Access 2016

These are my live conference notes of the ACCESS 2016 conference.

Note that these are being written live which means I miss all sorts of things as random distraction cross my screen. Please email if there is stuff I should correct.

Thursday, October 6th, 2016

George Duimovich: Going Dutch - Introducing ARTUDIS - A New Repository Platform

Duimovich gave us an introduction to ARTUDIS from Erasmus University, Rotterdam. He talked about how they are using it at Carleton University. He talked about how this helped them implement their Open Access strategy. Their focus is on identifying and promoting OA works. There is a lot of micro-publishing by academics that doesn't get preserved. Project web sites, blog essays, and so on. ARTUDIS is built on the Google Cloud platform which has pluses and minuses. The cloud allows a small team to focus on their application not the back-end infrastructure.

While they don't have big data, they do have a lot of dynamic log data. Persons, organizations and collections can also have their own dashboards.

He then gave a demo of their Institutional Repository implemented with ARTUDIS. He showed adding a publication and so on. What was really impressive was how much work they put into author data for their Carleton faculty. He then showed some dashboards for faceted browsing of information about downloads.

Michael Groenendyk: Searching in 3D: The Availability and Discoverability of 3D Models on the Internet

Groenendyk talked about 3D models and how they can integrated into libraries. He talked about haptic tech that lets you feel a 3D model and other related technologies including cultural scanning. He then talked about the history of 3D from VRML, early virtual museums, the digital Michaelangelo David, and the Georgia Institute of Technology Large Geometric Model Archive. In 2005 there is an emerging CD marketplace, especially for CAD models.

There is an important distinction between CAD models and animation models. CAD models are utilitarian and for engineers. Animation models are more for games.

The 3D era dates to about 2000. For a long time it was very expensive. The high range printers cost $200,000 or more. Then projects like RepRap came along that made affordable printers available. MakerBot made Thingiverse available where models are available, but there is lots of poor quality models. Now we have collections like My Mini Factory which carried professional models (for money.)

He pointed out that most buildings start as CAD models, but no one is archiving them. The Montreal Museum of Fine Art has a design collection and they are trying to collect the models. Animated movies and games are another area where fans are trying to collect and preserve models.

Groenendyk has put together a database of available models and organized it by discipline. There are a number of historical recreations and biological models. (And engineering and architectural.) Most models are free now.

Stratasys now owns MakerBot (and Thingiverse) and grabCAD. This has led to a backlash as Stratasys seems to be patenting user contributed content.

There is a lot 3D scanning now, Google Cultural Institute, Smithsonian 3D, and Emerald Cultural Institute.

There are a lot of challenges around preserving 3D data. Changing file formats, large files, technologies changing, lack of cuture of preservation ...

The home market seems to be over. Most of the low end printers bought were bought by libraries, faculty not home users. Now these printers gather dust. We need real applications after the waving octopus. Libraries have been focused on delivery, but not content. He doesn't think libraries should be running printers. Other units can run makerspaces better. Libraries can bring people in using 3D, but may not be set up for long term provision.

He closed with a plug for ChessPiece here in Fredricton.

During questions he mentioned why he didn't think open standards would emerge soon as most people are using high-end applications to make games/animations.

John Fink, Myron Groover: Splitting the AtoM: Implementing Access to Memory at McMaster University

They started by talking about the horror of web pages for their archival descriptions so they had to have students do it by hand. "Many hands make blight work."

They then talked about AtoM and how they use it.

They talked about problems with versioning and de-duplication. There is a known bug that generates false subject headings that take a lot of work in the SQL.

They then talked about MPLP (More Product Less Process). They run things through quickly which means that there are errors.

As for standards they taken a stripped down approach to RAD compliance. They talked about how they have a flat natural language taxonomy. They have also had to revisit a lot of their descriptions and update them.

Ignite Talks

Tim Ribaric talked about Twitter Bot that is library grievances. See https://twitter.com/lis_grievances . You can send suggestions.

J Jack Unrau talked about "Transmitting in Cleartext: Digital Privacy Education for Reluctant Technologists". He talked about privacy and how when we teach it we teach people to be obscure. Why is our net world made for people with secret arcane knowledge? It makes us techies feel good, but that isn't enough. Why should people need to take these privacy classes? Users need tools for default privacy and libraries need to support that.

Naomi Eichenlaub talked about Making MARC Actionable. Connecting text in records to URIs so that computers can do something with the information. They did a small experiment at Ryerson.

Trina Grover. You Made that Yourself?! Trina and others are running a knitting group called Knit Happens. Knitting is making, programming, coding. It has its own notation, and manipulates yarn. The LilyPad is an electronic kit that works with embroidery. Trina showed a punched card knitting machine. They combine knitting with Tinker CAD made things like handles.

Gillian Byrne talked about If 'Libraries are Software', What Does That Look Like?. She talked about what it would mean to think of the library as software. Cataloguers become Metadata Infrastructure specialists. We are now in a post-scarcity era. Subject Liaisons become Search Algorithm support.

Wiktor Rzeczkowski talked about Easing the Collaboration Curse for Access. He talked about different types of collaborators including very effective ones and bottleneckers (who want to be included in everything.) Technology can improve collaborations.

Sam Popowich talked about Clojure for the Perplexed. Clojure is a functional programming environment. In FP everything is a function. You don't have side effects unless you want them. Closure is based on Lisp. It is a new form of Lisp.

Andrew Nagy talked about PaaS (Platform as a Service). The library market is seeing aggregation, with fewer companies. We need open platforms for new open apps to be exchanged. Amazon S3 is a platform. Folio is an open source library services platform. The value is in the apps that you create and exchange.

Laura Wrubel talked about "Need Help with Your Code? Providing a Programming Consultation Service at the Library." Students come to them to figure out how to code for their projects. There is a gap on campus for coding support for humanities and social scientists. So they started a Programming Consultation service.

Pamela Carson and Janice Kung: Raspberry Pi Prototype Project: Measuring and Displaying Noise Levels in an Academic Library

Carson and Kung talked about setting up a sound measuring system developed with a Raspberry Pi. Noise is a real problem in libraries. Users have commented on excess noise. You can get cell phone apps to measure decibel levels. Over 90 is an issue for community noise. Classrooms should have less than 35. That said, subjective measures can be a better way for libraries.

Libraries have experimented with furniture and space to see how that affects noise. Other methods include staff monitoring, humorous cards, policies, signs and so on. At McGill they installed an electronic tool that would light a sign when noise went over a certain amount. Alas interventions don't have an effect and the interventions bothered students.

Research shows that for many acoustic environment choice is a physiologic need and a need that varies among people. So Carson and Kung what to display noise levels for different locations so people can go to the places with the levels of noise they want. They already display some live data. This would fit in with that.

Their prototype uses Arduinos to measure the sound that then communicate with a Raspberry Pi that is the server. They talked about how they solved all the problems on the way. Someone asked about security and privacy issues.

Jason Clark and Krista Godfrey: Addressing the “Uncanny Valley” of Search UX - Anticipatory Design, Privacy, and Tolerance for Systems that Speculate About You

Clark and Godfrey talked about how important search is and how there is more to search than just the words. There is a lot of context. Libraries need to think about how they interpret search queries more broadly. There may be a lot of tacit, inferred aspects as in the person is making the query from an iPhone on a highway.

What cues do you have? Word counts, stop words, cues from machine (there is lots of data in the HTTP Headers and Global Server variables. These can include language, time and location, referring URL, and so on. He then talked about Javascript information. You can also

See https://amiunique.org/ to see how much data you shed.

Their goal is to use the data to give much better results. They can give results based on time of year, location, and so on. There is, however, a tension between providing better results and the uncanny valley. Think of Clippy and other creepy customizations.

Then they talked about whether users care about privacy. Students don't seem to mind. According to Pew Research Center (2016) it depends. So how do you test an attitude about privacy? They are now planning testing on scripts and tool. They are using scenarios for testing.

Angie Fullington: Delivering Dynamic Structured Content Through User-Centered Taxonomies

Fullington talked about the web team they are part of. They are very user-centered and try to facilitate a cross-channel experience. They follow an interative design process. They build taxonomies for content types.

They work with stories for the different stakeholders. There is a library story and a user story. The library story needs to show that they understand the user story. How to organize all the neat stuff they do so users/students can see them and understand them. Long lists of services don't cut it. Users don't ask with "what is the full range of services you provide...." They start with, "I was told you can lend me a camera for an assignment tomorrow." Users tend to think of products not services. Libraries provide spaces, software, devices, expertise, and learning objects. You can think about what users make that the library can help with. They now create "Do" pages where people can see what the library can "do" for them around something they want to make.

Friday, October 7th, 2016

Mike Nason and Jacob Sanford: Starting From Scratch – The Technical, Logistical, and Political Challenges of (Re)Building an Institutional Repository

Nason and Sanford talked about building a new repository. They had a DSpace Repository and shifted to Islandora to create UNB Scholar. They commented on how you can't ever escaped your past. They talked about the problem of needing content in order to get content. One of the challenges they face is marketing. How do you take a photo of a digital repository? You put a book on a computer.

They talked about what they saw as their mission and how that differed from what their colleagues thought.

They talked about how the Tri-Council + call for stewardship of data was a game changer for libraries building repositories. Now their repositories are a solution to a broadly recognized problem.

Turnkey vs Latchkey - they talked about how no solution is turnkey. You have to adapt the code you get to suit your environment. A scholar-specific model needs to be decoupled from scholars. I didn't get this point, but it sounds important. Often services are organized around faculty (scholars) because we are powerful in universities, but perhaps we aren't the best way to organize things for the larger community.

Then they talked about mitigating technical debt. The architects need to keep a focus on sustainability which means reducing technical debt:

• Acknowledge the spectre of debt
• Plan, plan, plan
• Stay tight with contributed code
• Use best practices when customizing
• Run wild with tests
• Continuously clean up

Then they turned to lessons learned:

• Sometimes you have to say "nope"
• Repositories are never finished

In the question period there was discussion about forking code and the problems with that.

Hackfest Reports

The reports made it sound like they had a lot of fun, though many of the groups hit roadblocks.

Geoffrey Rockwell: The Ethics of Digitization

I gave the closing keynote and talked about ideas I have been developing with Bettina Berendt around the ethics of digitization.