...
- Manually linked MARC Authorities and MARC Bibs, and the impact on Data import (Khalilah - 55 mins)
- If you have time before the meeting, please add your comments and questions to the following wiki pages:
- Discussion
- Quick overview of the mockups for manual linking (see slide deck attached below)
- Then discussion of the the various cases documented in the above requirement docs
- Questions and comments
- Derive - often derive from print to create an e-resource record - would be good to have the links for any headings in the original record carry over to the derived record (Christie, Jacquie)
- Per Josh, Mich State uses on a limited basis - A would probably be best
- Concern about system load when importing updated authority records that trigger lots of impact to headings in linked bib records; comment that some large authority loads take days to complete in their current system (Jennifer)
- Concern about a linked bib heading/authority record heading just matching on $0 or full text string; would absolutely want to have control over any automated linking in the system, plus rich reporting on what did or didn't happen (Christie)
- Permissions:
- Manual linking will be separate from regular quickMARC edit/derive permission
- Automated linking will also have a separate permission
- Overlaying a MARC Bib via Inventory Single Record Import
- Maybe bringing in a new copy of the record from OCLC - if the values match, keep the links; if the values don't match (or don't existing in the new version of the record), then no links
- Matching: aim for $0 match if present in incoming record; if no $0, then exact text string match as secondary option
- Need to discuss further when we get to auto-linking
- Maybe until auto-linking, only preserve the links for existing fields if the existing field and the incoming field have the same $0; do not try to do any text-string matching; then make it more sophisticated (perhaps) when we get to auto-linking
- Maybe bringing in a new copy of the record from OCLC - if the values match, keep the links; if the values don't match (or don't existing in the new version of the record), then no links
- Data import updates a MARC Bib record, and a linked heading in that record conflicts with the authorized version of the heading in the authority record (scenario 7 in the bib record interactions doc linked above)
- see updated doc - several scenarios
- Might it be helpful to add other examples to tease out different handling preferences? e.g. in Scenario 7, we had
- 100 Angelou with $0 being updated to Angeloo with $0 (typo, different from authorized form, report - still TBD - update the Bib 100 or not?)
- 100 Angelou being updated to Twain, incoming field had different $0 (unlink, update, don't report)
- 100 Angelou with $0 being updated to Angelou without $0 (unlink, update, don't report)
- 100 Angelou with $0 being updated to 700 Angelou with $0 (keep linked, update, don't report)
- Derive - often derive from print to create an e-resource record - would be good to have the links for any headings in the original record carry over to the derived record (Christie, Jacquie)
- Mockups for manual linking (see slide 11 especially for how the links display in MARC Bibs):
Widget Connector url https://docs.google.com/presentation/d/14KaOeC04Y_UThBMGUnSMFjHjx_fykT-Q8NigFNqClfw/edit#slide=id.g116f1226fe9_1_104
Upcoming meetings:
- 31 August 2022 (Cancelled due to WOLFcon, or aim to have the mtg and cover from screen 6 onwards - TBD)
Comments (A-M forgot to copy the last bit of chat, but will get it from the Google Drive and add)
From Natalie Sommerville to Everyone 01:04 PM
Brb-- handling some office furniture moving logistics.
I'm back!
From Christie Thomas (she/her) to Everyone 01:05 PM
I will be there and would attend.
From Jacquie Samples -- Duke to Everyone 01:05 PM
I would attend virtually, depending on the day/time.
From Leeda Adkins to Everyone 01:05 PM
I'll be virtual as well
From Natalie Sommerville to Everyone 01:05 PM
+1 Jacquie
From Jacquie Samples -- Duke to Everyone 01:05 PM
I'll be there
From Raegan Wiechert to Everyone 01:05 PM
Possibly virtually
From Jacquie Samples -- Duke to Everyone 01:09 PM
I am out of the office (after WOLFcon sessions) so won't be at next week's meeting if it is at 1 PM EST
From Ann-Marie Breaux to Everyone 01:11 PM
No worries, Jacquie - we'll record and I'll add notes
From Jacquie Samples -- Duke to Everyone 01:11 PM
Thanks!
From Ann-Marie Breaux to Everyone 01:12 PM
For last week's notes - I still need to transcribe from the recording. Will do that by the end of this week
Super important - keep MANUAL linking in your head, done by a user, not the system triggering any automated linking
From Jacquie Samples -- Duke to Everyone 01:13 PM
Those interface changes are great, Khaliliah.
From Ann-Marie Breaux to Everyone 01:23 PM
I agree, Jacquie - and not on the mockup, but also in QM edit/derive screen as of Morning Glory - an indicator of which fields are protected
From Jacquie Samples -- Duke to Everyone 01:25 PM
Agreed! I want to just double, super, extra check that as an institution, I will be able to set configurations to never enable Bib to Authority connections. AND then to possibly change that configuration once we get to Entity Management. :)
This is why I didn't engage in answering Khalilah's questions.
From Jessica Janecki to Everyone 01:25 PM
I'm glad you brought up system load Jennifer
From Jacquie Samples -- Duke to Everyone 01:26 PM
+1 Jennifer. such important questions and concerns!
From Christie Thomas (she/her) to Everyone 01:26 PM
Yes, +1 to Jennifer.
From Ann-Marie Breaux to Everyone 01:26 PM
Jacquie - I'll doublecheck, but I'm pretty sure that the linking is a separate permission from other QM edits
From Jacquie Samples -- Duke to Everyone 01:29 PM
Thanks Ann-Marie.
+1 Christie robust reporting is needed at the same moment this type of linking is available.
From Jessica Janecki to Everyone 01:30 PM
linking on $0 should at least be much safer than string matching
From Joshua Barton to Everyone 01:30 PM
Re: deriving in QM, we will do it in limited cases at Michigan State. I think outcome option A seems best.
From Jennifer Eustis (she/her) to Everyone 01:30 PM
Great point Christie. Some linking issues come up between orders and inventory records when you need to move holdings and items for instance. There has also been issues with ghost links - people have seen this with package orders in the title list
From Jacquie Samples -- Duke to Everyone 01:32 PM
+1 again Christie, we use the print version of the record as the basis of records for scanned materials.
@Jessica I agree that matching on $0 should be safer, but I am concerned about the cascading nature of the linking as I heard it described.
From jeanette kalchik to Everyone 01:34 PM
We do it commonly with e-resources
From Jacquie Samples -- Duke to Everyone 01:34 PM
We overlay e-resource records all the time.
From Taylor Smith to Everyone 01:34 PM
overlaying happens a lot here with e-resources as well
From Jessica Janecki to Everyone 01:34 PM
if linking happened automatically, then like A-M says things that have links would be getting overlaid all the time
From Christie Thomas (she/her) to Everyone 01:34 PM
Yeah, we are always overlaying.
From Jennifer Eustis (she/her) to Everyone 01:34 PM
we overlay all sorts of records especially now after we went live
From Taylor Smith to Everyone 01:34 PM
not as much with phys but still not uncommon
From Jessica Janecki to Everyone 01:35 PM
and even if linking only happened manually as part of an artisanal cataloging process, you might still over lay that upon occasion.
From Jacquie Samples -- Duke to Everyone 01:35 PM
Serialists do overlay records somewhat regularly, as details of a title change.
From Natalie Sommerville to Everyone 01:36 PM
+1 to Jacquie and +1 Jessica . Lots of overlays for e resources, but we routinely overlay print records for many types of resources as well
From Jessica Janecki to Everyone 01:36 PM
good example A-M
From Joshua Barton to Everyone 01:39 PM
We also overlay all the time, including physical. And our monthly authorities processing is one of the big overlaying workflows, at least for now. But I don't know how much the introduction of this authority linking might change that. Lisa Robinson can't be here today, but she's the one I'd want to hear from for Michigan State.
From Jenn Colt to Everyone 01:39 PM
Would definitely want to see how all this comparing affects performance
From Jessica Janecki to Everyone 01:39 PM
Agreed Jenn
If it's too slow then it doesn't matter
From Leeda Adkins to Everyone 01:39 PM
+1 Jenn
From Jacquie Samples -- Duke to Everyone 01:40 PM
+1 Jenn
From Jennifer Eustis (she/her) to Everyone 01:40 PM
It might be nice to let the user decide like in OCLC Connexion to control and uncontrol headings.
From Jessica Janecki to Everyone 01:41 PM
K, don't make it too complicated since things are going to change with the implementation of autolinking
From Jacquie Samples -- Duke to Everyone 01:41 PM
+1 Josh. Currently, we do overlay all the time through the authority control vendor service (ACVS) process. We'd hit all types of potentially already linked heading. If we were going to turn this feature on at all.
From Jenn Colt to Everyone 01:44 PM
$0 (+field) seems reasonable
From Jessica Janecki to Everyone 01:46 PM
a vendor should be able to give you a $0
From Jennifer Eustis (she/her) to Everyone 01:46 PM
That's what I think too Jessica
From Jacquie Samples -- Duke to Everyone 01:47 PM
yes, I agree with Jessica. Our ACVS vendor provides the $0 and updates them when needed.
From Jessica Janecki to Everyone 01:47 PM
Yes, if we match, match on the $0, not the plain text string
From Ann-Marie Breaux to Everyone 01:47 PM
That's good to know, Jessica - that if libraries have not routinely received or retained $0s, they may be able to get updated versions of the records with $0s