2020-10-23 - Consortia SIG Notes

Link to recording

• Date

23 Oct 2020

Announcements

Attendees

• Lloyd Chittenden
• Sebastian Hammer
• Noah Brubaker
• Steve Walker
• Brooks Travis
• Rose Nelson
• Paula Sullenger
• Martina Tumulla
• Peter Sbrzesny
• Lucy Harrison
• Rick Burke
• Jill Morris
• Kirsten Leonard
• Marie Waltz
• Nicole Becwar
• Kyle Banerjee

Goals

• Draft vision document for an Open Source Bib Utility

Notes

We had a joint meeting with people from the International Coalition of Library Consortia.  Lucy Harrison gave a summary of what has been happening at ICOLC.  They have been having very similar conversations about the possibility of an open source bib utility.  They had a task force thinking about this topic for about a year.  They have had various concerns about OCLC.  They did a survey of ICOLC members to find out their issues with OCLC, and they met with OCLC at the January 2020 ALA Midwinter.  And they had a couple of folks from OCLC come and we came them and initial preview of their concerns including: pricing, quality of database, standards and interoperability, and general marketplace and culture concerns.  They thought OCLC's response was condescending.  They mostly dismissed the questions rather than answer them.  ICOLC does not think OCLC will make any changes based on this critique.  So they are starting to think about what they will do if OCLC does nothing.

Lucy shared their draft working document on this subject.  Including four areas of recommendations. 

• They complain that the pricing is not transparent, customers are required to buy things that they don't want, and they appear to be commercializing things unnecessarily.  For example, in 2017 EZ proxy cost GALILEO $30,000 a year.  Now OCLC wants $400,000 for EZ proxy for hosted.  Even non-hosted would be $200,000 a year.
• They are not happy about WorldCat database quality.  As fewer libraries can afford to keep their holdings in WorldCat, the value decreases.  They recommend that WorldCat should be open source so everyone can maintain holdings and see everyone's holdings.  It could be paid for with advertising. 
• There is a problem with standards interoperability.  WMS does not work with other resource sharing systems.  There should also be interoperability between resource sharing systems.  Holdings information should be exchanged, and there should be standard authentication protocols.
• The fourth area is marketplace and culture.  They are not living up to their cooperative values.  They are trying to be a vendor and a non-profit, but they are acting much more like a vendor than a non-profit cooperative.  Their CEO salary is now $1.67 million.  It was $260,000 in 2001.  They pay board members up to $80,000 annually.  Board members remain for 10 or 12 years.  They should separate cooperative activities from vendor activities.  Provide clear accounting of the actual costs of supporting each program and services.  Especially the core services on with they were founded.  Stop compensating board members.  Of 14 board members, 8 are appointed by the board themselves.  Only 6 are elected by members.  Elected members often only stay 4 years, while those appointed by the board often stay 12 years.  So the board is self-perpetuating.

They have considered alternatives to OCLC assuming they will not make any changes.  They have looked at now other countries do these same things.  There are already open source products that compete with OCLC such as Gold Rush.  There are others that are being developed to compete with the OCLC core areas.  Jisc in the UK is moving away from OCLC to a locally developed system.

Rick points out that ICOLC is not a legal entity.  It is an informal organization of Library Consortia.  It has no legal standing.  ICOLC has a virtual meeting of European members next month where they will talk about this.  Rick says that he has heard from OCLC that these things are issues for North American libraries, not international.  They hope to find out at the European meeting if that is correct.  OCLC does not operate as a non-profit outside of the US. 

There is no reason that we can't follow the example of OA2020.  There is enough money in the system that we can convert to another approach.  The ICOLC report makes it clear that there is enough money in the system.

Rick asks if people think there is an issue with EBSCO's involvement with FOLIO?  He doesn't think so, but if we are worried about OCLC acting like a vendor, then we go off an do something with another vendor.  Some librarians might think we are just in bed with EBSCO instead, so what's the difference?

Lloyd says that we have both been talking about the same problems.  However we didn't know the detail about how OCLC was spending it's money, we just knew it cost too much.  The Consortia SIG is currently considering placing this project in the ReShare sphere, not FOLIO.  ReShare is not as connected to EBSCO as FOLIO is.  But even with FOLIO, EBSCO has not been heavy handed at all.  They are putting their funding toward developing the parts of FOLIO that their subscribers need, but they are not pushing development otherwise.

Lucy says that they don't want to recreate OCLC.  We want to create tools that people need.  There are other ways that it could function.  We could avoid becoming another OCLC by how we design the product.

Travis asks if the report get into the issue of OCLC's capital investment problem?  No, it does not, what is that problem?  OCLC bought a lot of server capacity rather than using cloud based services which has driven thier acquisitions and service offerings to fill that capacity and recoup that investment.

Kirsten points out that they built their own data centers to comply with Patriot Act requirements.  They have a lot of compliance issues that drive up their cost.

Sebastian says that he used to work for the Danish National Bibliography, which is a Danish version of OCLC.  He thinks that they would have trouble due to the fact that becomes easier to manage large amounts of data over time with hardware costs coming down.  In Denmark they had some of the same dynamics.  He's worried about an assumption of a single central database, because then you just become another OCLC.  That requires a lot of central staff.  He thinks they treat WorldCat as a cash cow.  Some of their other activities seem to be burning off cash.  Such as the research department.  Once they get that staff then it becomes an end in itself.  They have many structural advantages because they have all the data from decades, but there is also baggage.  The other thing that strikes him is that reminds him of the mood that brought together the ReShare Community.   A community like that often starts by declaring itself in opposition to something.  Once the community is in place the thing you are opposed to is less relevant.  ReShare has been able to watch how much it influenced OCLC's thinking.  It has put pressure on them to change behavior.  He thinks they made changes to ILLIaid because of ReShare.  So this project could have similar influence on WorldCat.

Lucy says that they have been trying to work with OCLC to improve them.  They only recently realized that they need to start thinking about what they will do if OCLC does not improve.  The FOLIO Group has been working on a vision document.  There seems to be some synergy between the two groups.

Sebastian gives the perspective that ReShare is already focused on the space between libraries.  It is focused on consortia already.  FOLIO is focused on getting individual libraries up and running.

Jill says that strategically how we move this forward by bring the community together around that shared vision.  FOLIO has been controversial in some people's mind because of the EBCSO connection.  ReShare has remained independent from that.  ReShare is self governed.  They have no EBSCO money.  They are a group of consortia and commercial developers who have a shared vision to approach our work in a particular way.  This is no big player to content with. 

Lucy says it seems like with ReShare the workflows and information needed to perform resource sharing is the same workflow and same information that you need to copy a bibliographic record.

Sebastian says moreover there are different ways you can do that.  If you are presented with the problem of consortia resource sharing.  The path of least resistance if to have everyone dump there records into a central SOLR database.  ReShare had a desire to not just look at the solution consortia by consortia.  What does it look like for a library to be part of more than one consortia?  They tried to come at it from an infrastructure that would allow libraries and consortia to serve as aggregators for individual libraries and build further aggregations of those until you get something that has the scale of OCLC.  Or it could be a mix of different things.

Lucy: that could be a first step.  They have libraries that only contribute to OCLC would be fine not doing that, but then there are major research institutions that will continue with OCLC until there is a super robust alternative.  There is a lot of space between them.  The dilemma is that if we make it good enough to meet the needs of a lot of people, then enough people will join that it can become better. 

Sebastian: What ReShare tried to do is think about how to move the data around.  If we lean on standards as much as possible rather than the software or the organization, then we have an openness that can invite ExLibris or OCLC to offer that they join.  It doesn't need to belong to a single organization.  ReShare has experience with how difficult it is to get to OCLC's stuff without having a deep partnership with them.  ExLibris by comparison is far more open by comparison.

Kirsten: there have been conflict over who owns what data.  They have had trouble getting ExLibris and EBSCO to release discovery data.  The only thing OCLC has is the individual holdings data.  That's very important to them.

Lucy: but it's our holdings data.

Kirsten: OCLC is likely to sue or at least threaten to sue if people are taking their holdings from other bibliographic entities that they didn't create out of OCLC.  When they barter that is what they barter on.  We need to think about these things when we think about the governance structure.  A diffuse governance structure can be more expensive.  When we are working on the vision document we need think about what are the problems we are really trying to solve.  We need to be sure we are realistic about what we can put forward.  For us it's cost.  I don't want to recreate an environment where costs are uncontrolled.  Or that we create something that ends up having added cost because of staffing or complexity of the product.  Right now PALNI is all in on OCLC, but that is sustainable for now, but probably won't be in the future.  The last think I want to do is set up something that we can't manage.  OCLC is the behemoth that is very hard to move.  They have so much messy data that they can't pivot the architecture or governance or business model to meet changing environments.  We should figure out what it is we really need over the next ten years and build that.  And try to not perpetuate any legacy situation.  I want to be really pragmatic.  We need to think about who are the players, are there enough players, is there enough business reason for people to contribute, are we creating a free rider problem?  All of this is what vendors are doing now.  They may be doing it more efficiently because they have fewer people to consult.  The more diverse you make the voices in the environment, the less efficient.  There was a task force in 2010 that brought up many of these problems, but they haven't fixed them.  We probably aren't going to be able to fix OCLC.

Rick: I think it's really important to acknowledge what Sebastian said.  We have a concrete example of a resource sharing network that we build that OCLC has effectively destroyed it by the lack of compliance with NSIP.  They've sold WMS as a low cost alternative and then they undermine you because they won't work with anything else.  I'm disgusted with with they do.  When we talked to our non-profit lawyer he said we need a three year plan, a project coordinator, a tech team, and a business modelling team.  I think ReShare already has those things.  We probably need legal back up.  I don't understand how OCLC can sue us if we just put our data somewhere else.

Kirsten: they think it's their data in your system.

Rick: It's our data:

Lucy: Some of it is their data.  I don't think they would have much of a leg to stand on if OCLC didn't create the record or purchase it from a vendor.

Sebastian: The one place where they will go after people is with their identifier.  That one thing that we all need it the unique identifier.  The OCLC identifier is the one thing we need because it is manually maintained.

Lucy: It's pretty good, but we decided against using it for de-duplicating in Florida.

Kirsten: While we are talking about where the holdings will be held, having a co-located place where holdings are being held or at least the ability to run analytics across multiple holdings is pretty critical.  That's one of the things we see potentially lost.  As Lucy said the identifier is not the be all and end all.  There are other mechanisms to determine uniqueness.  We at least need an analytics model that will be a unifier.

Sebastian: ReShare shared index is using the Gold Rush algorithm.  If we can get to a place where we have a universal approach to how we establish identity, then there are all kinds of things we can do across services that are really exiting like getting a separate analytics platform that doesn't have to belong to the aggregation and holdings platform.  It is still to be determined if that does what it needs to do and exactly what roles it needs to fulfill.

Lucy: It is like the SSN.  It is supposed to be the identifier, but if I call my credit card company, they don't just rely on that.  They ask me four or five other questions to determine if I am who I say I am.  That's what really identifies me.

Discussion items

Action items

• Invite Theodor Tolstoy to next meeting Lloyd Chittenden
• Invite Kristin Wilson Lloyd Chittenden