Many of us in the US, and around the world for that matter, are now sitting at home working remotely trying to maintain some semblance of normalcy during quarantine for COVID-19.
I’m going to write up a few of the things that we have been working on her at the UNT Libraries to try and provide as many of the library employees, including student employees, with activities that they can do remotely.
On March 13th it became quite clear that there was going to be a large number of folks from the library needing to work remotely, many of us have activities that we can do remotely, and some of us even prefer to work remotely when we have the opportunity. There are however a large group of people at the library that don’t have jobs that directly transfer to working remotely, or in the situation where there aren’t students on campus to directly serve, the work that they would be doing isn’t available for them.
We wanted to provide an option for these individuals to create metadata for the UNT Libraries Digital Collections if they were interested in doing so. Additionally, this would give supervisors some meaningful activity that could be verified in the event that we had to move the workforce remotely because of this pandemic.
What we needed
There were a few things that we needed to have in place in order to move people online to edit metadata records.
- Web-based system for editing metadata
- Clear instructions and guides
- A way of identifying work to be done
- A way of coordinating activity
- A way of verifying/tracking activity for timekeeping
Web-based editing system.
There were a number of things that we have in place that allow us to move a large number of people into the metadata creation workflow. First, we have been using a web-based metadata editing system for the entire time that we have had our digital collections. Here at UNT Libraries we creatively call this system Edit. This system is built around the metadata format that we have been using locally called UNTL. We can add a user in the system, assign them to a subset of the collection, usually based on collections, and give them permission to begin editing metadata. We have a system in place to register new users in batches and then invite them to join the editing system.
In addition to our own work here in the libraries, we have made use of this metadata editing system for a number of metadata classes of library school students to give them real-world experience writing metadata records in a production system. Because of these previous experiences, the task of creating 60-70 new accounts for users in the library wasn’t too daunting.
For setting up user accounts we make use of a Django app that we wrote a few years ago called django_invite (https://github.com/unt-libraries/django-invite). Generally, the workflow we follow for new accounts is that we are given one or more new users who need accounts. The information we need is just a name, an email address, and the scope of the collection they need access to. We can enter multiple names at once if they have the same permissions. The permissions for this app are based on the standard Django permission and group concepts.
Once you submit users, the system sends out an invitation email for the user to complete the registration process by picking a name. We are able to see who has established their account and if needed, resend the invitation email.
This process makes it fairly straightforward to get people set up in the system.
As I said, we have been using a web-based editing system for the UNT Libraries Digital Collections for over a decade now. The editing system (Edit) starts a user with a view of all of the records that they can access based on their permissions. We call this view the “Search Dashboard”.
From here a user can search the metadata for a record, sort results, and limit their result sets based on any of the facets on the left-hand side of the screen. For those interested, the facets include.
- Resource Type
- Visibility (hidden or not)
- Date Validity (valid EDTF dates)
- My Edits (records you have edited)
- Record Completeness
- Location data (with/without placenames, geocodes, or bounding boxes)
- Recently Edited Records (Last 24h, 48h, 7d, 30d, 90d, 180d, 365d)
From there a user gets back their results where they can be further refined by sorting.
The current sort options include:
- Title (default)
- Date Added (newest/oldest
- Creation Date (newest/oldest)
- Date Last Modified (newest/oldest)
- ARK Identifier (lowest/highest)
- Completeness (lowest/highest)
They are also able to see basic information about the object including a thumbnail, the system, collections, and partner in which the item belongs. They can see the accession date and the last date that the item was edited. Finally, they can see a visibility flag, green for visible to the public and a red check for not visible.
They can choose to go to one of two places for a record, the edit view or the summary view. I will start with an overview of the summary view.
The Summary view provides an overview of the record including a compact version of the record itself. We provide an Edit Timeline to get a better sense of when the item has been edited, and when it became available online. Additionally, it has links and other information that are helpful for metadata creators. I will walk through a few of those links now. First up is the View Item screen.
This view is for the metadata creator to interact with the object itself, they are able to see all of the pages of the item and zoom in to look at the details. For audio and video, they have the ability to view the item in a player as well as download the media files as needed.
We also have the ability to see the history of edits that have occurred for a record. This history page presents information about who, when, and a high-level overview of what has happened to the item over time. You might notice a number of edits for this record. One of the things we have noticed in our metadata editing practice is that we tend to take a “column” approach to the editing of records instead of a “row” approach. We will find an issue, maybe an incorrectly formatted name, and fix all of those instances in the system. This results in many edits per record but allows editors to focus on a single task. As you can see, all of the record edits are versioned so it is possible to go back and view what was changed and by whom.
The final view is the metadata editor itself. I’ve mentioned this in a number of other posts over time so I won’t go into too much detail here. Basically all of the work of editing gets done here. Users can add, subtract, reorder, and edit elements. Most elements have qualifiers to designate the type of element being used such as the Main Title, Added Title, Serial Title, or Series Title for the title element. Some element such as creator, contributor, and publisher have a type ahead that pull from our name authority system (UNT Names) and include information about the type of name (personal/organization), the role (author, photographer, editor) and an info field for other bits of info about the agent. All dropdown values are pulled from a centralized vocabulary management system.
Some of the fields have popup modals for controlled vocabularies, picking locations from a map, or assigning bounding box information to an object. From here users can mark an object as hidden or visible, and publish the record in order to save it back into the system.
As I mentioned above there are a number of other components that are proving to be important as we move a large number of works into our metadata system. In the past week, we have created over 70 new accounts for students and staff in the library so that they can begin to incorporate metadata editing into their work. In the next few posts, I will go over how we are attempting to manage who is doing what and how we are providing social and technical infrastructure to help managers keep track of what is going on with the folks they are responsible for.
If you have questions or comments about this post, please let me know via Twitter.