 Maria Gould from California Digital Library. In this presentation, I will provide an update for the CNI community on the Research Organization Registry Project, also known as ROAR. ROAR is a community-led project to develop an open, sustainable, usable, and unique identifier for every research organization in the world. At its core, the project is a fully open and non-commercial registry of persistent identifiers and associated metadata for research organization affiliations. We have delivered presentations at previous CNI meetings that go into more detail on the background behind the project. So in this talk, I want to focus on the role that ROAR can play in scholarly infrastructure going forward. In the world of online scholarship, persistent identifiers are a key part of the infrastructure that we use to search, discover, access, and site research. And they're also a key part of how we tie different parts of the research landscape together. For example, if we want to be able to track all of the research products published by a particular researcher. ROAR emerged because in this research landscape, we already had existing mechanisms to consistently and unambiguously identify and discover certain things. We could already identify researchers using Orchid IDs and we could already identify research outputs using DOIs for articles and datasets and other publications. But without being able to identify the institutions linked to this research, it wasn't possible to have a complete picture of the research landscape. Affiliations were a missing piece of this puzzle. So we built the ROAR registry to be able to provide open identifiers for organization affiliations that can connect all of these pieces together. The ROAR registry launched in early 2019 and there are currently about 98,000 organizations in the registry right now, each with a unique ROAR ID along with additional metadata, including crosswalks to other identifiers. I'm showing on my screen here the search result in ROAR's front end search interface for my organization, California Digital Library. ROAR also has an open API where organization IDs and metadata can be searched and queried and filtered and the API can be connected to other systems. So here is a portion of the same CDL entry as represented in the API. I want to take a quick moment to acknowledge that ROAR is by no means the first or the only effort to establish an identifier for institutions. But it is different from other identifiers in a few key ways. First, ROAR is completely open and free. ROAR data is CC0 and our code is open source. We believe that institutions should not have to pay to access their own data about where their researchers are publishing. Second, ROAR's scope is different because it is specifically focused on top level affiliations for research outputs. So that means that ROAR is not mapping all of the hierarchies and departments within a given institution, nor is it a registry of all legal entities in the world. Lastly, ROAR is unique because it is being developed as a collaboration across community partners in the scholarly communication landscape, specifically to address the needs of libraries, research administrators and academic institutions so they can have clean and accurate reporting on their research outputs. ROAR is also distinct from other types of identifiers because for the most part, ROAR IDs are meant to operate in the background as a key piece of metadata that is collected and embedded in scholarly publishing systems and used to enrich persistent identifiers and paid infrastructure. The promise and the vision for ROAR is that it functions as this open standard and global identifier for research organization affiliations that can be implemented in any scholarly system where affiliations are collected and then deposited along with scholarly metadata in open indexes like Crossref and DataSite where the ROAR IDs can be searched and connected to other systems linking research organizations to research outputs and to researchers. So to give one example of how we see this playing out, I'm showing search results here from Crossref's API for DOIs registered for various publications that have one or more authors affiliated with Queen's University. And as you can see, I'm getting results for Queen's University and Queen's University Belfast and there are at least three different ways in which Queen's University is written and three different ways in which Queen's University Belfast is written. So if I'm just searching on the name, I may not be getting a complete or an accurate set of results. If ROAR IDs were used instead, it would be easy, much easier to find all of the publications registered in Crossref with affiliations for either Queen's University that happens to be in Canada or Queen's University Belfast. And I would be confident that we are getting an accurate set of results. In an ideal publishing workflow, when a researcher submits an article or a data set and provides an affiliation, the options can be based on ROAR's controlled list and then the ROAR ID can be stored in the submission system and the corresponding affiliation will be published along with the research output. And then when the metadata is deposited in Crossref or in data site, the ROAR ID can be included in those deposits. This kind of end-to-end pipeline is actually happening right now with data sets submitted in Dryad. When a researcher submits a data set to Dryad, as you can see here, the researcher is asked to provide an affiliation and when the researcher starts typing an affiliation, the researcher makes a call to the ROAR API that finds a match in ROAR's controlled list and then the researcher can choose their affiliation from that list of options. The researcher isn't even aware that ROAR is operating in the background, but Dryad is able to collect that affiliation and store the ROAR ID in its database. And then Dryad can deposit the metadata for this submission in data site and include the ROAR IDs as part of that deposit as well. So you can see here how this shows up in data site XML. Crossref's metadata schema is also being updated right now to support ROAR IDs in a similar way. And research and data site can now be searched on specific affiliations based on the ROAR IDs provided by data site members like Dryad. So this is the public search interface that anyone can use right now to look up research in data site and to look up research associated with specific affiliations like Queen's University Belfast. In addition to Dryad data site and the forthcoming Crossref integration, there are a number of other implementations that have already been completed or are in progress since ROAR's launch last year. This list shows the integrations that we know about. Of course, there may be more. And you can see just from this list that there are a range of different applications and different types of systems that are integrating ROAR. So in just over a year, the ROAR registry has launched and ROAR IDs are already being supported in data site metadata and being adopted in many systems. So as we continue our work this year and beyond, there are a few milestones that still lie ahead. As mentioned previously, Crossref metadata will soon support ROAR IDs. And this means that Crossref members, mainly publishers, can collect ROAR IDs for affiliations and deposit this metadata in Crossref. So this year and beyond, we're also working to drive wide adoption by publishers as well as by data repositories. Another key milestone this year is implementing a comprehensive curation model for ROAR. We launched the registry with seed data from Digital Sciences Grid Database. And we're now building the technical infrastructure and developing policies and workflows to support independent and long term stewardship of registry data so we can support additions and changes to the registry over time. And we're also focused more generally on ROAR's long term sustainability. ROAR is a collaborative cross organization effort that functions primarily on in kind resources. And we're more or less trying to keep it that way. We want to keep the overhead low and keep it community based. We don't plan to see a need to or we don't plan or see a need to establish a new legal organization for ROAR. But that being said, there are basic costs associated with sustaining the registry over time. So we're doing some fundraising right now to secure startup funds through community investments and grants. And then we ultimately plan to introduce an optional paid service tier in 2022 that will allow us to cover our basic costs while also ensuring that we can keep the core registry data open and free and perpetuity for everyone. So ROAR is a community effort. And if you are interested in being involved, we encourage broad participation and engagement. There are a number of ways to get involved or at the very least to stay on top of project news and updates. We publish blog posts semi regularly on our website. We have an open Slack workspace for discussions. You can follow and contribute issues on GitHub. You can follow our news on Twitter. We have a mailing list that you can sign up for in the footer of the ROAR website. And if you like to be part of a community call every other month or so where we discuss project updates and get started on various working group projects, you can email ROAR to join that group. So thank you very much for watching. I look forward to staying in touch.