<p>Many digital humanities projects are developed by organizations without large budgets, using funds from a "start-up grant". Consequently, tools and projects fail to be maintained after the funding runs out, and may eventually be taken down, although university libraries will sometimes step in to provide long-term server space. One workshop participant prioritized data preservation over tool preservation, on the grounds that tools are soon supplanted but the data created through the use of tools can be reused. Regardless of how one attempts to limit the scope of what must be preserved, sustainability is a daunting problem that still lacks a feasible solution: "how is it economically viable to permanently store data on a server for anyone anywhere?" (W2, Analyzing Directions, Group B.)</p> <p>Workshop participants saw two potential approaches that a consortium could take to ensure project sustainability. One proposal is analogous to libraries' efforts to preserve the printed record: the consortium could build the infrastructure necessary to curate digital tools and resources, and take on that responsibility for member institutions. The service-oriented approach would run tools rather than preserve them, and swap them out with new ones when they become obsolete. In such a model, institutions (or perhaps another consortium) might be responsible for curating the data, and tools would be treated as essentially disposable-- much to the chagrin of those who would like to see them preserved as objects of scholarship.</p> <p>The question of sustainability affects all tools and projects, regardless of their success. Arguably, the projects that feel it most acutely are those that fail to live up to the hopes and expectations of their PI, their funder, or their community. These tools and projects may never have resonated with their intended audience, and sit unused even before the technology behind them is obsolete. Workshop participants remarked that valuable lessons could be learned from projects that failed at any point in their lifecycle-- from projects that were planned but were unable to be implemented, to projects that were completed but withered earlier than anticipated. Mellon's lack of transparency in not posting projects that did not succeed was criticized. A "journal of digital failures" (<a href="http://www3.isrl.illinois.edu/~unsworth/">John Unsworth</a>, Ex 6b, 1d-D) where one could post write-ups or final reports from failed projects, was proposed as one potential step towards increasing transparency and salvaging what one can from failure.</p>