Leveraging Student Expertise and Learning to Produce a Sustainable Digital Collection of Music Composition Materials

Olivia M. Wikle, Iowa State University

Liam R. Marchant, DePaul University

A librarian-student collaboration benefits from a student’s subject expertise and willingness to learn data and web development skills via the creation of a static web digital collection of music composition material.

Introduction

During their composition process, composers often create nascent sketches of their musical ideas. These sketches and other draft material can be incredibly valuable to researchers, performers, and novice composers wishing to understand a composition or emulate the composer’s process. In most cases, however, the sketches are never published. Instead, they may be discarded or lost, or stored in the composer’s private archive. In the instance that they are preserved in an archival institution, their use is often limited to those who have the means to travel to the institution to view them physically, as copyright restrictions and questions of organization can stymie initiatives to digitize and publish them.

In 2020, University of Idaho (U of I) faculty member Daniel Bukvich, a composer and professor of percussion and theory at the U of I’s Lionel Hampton School of Music, donated a collection of his own sketches, handwritten and printed scores, and auxiliary material to the U of I Special Collections and Archives (Spec), located within the U of I Library. This donation, accompanied by Bukvich’s interest in providing context around his work and making the collection items publicly accessible, led to a rare opportunity for the U of I Library’s Digital Initiatives unit to organize, digitize, describe, preserve, and publish these items as the Daniel Bukvich Composition Digital Collection.

This process was undertaken in spring 2022 by U of I Digital Initiatives Librarian Olivia Wikle and Liam Marchant, a fourth year music composition student and Library student employee. Having found little external precedent for organizing, naming, describing, and publishing composition material such as sketches in digital collections, we relied on our own expertise (Marchant in music composition, Wikle in digital collections development) to design a project workflow and digital collection site using the static web digital collections framework CollectionBuilder. Our process privileges the differences between composition components such as sketches, scores, and parts via the use of custom file naming standards, controlled vocabulary, robust metadata, and engaging contextual content on the digital collection site. The resulting project demonstrates how an agile static web approach can enable student/faculty digital scholarship collaborations that are not only technically sustainable but also educationally-focused, providing students the opportunity to contribute their expertise to produce a tangible output while also learning fundamental technical skills that are broadly applicable to their future endeavors.

Background

The University of Idaho is a land-grant university located in Moscow, Idaho. The U of I Library’s Special Collections and Archives department specializes in the acquisition and preservation of archival material that documents the history and culture of Idaho and the University of Idaho, while the Library’s Digital Initiatives unit is engaged in digitizing, preserving, and disseminating that same material via digital collections and digital scholarship projects. The U of I’s Lionel Hampton School of Music (LHSoM) is an accredited institutional member of the National Association of Schools of Music, and includes undergraduate degrees in music education, music business, performance, and composition. Professor Bukvich is an active and well-respected composer: his musical compositions and arrangements are regularly performed by orchestras, choirs, bands, soloists, chamber groups, and jazz groups around the world. His composition papers include sketches, handwritten and printed scores and parts, and auxiliary contextual material related to his pieces’ performance.

In most cases, musicians and listeners come to know the work of a composer via their finalized, published works. The associated materials that were created in the process of drafting those final versions typically are not immediately available to the public upon the work’s publication, a tradition which masks and separates the composition process from the output itself. Yet these informal documents can provide insight for musicologists or performers wishing to interpret a piece or its performance, and can be powerful pedagogical tools for novice composers who are learning the trade.

Sketches of musical ideas, especially, serve a variety of purposes within the composition process, and have been preserved for almost every piece in the Daniel Bukvich Composition Collection. Bukvich himself describes his sketches as essential to his composition process: he sees the craft of composition similar to practicing an instrument; as he states, “if you practice your instrument in a variety of styles, keys, and environments, why not do the same for composition?” (Bukvich 2022). In this way, sketches can be perceived as small exercises or theoretical experiments exploring musical ideas which allow a composer to find their own style. The sketching exercise offers composers an alternative to composing exclusively by ear, a process which, subconsciously or not, involves copying and incorporating music they have heard before (Bukvich 2022). Instead, when sketching, a composer creates a “laboratory” of musical ideas that can be used to guide where the music should go without copying too closely the music of others. Sketching also allows for the composer to realize their vision of what a piece can be before committing to integrating more time-intensive aspects of the compositional process such as orchestrating or part writing. Because of Bukvich’s own reliance and emphasis on the importance of sketching, and the rarity with which sketches are published in general, we chose to foreground the sketches as a central part of this digital collection.

Despite the benefits they can provide to research, teaching, and learning, efforts by libraries and archival institutions to digitize and publish sketches and other composition materials can be thwarted by copyright complexities or by the time and expertise required to identify, organize, and describe the wide assortment of material associated with the composition process. Furthermore, those digital collections related to compositions and composers that do exist place an admirable focus on describing the context around a composer’s life and compositions, but rarely make the digitized composition components available for download in full.

At U of I, we were presented with a unique opportunity to overcome these obstacles: in 2022 Bukvich gave us permission to publish and make available for download the material associated with his original compositions. Additionally, as a seasoned Digital Initiatives student employee and composition student, Marchant had the musical and technical knowledge to interpret and organize the material, as well as a personal connection with Bukvich to obtain item-level metadata and contextual information directly from him when clarification was needed. With these advantages, our goal when undertaking this project was three-fold: (1) document our process for creating and describing the digital objects of this collection so that our staff can add to it in the future and so that it might serve as an example for other librarians hoping to disseminate composition materials; (2) craft an output that is more than just a database, working with the composer to creatively frame the content while making the collection data openly available to promote further research; and (3) engage Marchant in learning and practicing technical skills while contributing significantly to a stable website which can be cited on his resumé in the future.

We recognize that we did not face what is often the most difficult obstacle for a project such as this—that is, convincing a composer to make their material freely available online. Bukvich’s work has always contained strong educational ideals, and he sees value in making his compositional materials freely available for future composers and musicians to learn from. However, while we did not have to convince Bukvich, we did need to approach him with the idea for a digital collection and have conversations about how the final product would look and behave. While this project might have been something he thought about in a vague way, he did not have a concrete method to follow nor the skills and resources to execute it that the Library could offer. We think it is likely that there are many composers with a similar interest in teaching who would be open to putting their material online but have not encountered solutions for doing so. When approaching a composer, library, or archive with an idea like this, we suggest communicating that they can be selective about what is included in the collection and how it is accessed. There are many of Bukvich’s materials, for instance, that could not be included in this collection for copyright reasons. If composers understand that they have control over what is shared, they may be more willing to participate in a similar project. We also hope that the Daniel Bukvich Composition Collection website, and the steps to create it outlined below, will lower the barrier for convincing composers to participate in disseminating their compositional materials and ensure that the resulting projects are designed to be useful to a variety of users, including composers, students, musicians, and scholars.

Project Design

Bukvich maintains a reputation for being meticulously organized, a fact which is reflected in his collection of composition materials that he donated to Spec in 2020. Many composers’ collections do not have sketches saved with their works, but Bukvich put much thought into archiving and preserving all of his materials, not just his finalized compositions. His papers are organized in folders according to composition title, and almost every composition folder includes sketches and handwritten and/or printed scores and parts. Additionally, some composition folders contain auxiliary material such as programs or commission letters that offer additional context around the composition and its performance history.

File naming

Because the collection was so meticulously organized, our need to physically organize the items was drastically reduced. Yet we were still faced with the task of categorizing and developing a file naming system for the various material types. This process is more complex than might be assumed at first glance, in part due to the fact that a single composition consists of multiple components in varying formats which often require significant musical expertise to interpret and describe. A composition’s components consist of all materials—physical, digital, handwritten, printed—that were involved in the creation and/or performance of the composition. In our project context we determined that each composition would be represented as a compound object in the digital collection with multiple components populating a single composition’s item page. We are influenced here by Anneli Kivisiv and Kai Kutman’s approach to creating and maintaining the archive of the world-renowned living composer, Arvo Pärt, and their precedent in treating the myriad components of a composition equally (Kivisiv and Kutman 2017). To standardize classification of the composition components, we focused on the purpose of each document, organizing material into the categories Auxiliary, Scores, Sketches, and Parts:

Defining these categories was essential to our establishment of an abbreviation system that would allow us to consistently integrate meaning into our file names, identifying the type of content each file contains. Each of our file names begins with an identifier based on the composition’s title (for example, the identifier for the piece Moving Violation is movviol). Then, following Adriana Cuervo and Eric Harbeson who suggest using a series of abbreviations to indicate scores in finding aids (2011, 52), we developed a controlled set of suffixes based on abbreviations of the categories above that could be appended to our file names to distinguish the type of material in that file. Further suffixes were added to distinguish certain components that may have the same title but consist of a different medium (i.e. a typeset score versus a handwritten one):

Similarly, we created a controlled list of abbreviations to represent instrument names and part numbers. These were used as suffixes in the “part” file names:

Using our file naming system of “identifier_category_instrument_medium,” the file name for the handwritten first alto sax part for Moving Violations is “movviol_part_sx-a01_h.pdf,” while the filename for the handwritten transposed score of the same piece is “movviol_score_t.pdf.” These files have been preserved individually for archival purposes, but for ease of use and display in the digital collection, we ultimately combined access copies of all parts into a single “parts” PDF (i.e. “movviol_parts.pdf”) and scores into a single “scores” PDF (i.e. “movviol_scores.pdf”).

Though we were lucky to start with a relatively organized physical archive when we began the file naming and scanning process, we recognize many composition archives are not this organized upon accession. We think our component categories (Auxiliary, Scores, Sketches, and Parts) are general enough that they are still applicable for guiding the classification of the various material types that tend to be present in composition collections and provide a good starting place for those looking to organize or process a physical archive. For collections of living composers, we recommend obtaining as much contextual information about the collection’s content as possible, since it is likely that historical information can assist in the interpretation of this compositional material. Additionally, enlisting collaborators with musical or compositional backgrounds is imperative for interpreting material in a composition collection such as Bukvich’s since there are often references and technical terms that non-experts may miss.

Scanning

With the file naming system in place, we moved forward with digitization. Most material in this collection was scanned on a Fujitsu Fi-6770 24-bit color document feed scanner to 600 DPI jpeg files. If material was too large to fit the feed scanner, it was digitized on an EPSON Expression 12000XL 48-bit color scanner to 600 DPI jpeg files. Derivative 300 dpi copies were generated from these high-resolution files using PhotoShop. Adobe Acrobat was then used to first convert the low-resolution jpegs to PDF, and then apply Optical Character Recognition (OCR) to capture the documents’ text. We followed U of I’s standard digitization and processing procedures for this workflow, which can be found at https://uidaholib.github.io/digital-collections-docs/. At this point, the high-resolution jpegs, processed low-resolution jpegs, and PDFs were transferred to the U of I Library’s digital archive for preservation. The PDFs were also transferred to the Library’s web server in preparation for inclusion on the digital collection website. Our decision to feature the composition components as PDFs on the digital collection website rather than as multiple image files was largely to facilitate use of the material: whether they are playing, studying, or teaching with these compositions, users are likely to find it easier to download and sort through the material if it is organized into PDFs by component types that are standardized across compositions.

Metadata

Our collection’s metadata can be viewed in CSV format on GitHub, a popular cloud-based version control and code sharing platform. When it came to deriving a metadata schema that would allow us to describe the pieces in our collection while organizing them as a digital collection, the majority of potential examples we found were related to describing physical music archives rather than digital. Additionally, relevant scholarship on composition collections is often directed toward a decidedly research-focused audience and intended to describe a music archive’s content and provenance without much focus on the practicalities of its organization (Hunt 2011; Tsou 2017). Following machine-readable cataloging (MARC) standards and utilizing Library of Congress Subject Headings (LCSH), music catalogers have set guidelines for how to classify components of a composition (Iseminger et al. 2017). These guidelines are not often as granular as archival collections may require for describing the wide-ranging and sometimes surprising formats that these components can take (including sketches, programs, posters, musical diaries, articles, reviews, photographs, etc.). For this reason, we ultimately found that literature primarily focused on archival description and finding aids best informed our work in this area. Because each component holds unique value and purpose, correctly identifying and fully describing these components is imperative if researchers, composers, and musicians are to search for and discover relevant material (Cuervo and Harbeson 2011, 44). When creating our metadata we drew on the categories and organization system described in the File Naming section above as a controlled vocabulary for describing the types of material within each composition.

When establishing metadata fields, we used Dublin Core terms such as title, creator, date, source, format, type, and rights fields to record standard information about each piece. In determining the collection’s custom fields, we made an effort to consider our target audience based on Sean Luyk and Colette Leung’s account of creating metadata fields for inclusion in a sheet music collection, which was influenced by their anticipation of a target audience consisting of “those who will study the music” versus “those who will play it” (2016, 14). We anticipate that both of these audiences could conceivably find value in our digital collection, especially as many collection items contain full scores and parts. Fields we added in anticipation of prospective performers searching the collection include a “tempo” field, as well as “score_order” and “part_order” fields that denote the order of scores and instruments inside a composition’s “score” and “part” PDF files. Additionally, because of the fluidity of many ensemble and instrument names (jazz orchestra vs. big band or contrabass vs. double bass) we constructed a standard list of instrument names and ensemble names that will allow users to perform effective searches across the collection (see Table 7). This standard vocabulary was partially informed by the Getty Art and Architecture Thesaurus; however, due to several discrepancies in the organological organization of instruments, we altered it slightly to better fit our context.

Perhaps the most informative field in the metadata is “historical_background.” This field consists of brief, transcribed narratives about each piece that resulted from a series of interviews that Marchant conducted with Bukvich. These interviews took place after the initial digitization work was finished and provide valuable context around each piece, ranging from details about a piece’s commission to stories around its premier. Marchant’s pre-existing relationship with Bukvich and both parties’ willingness to capture this information by participating in interviews has been key to creating robust metadata. This situation is a unique factor compared to most composition collections, which more often document the work of deceased composers. Here we looked to guidance from Kivisiv and Kutman, who highlight the privilege and importance of communicating with a living composer to better contextualize the items in their archive. They prioritize interviews with Pärt even at the risk of impeding organization or digitization in order to meet their goal of contextualizing the components of each composition in their archive (Kivisiv and Kutman 2017, 159). As we move forward into additional phases of this project, we plan to continue breaking up digitization into manageable chunks of work and interspersing that work with time dedicated to recording interviews with Bukvich.

Documentation

Due to the large number of pieces in Bukvich’s archive, the scope of this project is too extensive to be completed in a single semester and will need to change hands to another student employee in the future. Because of this anticipated staffing shift, we have invested in creating clear documentation to ensure that the process of digitization, file processing, file naming, and metadata creation remains consistent. Part of this process includes documenting our category definitions and file name suffixes, as well as our controlled vocabulary for composition component types, score types, and instrument names. In addition to documentation represented by the categories and file name suffixes listed above, Marchant created the file naming flowchart below, which illustrates our file naming and processing workflow.

Dissemination

We used a customized instance of CollectionBuilder to make the Bukvich compositions publicly available via a digital collection. CollectionBuilder is an open-source, static web digital collections template developed in-house by librarians at the U of I. It uses the static site generator Jekyll to consume collection data and transform a folder of structured source code and data into a complete website, pre-rendering each page as a static asset. Though CollectionBuilder necessitates a steeper learning curve for beginners since it requires navigating a directory of files (typically through an integrated development environment such as Visual Studio Code), creating well-formed metadata CSVs, editing YAML files, and utilizing some command line rather than a GUI interface, there are several advantages to this approach, including sustainability and flexibility when it comes to customizing the site. Additionally, as Marchant’s participation in this project demonstrates, the technical workflows CollectionBuilder requires can be leveraged to create a meaningful and lasting educational experience for student collaborators who are likely to utilize these spreadsheet and data skills in their future careers.

Because CollectionBuilder is not a database to which digital objects are uploaded, it provides a sustainable approach to object creation and dissemination, encouraging the preparation of web-ready digital objects and their metadata in a way that is completely external to the platform itself. Once prepared, a collection’s objects can be placed on a simple web server, on the cloud, or in a content management system such as CONTENTdm. Metadata which points to those objects is prepared in a spreadsheet (usually in Google Sheets), and used by Jekyll to build out a site which automatically produces browse and item pages as well as word cloud, map, and timeline visualizations that provide engaging means of entry for collection users. The site’s static files can then be moved to a preferred web server or hosted for free on GitHub Pages or a similar service. Because it is composed of these static files, the website will remain functional and secure for the long term, making it ideal for academic digital projects like the Bukvich collection which often involve a relatively short period of intense design and development and then, once complete, receive little maintenance due to the realities of limited resources in the academic library. This project format also lends itself easily to preservation: the collection’s digital objects, metadata spreadsheet, template code, and static site files can be easily preserved in a digital archive, and the initial careful preparation of digital objects and metadata in standard formats ensures that they will be easily migratable to a new platform in the future.

We used the CollectionBuilder-CSV template, following the steps in CollectionBuilder’s documentation to produce the Daniel Bukvich Composition Collection. Our careful application of controlled vocabularies and composition dates in our metadata is reflected via CollectionBuilder’s automatic generation of word cloud and timeline visualizations, which allow users to access the site’s content from multiple vantage points. Development of the collection takes place on the code hosting platform GitHub, with Wikle and Marchant both contributing to the Buckvich Project GitHub repository. GitHub made it possible for us to collaborate and iteratively develop the site’s metadata and interpretive content: as Wikle edited the site’s code, Marchant could continue making edits to the metadata spreadsheet in Google Sheets and upload it to the GitHub repository to quickly update the site’s content. This iterative method of design has allowed us to customize the site around what its future user community might need in terms of content and aesthetics, as Brian Rosenblum et al. advise (2015). To accommodate the organization of the Bukvich materials by composition title, Wikle modified the CollectionBuilder template code so that thumbnail representations of each composition’s components are represented on a composition’s item page. A user may click on any thumbnail to view or download the PDF it represents. A “type” label (i.e. sketch, handwritten score, auxiliary material, etc.) is displayed on top of each thumbnail so that users can identify each component. Due to the importance of Bukvich’s sketches outlined above, we made the decision to foreground the sketches in this collection by adjusting the sketch thumbnail to be the largest thumbnail on a composition item page, as is demonstrated on the Harmonie du soir item page.

Finally, we took advantage of CollectionBuilder’s capacity to include interpretive text to create content about Bukvich’s life and composition style. These pages are written in Markdown (a lightweight markup text that can be transformed into HTML and simplifies writing for the web) and use Jekyll’s templating language Liquid to include thumbnails of collection items and external photos and videos that add to the narrative. We expect that future students employed on this project will add to the contextual material in this collection by creating additional interpretive pages investigating inaugural performances, commission history, and more.

Once the project metadata, code, and content were finalized, we moved the Jekyll-generated site files to the Library’s static server, where the site will be hosted long-term. When updates are needed, we push changes to the GitHub repository, then rebuild the site’s files and copy them over to the server again. Because Bukvich is a prolific composer and there are many more works of his to digitize and publicize through the site, we have tried to organize and contextualize items in a way that lays the foundation for what the project may become, while acknowledging we do not fully know what that might be. We are informed here by CollectionBuilder’s reliance and emphasis on the Collections as Data model (Padilla 2017). This model allows us to export the Bukvich collection’s metadata directly from the website in a variety of formats (including CSV and JSON) and subsets (including subjects, locations, metadata facets, and timeline data ready for import to TimelineJS), all for future exploration by users. Additionally, Bethany Nowviskie’s directive to re-envision digital collections as speculative—”not as statements about what was, but as toolsets and resources for what could be”—guides our commitment to engage the site’s user community (composers, musicians, and researchers) as co-creators (Nowviskie 2016) by striving to maintain a meaningful relationship with the composer, and continuing to provide opportunities for composition students to contribute to the project in a meaningful, valued, and compensated way. By beginning our project with well-documented procedures for digitizing, organizing, and preserving the collection’s material, and making use of the sustainability and flexibility that the static web approach affords, we have positioned this collection well for unimagined opportunities that may unfold in the future.

Pedagogical Value

Especially in light of recent discussion about student participation in digital humanities research which all too often renders students’ contributions to projects as invisible or undervalued (Keralis 2018), we wanted to ensure that Marchant’s participation on this project was compensated appropriately. To guarantee that Marchant was paid for the expertise he lent to the project, we conducted this project not as a course project or internship, but by funding his work as a Library student employee. As Di Pressi et al. recommend in their Student Collaborators’ Bill of Rights, Marchant is credited on the project website as a member of the project team, and the website itself can be listed and linked on his CV or resumé (Di Pressi et al. 2015). Additionally, due to the advantages of the static web approach discussed above, the website will not break or be removed anytime soon, so Marchant can count on this project being a visible addition to his portfolio into the future.

From an educational perspective, the skills Marchant used to contribute to the project are broadly relevant to a multitude of career paths. Apart from learning the scanning equipment, file naming best practices, navigating directory structures, and photoshop skills, Marchant also gained experience working in Google Sheets to create metadata and controlled vocabulary for the project. Finally, when it came time to customize the site, Marchant was an active collaborator on GitHub, writing interpretive content in Markdown and using Liquid to incorporate images that enrich the text. Because Markdown consists of just a handful of formatting guidelines, it is easier than HTML to learn quickly and lowers the barrier for students to participate meaningfully on static web projects. Beyond learning the technical skills to contribute to a website, the very act of writing content for the web can be a unique educational experience in itself. As Jeffrey McClurken argues, digital projects such as this one are a prime way to offer students a unique writing experience that prompts them to consider the wider audience for whom they write. In imagining and writing for the future users of a site rather than writing a traditional research paper, students gain real-world experience that they are likely to use again in a future career (McClurken 2011).

Due to his familiarity working with CollectionBuilder on other projects in the Digital Initiatives unit, contributing to the site in this way was not new for Marchant, but rather an opportunity to practice these skills. The repetition involving data management and web development skills in particular (Google Sheets, GitHub, Markdown) were a big advantage of using CollectionBuilder, which focuses creators’ attention on these components rather than on learning the intricacies of a specific database. Marchant may never build a digital collection again, but because so many career paths involve working with data in some capacity and writing content for the web, there is a large chance he will encounter these or similar tools in the future.

For those looking to follow the steps of this project with a student collaborator who may not have any experience with CollectionBuilder, we want to stress that student involvement in the technical development of the site can still be meaningful even if it looks different than what is described here. For students looking to start learning the technical workflows involved in CollectionBuilder, we recommend the Walkthroughs on the CollectionBuilder documentation website. Students who do not have time or capacity to learn CollectionBuilder can still benefit from preparing the collection’s metadata in Google Sheets, where they’ll encounter data formatting, controlled vocabulary, and formulas.

Conclusion

The identification and categorization of composition components, as well as our ability to capture metadata and contextual information directly from the composer himself, were crucial to the success of this project and would not have been possible without the collaboration of a music student. Similarly, the simple and sustainable digital representation of the items in this digital collection relied on our agile approach to digital collections using the CollectionBuilder framework and ensured the production of preservation-ready collection data in the process. Our focus on developing and documenting a file name system and controlled vocabulary for capturing the complexity of a composition’s components ensures that our workflows will be applicable not only to the current stage of the project but to future stages as well. We hope that our method of organization, description, and publication will serve to inspire work on similar music-related digital collections in the future and encourage similar student/faculty digital scholarship collaborations that prioritize sustainable data and websites as well as student learning. Finally, we anticipate that this approach will provide important access to compositions at a scale rarely encountered by composers, musicians, researchers, and critics, impacting a diverse user group and filling a need especially for those seeking to learn from sketches and other rarely published composition components.

This entry is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.