Including several very important topics were covered:
The weather
Water is dangerous but vital to life
What exactly is a prank? Did cave people play pranks? Apparently crows and apes play pranks on each other, so it is likely a universal trait??
We move on to more serious topics...eventually.
05:44
Housekeeping
We agree to cancel on August 1st and August 22nd.
Tara swears she will get to the directory project "today" and then failed to do so, but she will get to this.
Wolfcon reminder (07:30), virtual registration is free! Check out the sessions and sign up even if you are not going in person!
09:00
Loading users at ByWater
Ian explains how ByWater loads users:
Partners load a file to FTP, workflow processes pick up the files, the files are parsed and mapped, expiration dates are set (tied to patron group, generally), and then they use the User APIs to load users. To upload each user, they check to see if the user record already exists. If it does, they fetch the user record and merge the information so that nothing is lost. If it does not exist, they build a new record and create it in FOLIO. With that process, they update between 2,000 and 13,000 patrons a night, every night, forever.
How are the records formatted? (11:30)
Ian says that most of the time the library sends delimited files, with the content depending on the library. One library sends JSON. It is bespoke processing for each partner. Ian notes that he hopes someday he will be able to set up a webhook that will update FOLIO upon update/change from the institution's system, but no one has developed this functionality yet.
For staff members who are in FOLIO with permissions, how are those handled? (13:22)
Permissions are not managed through this process--they are not touched. Permissions records are separate.Â
What is the hold-up on creating an events-based update system? (14:37)
Ian notes that the source systems do not have this available functionality. Where it does exist, libraries often lack the bandwidth to do this programming on their side. The practice of loading users through files is well-established, common, and satisfactory at many libraries.
Julie notes that FTP sounds kind of "old school" but it's actually very solid as a way to do data transfer. Tod reminds us that folks sometimes say FTP and mean SFTP these days, and there is a big difference in terms of security.
Chicago's practice is similar to ByWater's. Chicago gets a nightly feed from their identity management infrastructure. It is a CSV file (or similar).
How long does it take for a very large file (over 10,000 patrons) to load? (18:00)
Ian believes it is under 5 minutes for a 10,000 patron load.
Does anyone keep archives of files? (19:28)
Bywater keeps two weeks by policy, but since the data is received so frequently, the files are almost immediately stale.
20:12
Cornell User Loads
Thomas explains Cornell's process.
Thomas notes that Cornell hits the user APIs directly, which he thought was less common. The library doesn't handle the patron loads directly--Cornell IT department handles it. IT pulls information from PeopleSoft and the Bursar's office. IT also pulls out patron groups, departments, etc from FOLIO as well. They push it into FOLIO on a nightly basis. Typically, these are updates. They only touch a selection of fields.
Thomas and the library get daily emails from the system with notifications of errors. If there are errors, Thomas has to fix them. Typically, the issue is duplicate barcodes.
Darsi asks about what happens with inactive users. (22:47)
Thomas notes that they don't use expiration dates, but they do make users inactive. Those records are not deleted, they are retained as inactive.
23:31
APIs vs mod-user-import
Maura asks about "hitting APIs directly." Does Cornell use mod-user-import, or the user APIs?
Thomas says that they forgo mod-user-import and are just hitting the user APIs.
Thomas touched base with the person responsible for this at Cornell about switching to mod-user-import, but that change was not desired. The original choice may have been due to functionality available when Cornell went live.
One issue that Cornell saw with this approach was that using the user APIs directly does not create a corresponding permissions user, which can cause significant errors in FOLIO--every user must have an empty permissions user, even if they do not and will not ever have permissions.
Maura asks if people use mod-user-import, or user the user APIs directly? (25:30)
Darsi believes that Stanford uses mod-user-import.
Index Data mainly uses mod-user-import.
Thomas says that on slack, it seems like the preferred way is to use mod-user-import, and that is what the majority are using. It's not clear if that's perception or reality.
You cannot upload a CSV of users in FOLIO currently. All of these processes are happening via API. This functionality may come someday to Bulk Edit, but there is no import for users. Bulk edit *can* currently modify users. But this is on the roadmap.
29:28
Public patrons and special cases
How are folks handling public patrons, special categories of users, etc who are not in the IT system, but are in a large enough number that manual creation would be painful?
Even for large numbers, libraries tend to create these manually. 😞
What happens with patrons who belong to multiple groups? For example, if a patron is a faculty member and a staff member, what do libraries do? (31:52)
Bywater has libraries take it up with IT--the file should be delivered as libraries want the patrons loaded.
At Stanford, Darsi says that they harvest the records and give patrons whichever status has "better" privileges.
Cornell does the same. (33:15) Voyager did allow multiple patron groups, which was helpful in some cases, but also sometimes caused confusion.
We note (34:07) that in some cases it's not easy to say which patron group is truly better, as one group may have longer loan times, but might not have access to certain materials (for example).
Tara asks about consortial set ups (34:19), when a person might have identities at multiple institutions within the same tenant.
We think it's likely that patrons will simply have two accounts, but we're not sure where this development is.
Any major mishaps ever happen with loading users? (36:04)
Not really...Cornell's process stops the load when it runs into a duplicate user.
Why is username required for importing, even for patrons who will not log in? (37:42)
Thomas notes that it is for linking with other systems--their username is their CNET ID.
Action Items
Tara Barnett  will restart the directory project and reach out to Darsi and Ian