This meeting is open to anyone who would like to attend; it will provide an opportunity for members of the Fedora community to connect, share information, and provide updates on local projects and initiatives. This meeting will be based around American timezones, but anyone is welcome to attend. We are also planning a future meeting that will be based around European timezones.

When

August 18-19

Where

Online via Zoom

Meeting Notes

• Collaborative Notes Document

Agenda/Presentations

August 18 - Presentations / Discussions

All times are Eastern Daylight (US)

August 19 - Presentations / Discussions

All times are Eastern Daylight (US)

Notes - Day 1

12:10pm-12:30pm Fedora Program and Community Update, David Wilcox, LYRASIS

• Slides: https://docs.google.com/presentation/d/1scgMTSc6ZRXrC2wkmXxHZejuve1De9Isf9cohSrmPpI 
• Fedora Migration Paths and Tools IMLS Grant, 2020-2022
• Tooling to support migration from F3 (unsupported) to F6
• Ultimately, goal is to have everyone on a supported version of Fedora
• Andrew Woods will talk about Fedora 6 and Oxford Common File Layout tomorrow

12:30pm-12:50pm Islandora Updates, Melissa Anez, Islandora Foundation

• Danny Lamb, technical lead for Islandora, giving an overview of software - Drupal + Fedora. Mark Leggott at UPEI wanted a front end for Fedora - Drupal version (originally idea was many front ends) is the one that stuck. But there’s a lot more to Islandora than just Drupal and Fedora…
• Original analogy (Islandora 7) of a cheeseburger - Drupal and Fedora buns, Islandora the stuff inside. 
• Islandora 8 - everything now horizontally scales, can be run on individual servers, completely Dockerized. Metaphor is now a Bento box (equally delicious, neatly organized).
• Lots of features available, with Drupal in the center. Not interacting with Fedora directly, but everything makes its way into Fedora. Drupal talks to all the different components (Solr, Tesseract, ImageMagick). Drupal puts requests in a queue and uses microservices, work is not all being done on the Drupal webserver.
• Props to Bethany Seeger for creating a full diagram of Islandora 8 (included in slides)
• Fedora used for: storage, fixity checks, Memento versioning, audit rail, Oxford Common File Layout (OCFL!)
• Flysystem - an abstraction layer for systems written in PHP. Fedora is a filesystem for Drupal, similar to how you would use Drupal and connect to AWS, all abstracted in the same shape.
• Melissa talking about Islandora community - active culture of peer support, given there are only 2 full time staff, makes people using software working together critical.
• Know of 320 users - though there are probably more that aren’t known about. 1 on every continent save for Antarctica
• Complex web of interactions in the community - need wide range of volunteers, developers, leaders, content experts, etc. And members - financial contributions are also extremely important.
• Kent State - first one using Islandora 8. Also creating OAI-PMH endpoint. Also using Open Journal Systems.
• Canterbury Stories in New Zealand - using Islandora 8 - making use of Drupal accessibility tools - fully compliant with New Zealand accessibility standards, including present site in both English and Maori.
• LADI (Latin American Digital Initiatives) using Islandora 8 - multilingual content and interface.
• Archives Central in New Zealand - migrated 200K records, using a complex metadata model, Records in Context Conceptual Model
• UPEI - deploying research data management in Islandora 8.
• Q&A notes
• UPEI work is helping with data model for multi-type works, e.g. oral histories + transcript, video file + transcript
• To play around with Islandora: https://github.com/Islandora-Devops/isle-dc
• Single sign on with LDAP & CAS? Will work if there is a Drupal module for that
• Islandora 8 to facilitate multiple instances? Multi-tenancy, single Fedora with multiple Drupals talking to it. They’re working on that now.
• Having a single Fedora repo, can have multiple “heads” on it, pulling from the same Fedora, some of which can be Islandaora

1:10pm - 1:30pm Member Update for the University of Maryland, Josh Westgard, University of Maryland

• Locally developed repository built on Fedora 4
• Focused on command line tools for batch import/export
• Full text search with highlighted search term
• UI for item metadata editing wasn’t top priority, but they did create one for editing a single item 
• User front end is CMS is called Bloomreach, formerly Hippo

1:30pm - 1:50pm Lightning Talks:

• NLM Repository Migration Considerations, Doron Shalvi, National Library of Medicine

• Want to have permalinks (URIs) for digital objects
• Q&A
1. More about vocabularies? 
• NLM has URIs for its subject headings, that we maintain. See https://id.nlm.nih.gov/mesh/ 
• Mange authorities in ILS, export for use in other systems. Currently use Voyager, migrating to Alma
2. Question about moving your metadata to RDF: How much do you plan to convert to RDF?  Does that include all your descriptive metadata?  MODS, DC, METS, etc?
• To some degree still TBD. Mostly use custom descriptive metadata scheme similar to MODS; METS used for packaging, DC not heavily used (simplified form for public to download)
• At first, will probably migrate over to Fedora as XML, then convert selectively to RDF
• (See a few more notes below in “Closing Remarks & Discussion” section)

• Using Riprap to perform periodic fixity validation on your Fedora repository, Mark Jordan, Simon Fraser University

• Integrated with Islandora to check/see fixity at the object level and repository level
• Another goal of this fixity check is that you can see if files are still where they’re supposed to be
• Could be reconfigured to integrate with Fedora, would need to write a specific plugin for Fedora.
• Q&A
• Is riprap multi-threaded for fixity checks?  Have you done performance tests?
• Haven’t tested this, but certainly can. Have to watch transactions in the database
• What is your goal for frequency of checking?  Have you ever encountered any failures?
• Can go through a cycle once every few months, but that frequency may not be necessary for others
• Think about what makes sense for your org (and storage infrastructure), should be part of a preservation policy
• Failures -- very rare occurrence, did happen a couple of times. 1) Bit in a tiff flipped, 2) Something disappeared; copying issue

1:50pm - 2:20pm Closing Remarks and Discussion

• RE converting metadata to RDF -- when you have a detailed scheme, can have explosion of data. 
• Once consideration is to convert a subset to RDF - identifying core fields, relationships
• Don’t necessarily need to convert in order to use Fedora, can use XML
• How some folks are doing full text search: store XML, consumed into Solr

Notes - Day 2

12:10pm-12:30pm Re-imagining Legacy Digital Content in (the) Spotlight, Jennifer Gilbert, National Library of Medicine

• Migrating “Profiles in Science” digital project
• Spotlight is UI for presentation, content pulled from Fedora repo (currently on F3)
• AMD = Archival Metadata
• Next steps include adding new content - everything so far was existing content that was migrated
• Q&A
• What roles were involved in this work? For the migration project, developers. The intent is in the future to have content owners build exhibits.
• More about annotations
• Annotations are mostly in the Lederberg collection. Not in production instance yet, but will be soon.

12:30pm -12:50pm Fedora and OCFL, Andrew Woods, LYRASIS

• New persistence layer based on OCFL (Oxford Common File Layout)
• Archival groups can have multiple hierarchies and nested files
• Q&A
• What's the status of export from F4 and import to F6?
• There’s an export tool for F4, finishing work on transformation tool, then rebuild in F6
• How will Fedora3 objects be represented? Is there any retention of the old Fedora PIDs?
• Pretty clean mapping from F3 to F6, PIDs are retained in the metadata (in F3 PIDs are in URIs as well). AW would like to see retaining URIs from F3, but haven’t worked through fully
• How will versioning play out in the objects?
•  May not want every change to represent a new version. How will this be managed so that it doesn’t get unwieldy? There is a config property where you can turn off auto-versioning, user says when to create a new version (there’s an API user request to create versioning on demand)
• If you don’t use auto-versioning, there’s an extension called mutable head that Fedora employs that maintains incremental changes -- all in OCFL in a specified way
• Will the original audit trails be maintained at all?
• Audit data gets migrated over

1:00pm-1:20pm UT Libraries DAMS Architecture, Tori Brown, University of Texas at Austin

• Digital collections frontend: https://collections.lib.utexas.edu/“Download metadata” is the IIIF manifest

1:20pm - 1:50pm Fedora as a Platform for IIIF and Solr Discovery, James Creel, Texas A&M University

• SAGE will aggregate from multiple Solr cores

2:10pm - 2:30pm BCDAMS Digital Collections Ecosystem Updates, Anne Washington, University of Houston

2:30pm - 2:50pm UT Libraries Exhibitions, Larry Yang, University of Texas at Austin

2:50pm - 3:20pm Closing Remarks and Discussion