Founders Online: Early Access: Reflections on Open Access, Crowd Sourcing, and Metadata Standards

Founders Online, a digital initiative of the National Historical Publications and Records Commission (NHPRC) of the U.S. National Archives, launched in June 2013. Since its debut, the site has attracted over a million visitors interested in learning more about the creation of the United States of America in the words of six of its Founding Fathers. Founders Online contains 177,000 letters or other writings of these men and their contemporaries. Widely used by academics and the general public, the site has demonstrated the value of digital humanities’ emphasis on free access. As a former assistant editor at Documents Compass, a program of the Virginia Foundation of the Humanities, I served as a project manager on the Early Access portion of the project. We worked directly with the staffs of the currently active Founding Fathers documentary editing projects to make preliminary versions of unpublished documents available for early viewing on Founders Online. These Early Access documents will eventually be replaced by fully vetted and annotated versions to be completed later by the documentary editing projects. Relying on a large staff of over thirty people, we transcribed or proofread over 50,000 Early Access documents from 2012 to 2015. My Early Access experience demonstrated the need to give employees constant feedback, to reward them for good work, and to encourage specialization among project staff. My experience also reemphasized the need for unified metadata standards when aggregating different sets of data from multiple projects into a single digital platform. Founders Online, une initiative numerique de la National Historical Publications and Records Commission (NHPRC) des Archives nationales des Etats-Unis, lancee en juin 2013. Depuis ses debuts, le site a attire plus d’un million de visiteurs interesses a en apprendre davantage au sujet de la creation des Etats-Unis d’Amerique d’apres six des peres fondateurs. Founders Online contient 177,000 lettres ou autres ecrits de ces hommes et de leurs contemporains. Largement utilise par les universitaires et le public en general, le site a demontre la valeur de l’emphase des humanites numeriques sur le libre acces. En tant qu’ancien redacteur en chef adjoint a Documents Compass, un programme de la Virginia Foundation of the Humanities, j’ai travaille comme gestionnaire de projet pour la partie d’acces anticipe du projet. Nous avons travaille directement avec les membres du personnel des projets de montage documentaire de Founding Fathers actifs a l’heure actuelle, pour rendre disponibles en acces anticipe des versions preliminaires de documents non publies sur Founders Online. Ces documents en acces anticipe seront eventuellement remplaces par des versions entierement approuvees et annotees qui seront completees plus tard par les projets de montage documentaire. Comptant sur un personnel nombreux de plus de trente personnes, nous avons transcrit ou relu plus de 50,000 documents d’acces anticipe entre 2012 et 2015. Mon experience de l’acces anticipe a demontre le besoin de donner aux employes une retroaction constante, de les recompenser pour leur bon travail, et d’encourager la specialisation parmi le personnel du projet. Mon experience a de plus souligne davantage le besoin de normes de metadonnees communes en transposant differents ensembles de donnees de projets multiples en une plateforme numerique unique. Mots-cles: Founders Online; Libre acces; transcription; metadonnees; externalisation a grande echelle; Histoire numerique

Founders Online, a digital initiative of the National Historical Publications and Records Commission (NHPRC) of the U.S. National Archives, launched in June 2013. Since its debut, the site has attracted over a million visitors interested in learning more about the creation of the United States of America in the words of six of its Founding Fathers. Founders Online contains 177,000 letters or other writings of these men and their contemporaries. Widely used by academics and the general public, the site has demonstrated the value of digital humanities' emphasis on free access. As a former assistant editor at Documents Compass, a program of the Virginia Foundation of the Humanities, I served as a project manager on the Early Access portion of the project. We worked directly with the staffs of the currently active Founding Fathers documentary editing projects to make preliminary versions of unpublished documents available for early viewing on Founders Online. These Early Access documents will eventually be replaced by fully vetted and annotated versions to be completed later by the documentary editing projects. Relying on a large staff of over thirty people, we transcribed or proofread over 50,000 Early Access documents from 2012 to 2015. My Early Access experience demonstrated the need to give employees constant feedback, to reward them for good work, and to encourage specialization among project staff. My experience also reemphasized the need for unified metadata standards when aggregating different sets of data from multiple projects into a single digital platform.
Founders Online contient 177,000 lettres ou autres écrits de ces hommes et de leurs contemporains. Largement utilisé par les universitaires et le public en général, le site a démontré la valeur de l'emphase des humanités numériques sur le libre accès. En tant qu'ancien rédacteur en chef adjoint à Documents Compass, un programme de la Virginia Foundation of the Humanities, j'ai travaillé comme gestionnaire de projet pour la partie d'accès anticipé du projet. Nous avons travaillé directement avec les membres du personnel des projets de montage documentaire de Founding Fathers actifs à l'heure actuelle, pour rendre disponibles en accès anticipé des versions préliminaires de documents non publiés sur Founders Online. Ces documents en accès anticipé seront éventuellement remplacés par des versions entièrement approuvées et annotées qui seront complétées plus tard par les projets de montage documentaire. Comptant sur un personnel nombreux de plus de trente personnes, nous avons transcrit ou relu plus de 50,000 documents d'accès anticipé entre 2012 et 2015. Mon expérience de l'accès anticipé a démontré le besoin de donner aux employés une rétroaction constante, de les récompenser pour leur bon travail, et d'encourager la spécialisation parmi le personnel du projet. Mon expérience a de plus souligné davantage le besoin de normes de métadonnées communes en transposant différents ensembles de données de projets multiples en une plateforme numérique unique. letters or other writings of these men and their contemporaries who corresponded, lived, and worked with them, the site utilizes extensible markup language (XML) encoded documents hosted on a MarkLogic server to provide free, fast, and userfriendly access to the public. Founders Online has received praise from the larger scholarly community, including the Society for History in the Federal Government, Kurtz: Founders Online 3 which awarded the site its prestigious Thomas Jefferson Prize for its "outstanding contribution" to scholarship (Ferriero 2015b). 1 This emphasis on reaching a much

Digital History Meets Documentary Editing
The appeal of publishing humanities work digitally cuts across disciplines and has increasingly become a priority for faculty hiring, cultural institutions hoping to celebrate and grant greater access to their holdings, and for scholars seeking to reach 1 For more information on the technical infrastructure that hosts Founders Online, please see Matt Allen, "Founder's Online: A Lesson in Performance" (Allen 2016) and "Founders Online Launches" (UVA 2013).

Kurtz: Founders Online 4
new audiences. In the field of history, the importance of writing and producing historical scholarship accessible beyond the walls of academia has become increasingly important even for those who do not consider themselves strictly public historians. In an era of decreasing faculty and research budgets for the humanities, proving that the study of history is valued by the American public is vital for history departments, universities, museums, and other historical entities that rely on public monies to cover their expenses. Certainly projects such as the University of The tireless efforts of project editors and staff to provide accurate and scholarly editions of these men's writing, however, are both time consuming and costly. In addition, the high cost of publishing their works in print ensures that only relatively few volumes are produced on a yearly basis. This means that the American public, which has demonstrated time and again its interest in the story of the creation of its nation, has limited opportunities to benefit from this important work. While the most important editions did start to go online in the early 21st century, these projects, such as University of Virginia Press's subscription-based service Founding Era Collection, often existed behind a paywall that limited their use to universities able to pay for access (UVaP Founding Era Collection 2017). website. These students demonstrated its use to an audience of editors, archivists, politicians, and journalists. Since its launch, the site has attracted over one million visitors both inside and outside of the academic and scholarly history communities (National Archives, 2010, 2011, and 2013a. Given that our staff was largely inexperienced in editing or historical work, we designed our processes to help ease employees through the various technical aspects of their work. For example, few of them had any experience with editing in XML.

A Digital Edition Success Story: Founders Online
We realized that we needed a user-friendly XML software such as Xmetal, a program whose interface closely resembles that of Microsoft Word. We found that the more popular and familiar program, Oxygen, even in its out-of-the-box "Author" view, was too technical and difficult for most of our staff. We also had to train our employees in the unique style guides employed by each project. We created instructions and processes that were simple and understandable in order to produce the best quality transcripts possible while adhering as closely as possible to project specific requirements. Despite our limited time and ambitious goals, varying accessibility of manuscript sources necessary for transcription and proofreading, and the different requirements and methods of each project, 4 we believe that by the end of our project our employees had improved their speed and accuracy considerably. Our proofreaders benefited from regular in-person and online feedback about their work from our project managers, and, as they gained experience, they became more adept at deciphering late 18th and early 19th century handwriting. They mastered the basics of XML editing and utilized historical techniques such as using other letters to decipher difficult words or using online resources to look up basic information about people, historic places, and past events. Ultimately, our staff produced 3 Here, I would like to recognize Early Access team members Mark Hawking, Jeffrey Diehm, Dena Radley, Jeffrey Zvengrowski, and James Ambuske for their exceptional contributions to the project.
Several of our proofreaders were so moved by their experience that they took part in a special episode of With Good Reason, a public radio show featured locally on National Public Radio (NPR), in order to tell the wider community of their newfound appreciation for early American history (With Good Reason 2014). 4 For example, our first goal was to proofread and publish 10,000 letters from the Papers of George Washington and Papers of James Madison projects. Both projects have unique style guides of differing complexity and specificity about how to handle superscripts, double punctuation, capitalization, etc. The need to train employees simultaneously in two different style guides greatly increased the difficulty of our work. Kurtz: Founders Online 9 preliminary transcripts that will be useful to Founders Online's users as they wait for the documentary editing projects to complete their work.

Lessons from Early Access
My experience in Founders Online: Early Access not only convinced me of the importance of taking into account a larger, non-academic audience in designing digital humanities projects, but I also believe it has provided several important lessons for digital humanists seeking to use crowd-sourcing, non-specialist volunteer labor as part of their workflow. It is unlikely that a similar project to Early Access with a large, paid staff will happen to be funded again anytime soon. Still, despite its use of paid employees, my experience as an Early Access project manager speaks to the need for constant feedback to project participants and methods of quality control for crowd-sourced projects relying on voluntary and free labor. Projects such as the Library of Virginia's Virginia Memory do provide a method of limited peer review, but a more robust method of communication between the editorial staff and volunteers, in my experience, is necessary. 5 It is not always helpful to create a one-size list of "do's and don'ts" for participants, for different volunteers will need more instruction or correction on different points of transcription. We found that constant individual feedback in person, through email, or via our project management software called Basecamp was the best way to ensure the best transcription and proofreading results possible.
I also believe projects need a reward mechanism to acknowledge their best contributors to keep them engaged and motivated to continue working through to a project's completion. Transcribe Bentham, a very successful crowd-sourced project

Kurtz: Founders Online 10
users by the number of images transcribed, or by the speed or accuracy of their work.
Needless to say when your labor force is working for free it is very important that top performers are encouraged and recognized for their exemplary contributions (see "About Us" and "Hall of Fame," Transcribe Bentham 2016a and 2016b).
We achieved this in Early Access primarily by giving our employees raises, an option admittedly not applicable to crowd-sourced projects. However, we also found non-monetary ways to reward employees. For example, I rewarded my best employees by assigning them particularly interesting writers and batches of correspondence such as Thomas Jefferson's letters with John and Abigail Adams.
I also made a point of acknowledging our best employees' contributions publicly through Basecamp, or praising them during our training meetings. Many similar opportunities for rewarding excellent work exist for crowd-sourced projects, whether that is using the principles of gamification as seen on popular language-learning programs like Duolingo or the digital crowd sourcing project Old Weather in which the best users are rewarded for their good work with increasingly high nautical ranks for transcribing weather logs from old ship manifests. Highlighting the names of individuals who made significant contributions to the project directly on its website in this way is very important. David Rumsey's listing of his top ten geo-coders on his Maps Collection site is a perfect example (Rumsey 2017).
Finally, it is a very good idea to encourage specialization. Particular writers' handwriting or some data sets may be harder to transcribe or record than others.
Instead of letting anyone edit or transcribe a document as is done on some Omekabased sites, volunteer transcribers should have to demonstrate a proficiency in deciphering the difficult handwriting of such historical figures as Abigail Adams or James Monroe before getting full access to these writers' letters. Not only does this ensure a higher quality of work in the finished product, but we also found that giving access to this important yet difficult subset of letter writers was a useful non-monetary reward that could be copied by others relying on non-expert, volunteer labor. Founders Online also demonstrates the difficulties inherent in bringing different projects together under one roof that follow different style guides, metadata conventions, and methodology. Differences across projects are inevitable confusing to users not familiar with the methods of documentary editing. For example, instead of using a single standard convention to indicate that a word or phrase is difficult to read (e.g. surrounding the word with [brackets]), there were several different ways to represent this common editorial issue across different projects. For better or worse, Founders Online did not attempt to standardize these conventions. Similarly, even though the same historical figures appear in different documentary editing projects, the projects have not agreed on a standardized way to spell their names. Hence the Marquis de Lafayette's long French name appears differently across projects, making a single search for his letters on Founders Online impossible. While adopting a standard methodology or style guide may prove impossible, the adoption of a names authority lists, perhaps by using the Library of Congress's authority files and adding new names as necessary, would help users. Another Documents Compass project titled People of the Founding Era, a collective biography of the 18th century U.S., has done considerable work to reconcile different naming conventions across projects, and thus might serve as another potential starting point for standardizing metadata across digital projects focused on Early American history.

Conclusion
Founders Online: Early Access was a complex, digital project with a large paid staff that was finished at the end of 2015. Early Access's successful completion provides a useful model for dealing with the technical issues of large-scale digital humanities work. It shows the need for rewarding hard-working, accurate volunteers while also imposing uniform standards across disparate sets of data. Ultimately, Founders Online's continued importance will be its emphasis on free access and user friendliness to professional historians as well as the general public. Its emphasis on digital methods of access will prove useful to more traditional scholars studying larger topics of interest to the public, while its emphasis on being useful to the public and thinking about the non-scholarly world will help make digital history even more appealing to potential funders. Such projects should be encouraged in the future, whether they are similar collections of freely available transcriptions related to a particular topic in American history such as the American Civil War, the