistics, etc.).
Excavating the 80s: Strategies for Restoring Digital Artifacts from the First Era of Personal Computing
John Durno – U of Victoria
Reviewed a number of 1990s articles about digital preservation that discussed format migration and emulation, noting that the former has become the default mode for many practices. Emulation is, however, quite viable. For his work with data from the 80s and 90s, he uses all of the options, often in combination.
Walked through a number of case studies. The first was text files written with AtariWriter. Found an Atari emulator, assuming that since it was a popular gaming platform that such would exist and that it would run AW. He said it wasn’t that challenging to find older software, although he noted that legally it’s something of a gray area to used abandoned software, although he suggested that research or educational uses are likely fairly safe. In this case, emulation allowed him to get to the text, which then allows him to do format migration.
Second example was WordStar. No conversion tool exists other than a flaky Word tool. Opened them using WordPerfect in an emulator. But what if you have thousands of them? Hand labour is tedious, file by file. Batch converters for WordStar, as he noted, are rare, actually non-existent. Scripted the conversion of files to ASCII, at least to do quick and basic conversion to get at the bulk of the content.
The third example was Telidon 699, related to the work he talked about at Access in Toronto in 2015, with regard to graphical works, i.e. artworks, created for screens. Telidon had specific hardware. Few made, nearly impossible to find. Found a Telidon 699 in a radio museum outside Vancouver, which also had a volunteer who knew how to run it and fix it. Ended up filming the images using a video camera to capture them, while blowing air across the Telidon with its case removed to keep it from blowing tubes.
Finished by noting that we will have to hold on to old hardware to be able to read disks that use drives for which ports and drivers are no longer available. We don’t all need to keep hardware, of course.
Opening the DAMS : Open Systems, Open Data, and Open Collaboration with Samvera at UVic
Dean Seeman, Lisa Goddard – U of Victoria
Samvera is an amalgam of multiple tools. She noted that Hyku is a more turnkey option for installing it. They migrated from CONTENTdm; it has served them well, but it doesn’t align with their strategic directions, particularly making their infrastructure more available for faculty research, i.e. make it available for hosting faculty-generated materials. They need it to be entirely Web-based, support multi-tenency, better tools for working with images (e.g. IIIF), and offer better integration with exhibits. They also wanted something with open development. Many other advantages, such as being linked open data friendly, with all of their assets addressable via a URI.
Wish I could have followed Dean’s comments about how they are working with the metadata better. I just don’t know enough about the details to take sensible and coherent notes.
The Way Leads to Pushmi-Pullyu, a Lightweight Approach to Managing Content Flow for Repository Preservation at UofA Libraries
Weiwei Shi, Shane Murnaghan, Matt Barnett – U of Alberta
Pushmi-Pullyu pulls content from their Fedora repo when a user makes a request via the interface. Why do they need it? Integration with Archivematica is in their future, but they need baseline preservation today. They also have IT security restrictions on access to their preservation storage directly via a user interface. So the application manages that flow while minimizing risk.
Use a paid service, Rollbar, to monitor their logs.
“No, we can’t just script it.”
Danielle Robichaud – U of Waterloo; Sara Allain – Artefactual
Sara outlined differences between archive and library data. Major difference: all archival descriptions are original work. There is no copy cataloguing in archives. Records are organic and interrelated; descriptive records may point to an object or a set of objects. There is also not enough time to describe everything, so item-level description is rare and expensive. The adoption of standards has been slow and uneven.
Danielle spoke about a concrete example of a collection of photographic negatives, which arrived in the archives with very Spartan metadata. Used an example of a descriptor to show that the description of an envelope had zero connection to the images in a sense that we would expect or understand. One cannot just take an envelope level description and iterate it across all of the objects in the envelope; to do so introduces error/confusion and also makes it looks like someone made a really obvious mistake.
Supporting Media-Intensive Digital Scholarship: The Development of a Streaming Media Repository at the University of Alberta Libraries
Sean Luyk, Weiwei Shi – U of Alberta
ERA A+V is a media hosting and streaming service based on Avalon. Other repositories don’t handle time-based media well, offering download rather than streaming. Media systems such as Kaltura et al., conversely, don’t work well for library purposes.
Deposits are mediated for now, largely due to file size limitations for uploads. They take requests via a Google form that is open to faculty and graduate students. Many core features in Avalon, including LTI (Learning Tool Integration) to the LMS. Their customizations include augmenting the batch metadata updating, adding Rollbar monitoring, modifying permissions, etc.
What’s not going so well? Some legal issues around DVD streaming. The LTI integration took some work.
ERA (Education & Research Archive) is currently just branding, i.e. the backends of their IR and ERA A+V are separate, but by using common branding they leave the door open to future integration.
The Academic Library Commons: Reimagining Collaborative Learning Spaces to Support the Scholarly Journey from Curiosity to Discovery to Publication
Michael Courtney, Angela Courtney – Indiana U Bloomington
While rethinking their spaces, they discovered that a lot of their computer workstations were not used, so they removed 75% of their fleet, but they replaced it with stronger technical and technology support that is integrated with other forms of support.
The UX of Online Help
Ruby Warren – U of Manitoba
They have created 55 online tutorials (video and text) and present them to users in clusters based on user category. They can show that they are used, but they wanted to know if they were effective, so ran a fairly large set of usability tests. They wanted to know if people could find them, navigate them, and if the language made sense. Also are conducting interviews, asking users to compare videos and also general questions about online learning.
What did they discover? Their terminology had issues. No clear links to talk to a person. Videos had non-intuitive titles. The help area didn’t have “help” in the title. Some videos were too long. Old, outdated content ruins everything. User group divisions are non-intuitive. They also discovered that top nav does not work, ever. They look down pages, not up.
What attributes did videos they like have? Clear, unobstructed views of someone demonstrating what to do. Slower pacing and clarity. Small tasks, short videos (under three minutes). Also appreciate consistent voice across videos: “the voice of the library.” What about text? Text needs to be broken up with images. Bolding, headings, dynamism. Short.
What’s next? They are going to daisy-chain their videos by linking them at the end.
Can Link: A Linked Data Project for Canadian Theses / CanaLien: un projet de données liées pour les thèses canadiennes
Rob Warren – Carleton U; Sharon Farnel – U of Alberta
A number of Canadian institutions contributed about 5,000 theses records from their library systems. These arrived in various formats: MARC, CSV, XLS). They chose to create their own ontology, but they did some research that showed that people and computers “aren’t psychic.” So they created their own ontology that reuses a lot of other vocabularies, which restricts what they can do. They are only trying to communicate what a thesis is and what it relates to. Ended up using bits of various ontologies to describe specific aspects. He showed this graphically on slides and it really made good sense visually.
The core functionality they sought to deliver includes a Web tool for transforming MARC data. Others: Github integration (missing URIs can be added through Github tickets), sparql queries, basic UI queries, tweets for newly added theses.
Speaking about the why of the project, Sharon noted the high cumulative value of theses in terms of the investment in their creation. A lot of knowledge resides in them, but they are often read when submitted and then locked away and forgotten. They want to surface the intellectual contributions represented in these documents.
At the end of their talk, they noted that they are looking for theses records from anyone. We should consider adding ours from McMaster.
User Experience From a Technical Services Point of View
Shelley Gullikson – Carleton U
One of their motivations was that technical services and user experience do not go hand in hand. A technical services supervisor, however, wanted to find ways to make their work more visible and relevant. Their method was to run usability tests, where they asked students to search for information to complete and assignment, letting them know that they didn’t have to use library resources if that’s not typically what they would do. Most subjects searched for 20-30 minutes. Subjects were 10 each, graduate and undergraduate students.
Major themes:
- Overwhelming use of the single search box (15 – Summon in their case). Some used Google Scholar (12). Those two came out on top. Catalogue used by two undergraduates.
- Popularity of the “Get it” button. Students quite verbal about it.
- Metadata. Students searched keywords, but examined titles for those words. Publication date often matters. They like to scan abstracts, so monograph records confuse them and they move past them. Few browse searches (author, generally).
- They want fast and easy access. Rarely go past top ten results, regardless of how long the list might be. They skip hard-to-access content, taking the next best available thing.
Feedback from staff was somewhat disheartening. Staff assume they don’t use the catalogue because they can’t find it. Suggested that students need to be trained to use it, or given a booklet that explains it to them. [Groaning and laughter in the audience.] More positively, they saw that they need to push vendors to supply accurate data and to resolve issues. Began to see that if that is what students use, their job is to make it better.
Lightning talk tidbits (I gave a lightning talk, so my attention was a bit all over the place while others were speaking as I prepped my talk.)
Using Slack to create a in/out board, a virtual version of the white board where people record their status or location. Created a Web interface that pulls it from Slack via their API. (U Sask)
Amy (Thompson Rivers) urged conferences to consider offering childcare. Yes! She noted that this issue impacts women more than men, generally speaking, as well as having more impact on people in precarious employment situations. As she noted the how is complicated, but it is a solvable issue. Even if we can’t solve the childcare issue, we can work on messaging, talking about what we allow and what we need at conferences.
Sara (Artefactual) gave a pitch for using Gherkin for documentation.
IIIF You Can Dream—And Not Make Dreams Your Master: IIIF in the Real World
Peter Binkley – U of Alberta
IIIF is aimed at making access to images better for scholars, supporting interoperability between repositories via APIs, and creating a world-class user experience in viewing, comparing, manipulating, and annotating images. There is a family of APIs: Image API, Presentation API, Search API, and Authentication API. Peter appreciates and notes the care with which these APIs were thought out and implemented.
Cantaloupes and Canvases: Adopting IIIF Specifications
Sascha Adler – Canadiana
IIIF allows Canadiana to consider and add new features. They’ve been working at this for about six months. Cantaloupe is a tool for derivative image generation and works with the IIIF Image API.
Kenny: A Web Based Tool for EZproxy Authentication Management
Darryl Friesen – U of Saskatchewan
Created a tool to help tackle illegal downloading and the need to override a user’s access level, such as when a student becomes an alumna/alumnus. EZproxy does have some functionality that handles this, such as the RejectIP directive, as well as baduser lists in user.txt. For staff to do this, they must have access rights to this system, and EZproxy has to be restarted to put changes into effect, so this limits who can have access and do the work. There is also no record or tracking of who did what when you use these tools, i.e. no logging of user actions or their justifications.
Kenny (Logins, get it?) is intended to enable doing this work via a Web-based tool. Having sys admins do all of this work means it is slower than if it’s handled by frontline staff. With Kenny, users do not need access to EZproxy on the server. EZproxy does not need to be restarted because it integrates into the authentication process. It maintains a history of actions: who was blocked, and when, why, and by whom. Last, it can also terminate current EZproxy sessions (if configured properly, he added).
How is it done? EZproxy supports two custom auth methods, CGI and external. Kenny uses external authentication. EZproxy sends the username and IP address to Kenny, it checks for overrides, and just returns OK or deny.
Did a live demo. Really great tool, very basic interface that is very easy to grasp at a glance. Coming soon to github (link doesn’t work yet).
Sadly, Nora Young had some flight issues so wasn’t able to deliver her closing keynote. Really sad to have missed that.
