The following is a guest post by Morgan McKeehan, National Digital Stewardship Resident at Rhizome. She is participating in the NDSR-NYC cohort.
I began my National Digital Stewardship Residency at Rhizome — NDSR project description here (PDF) — by leading a workshop for the by leading a workshop for the Emulation as a Service framework (EaaS), at “Party Like it’s 1999: Emulated Access to Complex Media Collections,” a panel about software emulation organized by NDSR-NYC alum Julia Kim, at the 2015 AMIA conference in Portland, OR. Software emulation is an important tool for preservation of digital artworks because it allows researchers to experience complex digital materials in their native creation environments, and can thereby enable full access to “software dependent content,” the term offered by Euan Cochrane, Digital Preservation Manager at Yale University, for content that is integral to the overall meaning of a work, but which “requires a particular and limited range of software environments in order to be interacted with, rendered, viewed or consumed.”
The EaaS framework provides a streamlined user experience for accessing a number of different emulated software environments, and the AMIA panel provided an opportunity to examine this approach alongside other recent projects using software emulation, including computer games within the Timothy Leary Papers that are available to play at the New York Public Library Brooke Russell Astor Reading Room, Cornell University Library’s Preservation and Access Frameworks for Digital Art Objects project (PAFDAO), and the Jeremy Blake Papers at New York University. In my workshop presentation I discussed Rhizome’s collaborative development with bwFLA of a cloud-based EaaS implementation allowing users to access artworks via Rhizome’s emulator site, which can also be embedded into other websites such as blogs and social media sites. Since April 2015, Rhizome’s online EaaS implementation has allowed visitors to immerse themselves in the interactive narratives and richly detailed audio and graphics of three video games/artworks created as CD-ROMs in the 1990s by the artist Theresa Duncan.
Even with such promising technical developments for emulation as an access strategy, however, at present intellectual property (IP) rights and software licensing issues remain as primary obstacles to widespread implementation. For example, one of the case studies I examined, Cochrane’s August 2014 post in The Signal analyzing the EaaS implementation at Yale University Library, described three compelling use-case scenarios for emulation across different research contexts, but noted that even given the potential usefulness of EaaS for each project, licensing issues ultimately prevented researchers from actually using EaaS beyond their initial research stages.
Meanwhile, examining and addressing the complex challenges of software IP rights has been included as a key part of the work of the Software Preservation Network (SPN), one of the national digital platform priority projects among nine recently selected for funding by the IMLS. SPN’s project proposal (PDF) includes “articulating a business model and organizational structure through which software licensing and legitimate, ongoing non-profit re-use can operate harmoniously” as a critical step toward preserving software, and therefore a core practice within digital preservation.
To learn more about IP rights issues related to software emulation, I reached out to Euan Cochrane, and to SPN’s co-PIs: Zach Vowell, Digital Archivist at California Polytechnic State University, and Jessica Meyerson, Digital Archivist at the University of Texas at Austin, with a few questions about broad challenges within software IP rights, specific licensing scenarios they are addressing, and potential future licensing models. Their observations shed light on the range of avenues they are pursuing to further establish emulation as a viable and practical method for enabling researchers’ direct interaction with valuable software-dependent content that is locked inside legacy media or inaccessible file formats.
Morgan: The SPN’s IMLS project summary cites a 2014-15 survey of archives professionals in which 51% of survey participants identified access to licenses as an obstacle for emulated access to born-digital materials. Could you share your observations about the implications of these findings, and what approaches SPN may take to address some of these challenges?
Jessica: At the most basic level, that survey response pattern indicates that even in a scenario where staff resources and technical constraints are not considered the most significant barrier, organizations are still hesitant to endorse emulation as an institutional preservation and access strategy unless they have some way to reconcile the software licensing requirements. With no systematic software collection program, most institutions do not currently know how they would access the software titles required to provide emulated access. And in terms of systematic software collection, no single institution can possibly collect all of the software titles needed to render/access the digital objects in its collections present and future.
So, unless we want restrictive software licensing to play a crucial role in appraisal and collection development decisions moving forward – as a community we might think about systematic software collection in terms of documentation strategy. In other words, art and design organizations could focus on creative design software such as Adobe Creative Suite and business history focused collections might focus on accounting software such as Quickbooks or business intelligence software such as Tableau. Meanwhile, digital objects created in software only available as a cloud-hosted service will likely require participation from the software industry.
Morgan: How will SPN address software licensing as it relates specifically to emulation? For example, what priority might be given to the creation of allowances for emulation use-cases?
Zach: Academic licensing may provide some protection from infringement claims, but the extent of that protection remains to be seen. (That is, is it good enough for an institution to show a receipt for its license purchase and that covers all software under its care?). Regardless, licenses will need to be well documented and linked to the license keys used to activate the software.
Ideally, we want to establish a licensing framework that directly addresses the emulation use-cases that cultural heritage organizations are facing now (and in the near future). Pursuing such a framework is a high priority for our project, and hopefully, our project’s research can develop the literature on use cases emulation and other born-digital access strategies. Of course, establishing such a framework presupposes a working relationship with software rights holders — to that end, we’ve been discussing three distinct licensing scenarios:
- rights holders are interested in collaborating on a new licensing framework
- rights holders are not interested in collaborating on a new licensing framework
- orphan works
The SPN project will analyze and document the liability to copyright infringement present in each scenario — and potentially produce best practices based on that analysis.
Morgan: Euan, at Yale, you are working to implement the EaaS framework to enable researchers’ access to digitized materials from the libraries’ collections. For example, you’ve described one project that aims to increase overall use of the wealth of (currently underused) digital materials stored on floppy disks and CD-ROMs in Yale’s general collections by making these materials directly available to users on library computers running legacy software via the EaaS framework. What kinds of licensing issues have you discovered in the course of this process, and how are you addressing them?
Euan: Since writing the Signal article, we’ve discovered that existing agreements with Microsoft allow us to use their older software in many contexts, which is great. But, while very useful, that leaves many thousands of other products that we don’t yet have any solution for; and furthermore, our existing agreement with Microsoft may need to be updated to take into account uses that aren’t covered by it. We will implement EaaS into a production environment in coming months, and start using it with the disks that are clearly allowable within that purpose and which only require either Linux or Windows to be accessed (within the terms of our current Microsoft agreement).
To make digitized content from the general collections available via the EaaS framework, we need the following:
- Either: the ability to restrict access to only as many users as we have copies of the software—something the bwFLA team have been working on recently; or: permission to make the content available to as many people as want it, regardless of how many copies we have in our possession/own.
- Either: an understanding of any end-user license agreements (EULAs) associated with each item and what they allow us to do (where applicable); or: permission to ignore the EULAs; or: permission to use the items in this context, even when it is not mentioned in the EULAs.
- Permission to make enough copies of the data to enable this to technically happen and to properly preserve the bits.
- Sufficient software licenses to support concurrent users of the software needed to interact with the content on the disk(/c)s.
Morgan: Could you describe any licensing models you’re looking at which may provide effective approaches to these issues?
Euan: Tools like EaaS and cloud gaming enable software to be treated as a performance, and as a service much like music is treated by streaming music services. As such, we could address licensing by establishing similar organizations to manage it as exist for music. i.e. performance rights organizations (PROs). So my suggestion is that we set up such an organization for licensing old software.
Some of the significant benefits to a performance rights model for software would be:
- Memory institutions could archive software and only ever pay anything (if anything) when it is used, not just for archiving it.
- Appropriate (e.g. academic/non-profit) organizations might be provided with royalty-free licenses to older software, but these could be managed/validated by the PRO.
- Commercial services could go to the PRO for licenses for commercial uses.
- PROs would alleviate the current inefficient situation of organizations issuing separate access requests from each vendor for a needed software product.
- PROs could deal with transnational licensing. For example, a single PRO could enable French-licensed software to be made available in other countries for use in accessing content produced in France using that software.
Morgan: Jessica, what kinds of licensing models is SPN exploring? I’m also curious about what you’re learning about how these issues factor into the use of emulation for creating “virtual reading rooms” for digital materials, and what kinds of models might inform the future development of software emulation. Are there examples of successful current implementations that you’re looking at?
In terms of licensing models, one that has come up in our conversations (and during the 2014 SAA session) was the performing rights licensing model. Again, that model presupposes that copyright holders are willing to opt-in and embrace the model.
In terms of virtual reading rooms, Dragan Espenschied’s work at Rhizome with Theresa Duncan’s Zero Zero is one exciting example of current emulation projects. The 2014-2015 Born Digital Access Survey (referenced in the SPN IMLS project narrative) investigated current efforts to provide access to born-digital collection material and interview questions specifically addressed any attempts by participants to experiment with emulation as a preservation/access method in their institutions—that data set should be available to the community in the spring. Others are working on similar research—but so far, the general impression is that institutions currently engaged in emulation projects are taking a calculated risk by relying on the section 108 fair use exemption for libraries and archives. However, we can only apply that logic to physical installation media (either already in our collections or available by some other means, i.e., purchase or donation). Again, in the current era of cloud-based subscription software we can apply fair use, procurement of the actual software (once the individual or institutional subscription has lapsed) is introduced as part of the larger challenge.
This underscores the need for the SPN project—to serve as a mechanism for the coordination of existing but disparate collection efforts, as well as a vehicle for collective licensing negotiations. SPN’s emphasis is distinctive but complementary to the UNESCO Memory of the World PERSIST project, which is working to address the broader challenge of “sustainability for the information society” by focusing on several key areas, including the development of technical infrastructure in the form of the Software Heritage Platform (PDF).