Despite decades of effort, digital scholarly editions (DSEs) and digital humanities research software suffer from profound sustainability problems that remain largely unresolved. Empirical studies show that DSEs have short lifespans and high obsolescence rates: out of 466 tracked editions, 376 disappeared, and the average lifespan is only 8.5 years (Neuefeind et al. 2020, Helling, Schildkamp, and Mathiak 2019). Projects frequently vanish when funding ends, institutions shift priorities, or personnel leave; these outcomes show that even robust initial infrastructure cannot ensure longevity. The maintenance burden imposed by continual changes in technologies, standards, and interfaces often exceeds what scholarly teams can provide (Viglianti and del Rio Riande 2025; Boot and Zundert van 2011). Many textual scholars do not possess, and arguably should not be expected to possess, the requisite software engineering and infrastructure skills to sustain functional lifetimes (Pierazzo 2019; Smithies et al. 2019). Graphical user interfaces are costly to build and maintain, yet arguably they represent more often lessons learned than enduring scholarly value; and their obsolescence rate is frightening (cf. Andrews and Van Zundert 2018 and Turska, Cummings, and Rahtz 2016 for various perspectives). Numerous editions rely on project-specific, experimental stacks with inadequate plans for maintenance or handover, and coordination between editorial teams and infrastructure providers is weak. Scholars note a “digital entropy” that increases the risk of decay and abandonment from launch (Viglianti and del Rio Riande 2025).

This fragility contrasts starkly with the durability of codices. Medieval manuscripts, despite centuries of wear, often survive intact and legible; survival ratios for medieval textual traditions in comparison to DSEs are remarkably high (Kestemont et al. 2022), while digital editions fail at astonishing rates in under a decade. The superior durability of codices is rooted in specific aspects of their materiality — pigment on paper has projected longevity on the order of millennia, whereas the lifespans of digital objects are measured in years (Bollacker 2010).

The prevailing discourse in digital textual scholarship treats sustainability primarily as an infrastructural and procedural problem, advocating for community standards, institutional solutions (cf. e.g. Edmond and Morselli 2020; Pierazzo 2019), FAIR principles (Tucker 2022), and software/data good practices such as those promoted by the Endings Project (The Endings Project Team 2023). But these proposals generally assume the dependable continuity of the very infrastructures meant to host data and software. In reality, research infrastructures — national platforms, academic clouds, discipline-specific services, etc. — depend on volatile funding and institutional backing, making them vulnerable themselves (Eghbal 2016). Even funding bodies acknowledge the difficulty of ensuring a financially sustainable portfolio of research infrastructures (Van den Berg et al. 2023). As a result, warnings about a looming “digital dark age” should not be dismissed (Jeffrey 2012; Ghosh 2015; Migowski 2016). Archiving born-digital resources is extraordinarily difficult precisely because their platforms and dependencies change or disappear so quickly (Nielsen 2016, Teszelsky 2019).

Durability and longevity should be an urgent research problem for a field that is preoccupied with the cultural memory and textual archive of humanity (McGann 2015). Philologists should not assume future technologies will provide solutions. To table the urgency of the problem we introduce QUEDEE (pronounced between “Q.E.D.” and “cutie”), a long-term, extremely low-cost, maker/tinkerer-oriented experiment designed to foreground and investigate the durability needs of digital scholarly editions. The central question is whether we can design a digital scholarly edition that plausibly rivals the durability of print editions.

The existing literature contains calls for minimal computing and open, baseline web standards (HTML, CSS, JavaScript), emphasizing static, framework-free editions as a sustainability measure (Risam and Gil 2022, GO::DH Minimal Computing Working Group 2015, Holmes and Takeda 2018, Viglianti and del Rio Riande 2025; Wikle and Williamson 2025; Moran 2025). While this approach makes sense for software longevity, there has been relatively little concrete, hardware-oriented work on fully self-contained editions. One earlier effort deployed a digital library on an Apple iPod, but the closed nature of the platform prevented open sourcing and long-term viability (Bainbridge et al. 2008). We argue that true durability demands self-contained solutions that emulate the codex’s characteristics: they should be self-reliant, portable, and physically similar in size to a book. The project also commits to widely endorsed good practices—open standards for data, extensive documentation, version control and continuous integration, independence from server-side software, avoidance of short-lived frameworks and libraries, massive redundancy, and graceful failure.

The concepts of QUEDEE's design have been presented at the DHd2026 in Vienna (Van Zundert, 2026). Currently QUEDEE's first implementation iteration has been finalized. We will use the occasion of the presentation at DH2026 as a moment to unbox and demonstrate QUEDEE's first material instantiation. The device uses a Raspberry Pi as the hardware backbone because it is open, affordable, and well documented. It features a 480×320 pixel touchscreen and runs Debian 12 based Raspberry Pi's OS “Bookworm” (“Documentation” 2025). On boot, it launches the NGINX lightweight web server (Young 2024) that serves several digital editions viewable through a Chromium browser (Li et al. 2025). The device also starts an open, stand-alone Wi-Fi network, allowing external devices to connect and browse the edition without relying on the touchscreen. Although the software will be archived online, the intent is that once installed, neither hardware nor software changes; this fixity is meant to reduce future breakage from evolving dependencies. The unit is housed in a codex-like case that contains not only the device but also a manual explaining setup, use, and the project’s rationale, underscoring the aim of mimicking a print publication.

Longevity is pursued through redundancy and graceful fallbacks. The touchscreen and the Wi-Fi network provide alternative access modes. Because the Raspberry Pi relies on a solid-state memory card, multiple duplicate cards are included in the case, so a corrupted medium can be replaced. To account for environments without stable electricity, the case includes two solar panels and a solar power manager to regulate input; this allows off-grid use. All documentation and data are stored redundantly on the device in plain text formats to maximize future readability. As a last resort, the case also includes two stacks of minified facsimiles of the editions and their primary source, so in case even if all hardware and software fail, a user still has access to the scholarly content.

Currently one digital edition is hosted on QUEDEE. A static, HTML/JavaScript-based scholarly edition of the Middle Dutch classic “Van den Vos Reynaerde” (Bouwman and Besamusca 2009). Eventually we want to host multiple SDEs on the device, demonstrating the ability of the device to act as a digital library and archive, as well as a reading device. For this step we are targeting a TEI-XML based edition of the works of Anne Frank (De Bruijn, Bleeker, and Scherer 2024) and the Online Variorum of Darwin's Origin of Species (Bordalejo 2025). To preempt any potential copyright infringements, the editions only consist of the open source data of the originating projects that is available in the public domain.

In collaboration with the National Library of the Netherlands possibilities will be explored to keep the device in realistic archival conditions and retrieving it for evaluation after 1, 2, 5, and 10 years. Each retrieval would assess functionality and the viability of fallbacks, with results and any design revisions published over time. This longitudinal, empirical testing is meant to move the conversation beyond theoretical prescriptions to practical evidence about what actually endures.

The current implementation has several known flaws. Solid-state memory may suffer from data retention decay, with anecdotal survivability ranging from a few years to over a decade. Additionally, including batteries in the case raises preservation concerns as libraries obviously generally avoid storing chemical batteries with collections. However, we argue that any future user capable of understanding the device would likely be able to supply or fabricate a compatible battery. These limitations do not negate the project’s value; rather, they highlight the difficulty and necessity of addressing hardware and energy dependencies directly when designing for long-term survival.

The design and implementation of QUEDEE also raises critical questions about knowledge transfer by digital humanities concerning digital infrastructure. Longevity and durability are strongly determined by materiality. While materiality is a key aspect of textual scholarship when it comes to the sources studied, the materiality of digital infrastructures that textual scholars rely upon, the technical knowledge related to that, and consequences for durability are scarcely discussed in depth as part of this knowledge transfer.

In conclusion, our paper contends that DSE sustainability is in crisis and that current strategies underestimate the fragility of digital infrastructures. By proposing a self-contained, codex-like, standards-based device with multiple fallbacks and a concrete testing regimen, QUEDEE aims to reframe durability as a design goal and research agenda. The project is offered as both a practical prototype and a request for comments: an invitation to the community to engage in earnest with the material, organizational, and technical realities that determine whether digital scholarly editions will be accessible decades from now.