Pioneers of Open
Pioneers of Open is a collaboration between Creative Commons and the Engelberg Center. It explores how three early open access champions - the Statens Museum for Kunst (SMK), New York Public Library, and National Gallery of Art - created open access programs that continue to act as blueprints for programs developed today.
Introduction
This report is about how open access happens.
The past five years have been good to the open access movement. In 2017, the Metropolitan Museum of Art announced an open access program including 375,000 objects built on the Creative Commons Zero (CC0) Public Domain Dedication. The Smithsonian Institution followed in 2020, with 2.8 million objects and metadata released under the same CC0 dedication.
These announcements were historic. But the Met and the Smithsonian would be quick to say that they were not the first. Their open access programs were built on the work of other institutions and organizations, large and small, that had been working on open access for years.
This report examines three of those institutions, flowing from discussions with individuals who helped to shape the open access journey. The National Gallery of Art (United States), _Statens Museum for Kunst_, and New York Public Library are different institutions. They have different funding models, different relationships to government, and different styles of public engagement. In the years since they started, their open access programs have taken different directions. However, all three pioneered their own versions of successful open access programs.
What follows is not a comprehensive analysis of each institution’s open access program. Instead, it is an exploration of how some of the people who created and operated these programs understood their work. The goal is to provide a window into the process. This window might help those who want to follow similar paths.
While each case study has conclusions specific to the institution, a few points of commonality do begin to emerge:
Digital Infrastructure Matters. Successful open access programs are built on digital foundations that directly incorporate rights and rights awareness. Digital systems redesigns were opportunities to build the possibility of open into an institution’s DNA. Well designed digital backends also made it easier to experiment with smaller projects that were not true one-offs, but rather closely integrated into the institution’s technology infrastructure.
Experimentation is Important. Collections are diverse, as are the users who are interested in them. Open access programs succeed when there is space to try new things, and create multiple points of entry into an institution’s collections. This is true for members of the public who want to explore the collection. It is also true of internal stakeholders who want to understand how open access can help them achieve their own goals. Space takes the form of financial support from within and without the institution. It also takes the space of an institutional environment that is welcoming to experimentation.
Make the Easy Things Easy. Open access programs can be challenging to construct and sustain. Technology must be built. Collections must be designed. Rights statuses must be documented. That makes it important to use tools that make things easier whenever they exist. Those tools include legal tools such as the CC0 public domain dedication, and technical tools such as open source software. The reliability of these tools allows teams to focus on the hard parts of creating open access collections.
These common themes also point to opportunities for the future. The programs described in this report can trace their technical and cultural existence to the web 2.0 moment of the early 2010s. At that moment, various pieces of technology were making it easier, cheaper, and faster to build interesting things online. This enhanced technical capacity merged with cultural excitement about collaboration, creation, and access.
We may be in a similar moment today. Software tools and online services have made building innovative online experiences faster, cheaper, and easier, even compared to five years ago. The permanency of social networks like Twitter appears less certain, elevating a discussion around what kind of digital spaces we want. Every month a new algorithmically powered tool for creation is refocusing cultural discussions on ideas of collaborative creativity. A small team building open access projects today will create experiences that are significantly different, and potentially richer, than those described in this report.
Finally, we can continue to simplify the legal plumbing of open access. That means maintaining and improving tools like CC0. It also means understanding the policies and practices that form the basis of successful open access programs, and making them available to institutions interested in following the same path.
That process of building the next great open access programs will benefit from understanding the work that has already been done. Let’s dive in.
---
I. Statens Museum for Kunst (SMK)
Key Takeaways
The open access program is a key part of projects unrelated to openness at SMK SMK’s history gives it an enhanced institutional obligation to share and make its collection available to the public
The relatively flat hierarchy of SMK, paired with larger donors interested in supporting experimentation, creates a space within SMK to try new things
SMK’s internal data management systems make it easier to build creative projects
SMK went into the process of developing its open access program with an understanding that image licensing fees were not a significant driver of revenue
SMK at a Glance
The Statens Museum for Kunst (SMK) is the national art gallery of Denmark. Denmark’s largest art museum, SMK is located in Copenhagen and is home to approximately 260,000 works of art. The collection began as the private collection of the Danish royal family and is now state-owned and receives significant support from taxpayers. SMK aspires to be “a museum for everyone,” and sees its position as a national institution as giving it a special obligation to make its collection available to the public.
SMK’s open access program, “SMK Open,” has been a leader since its earliest days. The program releases digital versions of public domain objects under a CC0 Public Domain Dedication. SMK now views openness as a core function, proclaiming that “[r]adical openness is the future for our digital cultural heritage.”
SMK’s Journey to Open Access
SMK’s Art Stories project became a catalyst for SMK’s open access program. The project began as an effort to use the internet to increase the portion of the collection that was accessible to the public. The project quickly ran into licensing issues, with copyright restrictions effectively limiting the impact of the effort.
In response to these challenges, SMK secured a grant from the Danish Agency for Culture to support a pilot program to explore how Danish art museums could freely share images. As part of the discussions with the four other institutions involved in the program, SMK realized that its own practice of charging institutions for images established a fee-based norm for all of the institutions. Moreover, the fee-based license practice acted as a significant barrier to innovation without providing any meaningful profits to any institution involved. Moving to a free sharing model allowed SMK and the other institutions to free up capacity, enable new uses, all without creating negative financial impacts at the institutions involved.
In this way, SMK’s early interest in open access was more of a means than an end in and of itself. The open access program enabled it to move forward with the types of interactive, publicly engaging digital projects that were becoming an institutional priority. The interest in open access was not merely in open access for open access’ sake.
Nonetheless, once it had identified the importance of open access to other institutional goals, SMK quickly moved to place openness at the center of its work. The 2011 Sharing is Caring Conference helped to establish SMK as a leader in openness-related conversations.
This conference, which continues to be held regularly, was a major step in building open access into the brand identity of SMK itself. SMK also understands open access as a complement to its ethical obligation to make the collection available and accessible to the public that helps support the institution. SMK is a steward of Denmark’s national collection. Part of SMK’s stewardship duties involves making sure that the citizens of Denmark can access and make use of as much of the collection as possible.
Implementing Openness
Open access has been an ongoing process at SMK.
Context
When considering open access for the first time, many institutions (and institutional leaders tasked with the fundraising required to support the institution) worry that making images of objects available for free will undermine existing image licensing revenue.
This was not a barrier for SMK. Influenced by a study on reproduction charging models commissioned by the Mellon Foundation, SMK leadership recognized that image licensing was not a significant source of revenue for the institution. The absence of countervailing revenue pressure from lost image licensing revenue provided a space for experimentation and testing for the open access project.
The period of experimentation matched a moment of increased interest in digital transition from foundations and other funders. These partners were hungry for imaginative digital projects. SMK’s early experience had taught it that open access was key to such projects, and was able to bundle open access infrastructure into a larger program of digital innovation. This experience, developed as a result of early institutional decisions to position itself as an inclusive, accessible, and visionary institution, allowed SMK to raise funds that supported a multi-year digital development effort. These funds made it much easier to say yes to new ideas, and helped to create the space to experiment and test.
The fruits of this experimentation helped SMK emerge as a leader in open access and digital innovation. This leadership role did not require that SMK’s directors were themselves deeply fluent in the possibilities of a digital transformation. Instead, the highest level leadership could support a broader concept of visionary digital change, allowing others within the institution to push forward with new ideas. The relatively flat organizational hierarchy of SMK further empowered subject area specialists to try new things that were exiting within the context of their area of expertise.
Implementation
SMK’s early groundwork in support of open access created an environment with the space and support (both managerial and financial) to try new things. Initially, it adapted the startup-style mantra of “think big, start small, move fast,” SMK was able to fall forward and rapidly experiment.
Over time, SMK revised the mantra to “think big, start small, reflect.” This shift from “move fast” to “reflect” is motivated by a recognition that digital transitions can impact a diverse set of stakeholders. SMK strives to take these questions seriously during the development process, not merely after the fact.
This mantra manifested itself in an incremental approach to implementation. By testing many small projects, approaches, and tools, SMK was able to identify the ones that worked and support them in order to become viable. These viable projects drew public attention, allowing SMK to position itself as a leader within the open access community. This resulted in SMK being recognized with other open access leaders, creating an innovation flywheel where institutional confidence in open access supported new projects, the success of which further increased institutional confidence in open access.
Importantly, this trajectory did not require a 100% success rate, or for the institution to be able to justify the entire open access program by pointing to a single project. Instead, by creating the space to experiment, SMK fostered an overall successful approach to open access.
Challenges
As this success built on itself, the ambitions of SMK and the larger open access community expanded. At one point, SMK entered into a larger, more ambitious agreement to collectively develop an integrated collections management system with an international group of GLAM institutions.
The benefits of such a system could have been significant. A unified data structure and technical backend would make it easier for projects to draw information from a range of sources. This would allow each institution, as well as outside organizations, to build larger, more ambitious projects.
Unfortunately, this approach proved itself to be fragile. It was scaled and scoped to require an ongoing commitment from each of the partners in order to succeed. As partner institutions began pulling out of the consortium, the remaining institutions were unwilling or unable to make up the shortcoming. The project was never completed.
Partially as a result of this experience, SMK made an institutional decision to move towards a more agile approach to content management. Instead of trying to build a digital collection system to end all systems, SMK developed a data system that worked for SMK. It also invested in documenting its Application Programing Interface (API) so that others could easily access SMK’s collection from the outside. Today, this internal system acts as a backbone for SMK as a whole. Everything within SMK relies on it, and outside partners access it via the API.
Although it has allowed SMK to continue its successful open access program, this approach is not without its own costs. SMK supports its own in-house developer. It also works with a team of outside contractors to help grow and maintain the system. SMK has prioritized working with smaller companies because those companies are often more invested in SMK’s priorities and success, as opposed to being focused on pushing SMK towards the company’s preferred solution. In addition to its in-house developer, SMK relies on a Head of Digital within the institution with the skills required to act as a bridge between SMK’s needs and the capacities of outside contractors.
Institutional Lessons
SMK’s internal content management system represents one of its key institutional lessons. Building a sustainable digital backbone has made it much easier to try new things and experiment with one-off projects. It also effectively makes experiments less “one-off” because they are linked to a core system that is actively maintained and relied upon by a host of other initiatives at the institution.
Building and maintaining a robust content management system also allows SMK to focus on what it understands as its core competency of using its open dataset to empower other users who are doing interesting and impactful work. That means SMK spends most of its time on research, education, and maintaining its data to make sure it is as accurate as possible.
This focus is partially the result of negative lessons that SMK has learned, specifically to avoid projects that require skills that the institution does not have. SMK intentionally avoids projects like apps, QR codes, and screens and digital tables in the institution itself because they are outside of SMK’s core institutional expertise.
SMK’s wariness of screens in museums can run contrary to expectations for an institution that is so invested in digital openness. However, over time SMK has concluded that visitors to the physical museum are looking for specific types of experiences that directly connect them to the works in the collection and to each other. More often than not, screens on those spaces end up detracting from those types of experiences. Although SMK supports many screen-based projects outside of the institution, within SMK’s galleries the objects themselves are at the center of visitors’ experiences.
Finally, SMK as an institution has recalibrated its understanding of the value of digital and open access. Instead of expecting digitization to make things cheaper, it understands that digitization allows SMK to work smarter and more effectively. Similarly, embedding openness into SMK’s institutional DNA - making open access part for SMK’s brand - has made all of these projects easier. The institutional alignment allows for a persistence and stability that supports the long term growth of effective uses of its collection and digital assets.
---
II. New York Public Library (NYPL)
Key Takeaways
Interest in open access took advantage of opportunities presented by a larger institutional shift to digital infrastructure
Upper-middle level management was able to create a space for innovation, but the lack of full institutional buy-in made that space hard to sustain
Staff interested in innovation generally were also interested in open access as a value
Combining open access with a digitization service fee model can build revenue The “lab” model helped to create a team that worked to inspire each other
NYPL at a Glance
The New York Public Library (NYPL) is the largest public library system in the United States, with branches in the Bronx, Manhattan, and Staten Island. In addition to its branches, NYPL is home to research collections that contain tens of millions of objects (not only books but also sculptures, toys, wax cylinders) organized into 100 major topics. Some of these collections are housed in purpose-built facilities, such as the Public Library for the Performing Arts at Lincoln Center, the Schomburg Center for Research in Black Culture in Harlem, and the Thomas Yoseloff Business Center at the Stavros Niarchos Foundation Library in midtown Manhattan.
NYPL’s digital public domain collections currently include over 180,000 objects, including books, posters, prints, sheet music, photographs, personal paper, and more, released with a statement indicating that it believes that they are in the public domain. It also includes a number of open source tools to use objects and data from the collection.
NYPL’s Journey to Open Access
NYPL’s open access program can trace its roots to an internal memo about the library’s broader digital strategy. This memo set in motion a number of efforts to digitize the library and its collection —- to create a library of the internet, not just on the internet. Part of this process involved setting up the Digital Library + Labs in 2015, an internal team dedicated to exploring what was possible at the intersection of NYPL and digital technology.
The NYPL Labs team, and the digital effort more broadly, quickly identified a connection between their work and the value of open access. Without internal open access structures, every digital experiment involved exploring, navigating, and documenting rights for any work involved. This was not sustainable, and created urgency to accelerate an open access program within the institution.
The practical need for open access to streamline digital projects dovetailed with a cultural affinity between team members involved in NYPL’s digital transformation and larger open access movements outside of the organization. In some ways it was a happy coincidence that the digital transformation of NYPL was happening at a moment when open-friendly sharing programs like Flickr Commons were in their ascendancy, peer institutions such as Cooper Hewitt and the Brooklyn Museum were experimenting with open data, and open access programs had been established in academic research libraries for some time. As a result, the people who were shepherding NYPL’s digital transformation also happened to be personally interested in discussions around open access that were happening more broadly online. This meant that the process of digitizing NYPL integrated open access concepts from the beginning.
Implementing Openness
NYPL’s initial implementation of open access was deeply tied to staffers’ perception of NYPL's efforts in the context of broader sector initiatives and their desire to build tools and programs to put NYPL digital objects in conversation with those at other institutions.
Context
Revenue is often a concern when institutions shift towards open access programs, especially for those that have licensed digital files in such contexts as publishing. In the case of NYPL, image licensing revenue had been in decline prior to the start of the openness discussion. Perhaps just as importantly, revenue licensing itself was a relatively modest part of the overall NYPL budget. This combined to make it easier for NYPL to experiment with new open access initiatives without becoming overly fixated on how they might impact an important revenue stream.
This freedom ended up producing new revenue for NYPL. First, NYPL featured prominently a “buy the print” button on every open access item page. While users were free to download the files themselves and print the images elsewhere, the ease of printing directly from the item page resulted in a steady stream of orders.
Second, NYPL effectively turned its open access collection into an advertisement for its paid digitization services. Any user could download the digital files for objects that had already been digitized. That included tens of thousands of artifacts.
At the same time, NYPL’s collection includes millions of objects that had not yet been digitized. NYPL offered digitization services for a fee. Anyone could request that a public domain object be digitized, pay the fee, and receive a digital version of the file. That file was then also added to the publicly available open access program. High profile open access projects drove interest in the collection, which inevitably surfaced undigitized objects of interest. This allowed the team to continue to digitize works in a way that was responsive to user demand, paid for by that demand itself.
The leadership structure of NYPL Labs also played an important role in the development of the open access program. The head of Labs had some degree of budgetary authority, which created space for experimentation within the group. This may also have allowed the head of Labs to move forward with some projects without fully engaging the highest levels of NYPL leadership. While useful in the short term, this lack of full institutional buy-in may have created longer term weakness for the initiative.
Finally, the Labs-driven open access initiatives were able to expand rapidly at launch because they built on years of internal work within NYPL. The NYPL Digital Gallery debuted in 2005. That technical system had been mostly rebuilt by 2015, but at that point it still did not contain a system for displaying rights. A 2015 redesign was an opportunity to integrate the years of internal rights clearance work by the NYPL into the publicly-facing collections system. Combining existing legal review work with a user interface redesign made it feel to the public as though NYPL was suddenly bringing a large portion of its collection into the world of open access.
Implementation
NYPL’s open access program implementation was guided by the slogan “free to use and reuse.” The team hoped that the slogan was so simple that it could not be misinterpreted —- everyone should be able to use and reuse the collection as they see fit without any caveats or hidden processes. This concept — simple, inviting — helped guide many decisions in the development process.
By 2015, NYPL had been sharing digital objects online for a decade. Other institutions were also taking steps to put their collections online. As a result, the open access team wanted to go beyond simply sharing the collection —- they wanted to actively promote reuse. The “free to use and reuse” slogan helped to communicate that goal to the public.
The 2015 effort was also an opportunity to move away from independent microsites used to show off the collection. While individual microsites might make a splash when launched, they quickly became hard to maintain. NYPL’s new approach would build a more holistic ecosystem of smaller projects tied together by a robust asset management system. A unified backend would make it easier to launch new projects, and to maintain them once they had been launched. One-off sites would no longer truly be one-off — they would be linked to the core content management system that NYPL relied on for its own operations.
NYPL also found that it needed new tools to work with newly accessible collections objects and data. These tools themselves became part of the program, released under permissive licenses and available to anyone else who needed them.
Combined, these efforts moved away from a vision of open access where an institution posts images and calls it a day, and towards a more robust, complex approach involving data, tools, and projects that invited users to explore the collection. In order to make use of this new capacity, public outreach became a core part of NYPL’s open access program.
The outreach strategy was oriented around smaller projects that functioned as examples of what could be built on NYPL’s content management system. The open access team believed that every collection at NYPL had an interesting project hidden within it, and worked with curators to find them.
This small project approach also acted to maximize the number of entry points for media and the public. Each project was an opportunity to connect items in the collection to a new audience in a new way. Combined, these projects allowed NYPL’s open access program to go viral in many different directions, connecting with a range of audiences. In some ways, the projects also acted as a representation of NYPL’s broader capacity to innovate.
Challenges
NYPL Labs and its open access initiatives found a significant amount of success during 2015-2016. However, that success was often understood as a success for NYPL Labs, and not necessarily for NYPL as an institution. The organizational bubble that had allowed NYPL Labs to flourish also created distance between NYPL Labs and the rest of NYPL. When NYPL underwent a leadership change, the equilibrium that had sustained the Lab became unstable and Labs was absorbed back into the rest of the institution. Although NYPL continues to support its open access program and contribute to its growth, the specific focus on research collections and the innovative propulsion that made use of the open access collection has diminished.
A move towards larger, consortium-based projects may have also contributed to the reduction in innovative internal projects. These larger projects had a tendency to draw budget and operational capacity away from individual projects, reducing NYPL’s ability to create one-off entry point projects for individual collections.
Finally, the success of the open access Labs projects opened a discussion about resource prioritization within NYPL. While they proved popular, many open access projects were targeting a group of users that may have been distinct from the patrons visiting individual NYPL branches and research centers. A wariness of this discrepancy may have made NYPL less enthusiastic about continuing to support these projects.
Institutional Lessons
Although NYPL’s open access program is ongoing, the work of NYPL Labs building tools to develop new ways to bring users to those collections covered a relatively brief moment in time. The success of NYPL Labs relied on the administrative and budgetary space to test new things. This space produced a number of highly successful projects that relied on the NYPL’s open access collections. Unfortunately, it ultimately proved to be unstable and unsustainable.
These projects, and the open access program more broadly, are a testament to the power of digital infrastructure. NYPL’s homegrown asset management system integrated rights management directly. This digital infrastructure allowed NYPL and the public to benefit from the significant work done by NYPL’s rights clearance group to investigate the rights status of works in its collection. The asset management system was also able to act as a backend for a wide variety of initiatives. The technical system and rights information formed a foundation that allowed the NYPL Labs team to focus on creating new ways to explore the collection itself.
Finally, the most successful projects illustrate the power of working (although not necessarily thinking) small. While completely one-off microsites will always be hard to sustain, individual sites and projects that can draw from existing infrastructure can be incredibly powerful models for what is possible with an institution’s collection. They can act as new entry points, drawing attention and enthusiasm from users. Those entry points, in turn, can encourage new types of users to move deeper into the collection, discovering new treasures along the way.
---
III. National Gallery of Art (NGA)
Key Takeaways
Understanding public access to the collection as a core value of the institution made it comparatively easy to create an open access program
An approach grounded in stability and incrementalism creates an open access program that works within the institution
Linking open access to the success of internal projects can fuel open access growth
NGA at a Glance
The National Gallery of Art (NGA), located in Washington, D.C., was created by the Congress of the United States in 1937. Although it shares a location on the National Mall with the Smithsonian Institution museums, NGA is an independent entity. It receives annual funding from Congress. This public support helps to create an institutional culture built around serving the public good. NGA is home to over 150,000 objects.
NGA’s Journey to Open Access
NGA’s interest in open access is deeply linked to its larger institutional mandate to serve the public good. The organization itself receives significant levels of financial support from the United States Congress and describes its history as “founded as a gift to the nation.” This has created an institutional culture that prioritizes public service. There is an institutional understanding that open access is deeply intertwined with that public service mandate.
In 2012 the NGA launched the NGA Images program. This program made high quality images of public domain works in the NGA collection available to the public for free online. The program grew out of a recognition within the NGA that it was losing control of how images of works within its collections were being used online. This resulted in objects from the collection being represented by poor quality images. Giving the public high quality images allowed the collection to be represented online in the best possible light.
The program was shaped by the needs of different parts of NGA itself. There was, and continues to be, a robust curatorial interest in linked data about the collection. This interest drives work to enrich collection metadata, both by adding information and connecting bits of existing information.
This curatorial interest in data and linked data has increased internal expectations for what can be possible with NGA initiatives. The launch of the NGA Images program instigated a long-term shift in how curators thought about data sets, with curators coming to the team with datasets they were especially proud of. The curators wanted to think more broadly about how to get that dataset out into the world.
The program also created an entry point to educate curators about tools. As open access became more popular within the institutions, curators developed an organic interest in learning how to use tools like Wikidata and OpenRefine to explore their own datasets. The fruits of these explorations would then become new contributions to the open access program.
Implementing Openness
The NGA’s open access program is characterized by its steady development trajectory. Its initial goal was simply to enhance user experience of the collection, and to increase the collection’s (and, by extension, NGA’s) reach.
The strong relationship between the lead of the open access program and the NGA’s General Counsel formed the foundation of the open access project. They formed a collaborative team that eventually drew in a constellation of supporters within NGA. This constellation of individuals worked in an atmosphere of trust and support, aligned by a common understanding of the program’s goals.
The open access program was also able to leverage infrastructure that was already in place at NGA. Prior to the launch of the program, the NGA already had a well-functioning photography department. This made it relatively easy to create the high quality images of objects in the collection that are at the core of the open access program.
Perhaps equally important, the NGA’s relatively small collection (for an institution of its prominence) bounded the effort. Comprehensive open access at the NGA would require the digitization of 150,000 objects, not tens of millions.
Like many other institutions, NGA also took advantage of redesign of its digital infrastructure to make open access a more prominent part of the institution’s internal and external identity. Rebuilding the NGA’s internal infrastructure with open access in mind created the capacity for long term change.
That change was understood from the beginning as an interactive process. Once it had a large enough collection of open objects to launch, it launched and listened in order to understand how the collection was used, how future development should be prioritized, and what could be done to make it even easier for the public to access it. After the program’s initial launch, the NGA continued (and continues) to add more objects and more images.
For example, in 2014 the NGA launched “Online Editions,” an online project to enrich CC0-licensed images of works in the collection with rich scholarly context and commentary. The first Online Edition combined a range of information into a single online resource. That included an updated version of a book on Dutch paintings by an NGA curator, as well as the open access images of those works from the NGA’s collection (those images were originally licensed under a bespoke, NGA-drafted license and have since shifted to CC0). Online Editions have since expanded to include a number of different collections, scholarly authors, and artists. Each of these editions helps to showcase the NGA’s collection and relies on its open access images
NGA has also worked to make it easier for the public to understand its licensing policies. That includes using the CC0 Public Domain dedication for object metadata and images. Clear, simple licensing allows NGA to lean on partners like Wikimedia Commons, Art Store, Google Arts & Culture, and Github to distribute the collection as widely as possible in formats that people find maximally useful.
Institutional Lessons
The NGA’s open access success is built on a combination of digital infrastructure that makes it easy to track rights information and an institutional culture that centers and values the public’s access to its collections.
This institutional DNA of service combines with having a stable combination of the right people in the right positions at the right time. The early constellation of open access supporters helped to shape an environment that created a long-term commitment to open access approaches. This commitment creates the space for open access to feel less revolutionary and more fundamental to the NGA’s decisions, both large and small. More recently, the NGA has worked to bring its entire open access collection to platforms such as Wikimedia. By February of 2020, it had contributed 120,000 art object collection records to Wikidata.
---
© Creative Commons, 2022, made available under Creative Commons Attribution 4.0 International (CC BY 4.0).
