Music DH 2021: A Directory of Digital Scholarship in Music

Music DH 2021: A Directory of Digital Scholarship in Music (hereafter Music DH) (https://rutgersdh.github.io/musicdh/) represents the work of a collaborative project of the Digital Humanities Interest Group of the Music Library Association. Librarians and scholars, many of whom have experience either creating or supporting digital humanities work, collected, collated, and catalogued more than 200 projects that now comprise Music DH. Since its launch in 2021, it has been maintained by Francesca Giannetti, who is the Digital Humanities Librarian at Rutgers University and the Co-Director of the Rutgers Digital Humanities Institute. The Sources page in the site acknowledges, however, that Giannetti is standing on the shoulders of giants in championing Music DH.¹ This project grew in part out of existing bibliographies that catalogue digital projects related to musical topics.² The directory platform itself also leans heavily on designs used in other digital scholarship websites. The clear acknowledgement of collaboration and antecedents is in keeping with the spirit of born-digital scholarship: it takes a village to conceive, launch and support born-digital work.

For anyone seeking to expand research based in traditional publication formats, Music DH should be a first stop on the way to augmenting both primary source and secondary research. It is currently the only open-access directory of digital resources that specializes in musical scholarship where each of the indexed items are catalogued with a standardized and searchable bibliographic record.³ Most of the resources catalogued therein are also identified as open-access as well (only four projects are not open access).⁴ And, it continues to grow! New submissions are still welcome, which means that this project will continue to be relevant for future scholars not only as an archive of born-digital scholarship in music but also as a way to promote new projects.

Content in Music DH

Careful thought has been put into choosing the content in Music DH. The projects catalogued in the list were chosen first and foremost based on their ‘digital-first’ nature, relevance to music as a field of study (broadly defined), and the benefit the project would receive from being part of this collection (the more widely used, the less likely it would be found in this collection). Music DH is also not a project dedicated to public digital scholarship: even though it champions open-access content, its contents are largely by and for scholars.⁵ The particular scope of Music DH means that it is not necessarily a comprehensive collection of digital projects, but it is a carefully curated list designed to provide more visibility to perhaps overlooked or previously unknown digital projects that are relevant to musical scholars.

One particularly wonderful thing about digital scholarship is its ability to produce detailed, carefully researched projects that are not bound by the traditions of scholarly writing and publication. A wide variety of projects are indexed in Music DH; they are produced by scholars around the world and run the gamut of topics and research methodologies. Many are inherently interdisciplinary. This variety is a blessing for researchers and students, but it defies easy categorization for a project team. The Music DH project team seems to have faced this challenge head on, because the content is not organized according to time period, genre of music studied, or geographic region covered, but according to type of project. All 213 projects in the collection (as of early 2023) are classified under a primary project genre: 83 are identified as collections and exhibits, 15 are considered collections of data, 28 are editions of manuscripts or digital editions of music, 20 are laboratories, 39 are reference works, and 28 are software tools.

The breadth of the content indexed in Music DH is admirable and the project itself could easily serve as a research or teaching tool for both experienced digital scholars and novices just learning about digital scholarship. The mix of projects chosen also reflects a growing trend in digital scholarship more broadly: tool and laboratory projects are becoming increasingly important.⁶ Based on the genre categories assigned by the project team, about 75 per cent of the projects currently indexed in Music DH fall into one of three categories of digital scholarship: digital archives (those that encompass exhibition-style projects and editions of music), databases (sites for information retrieval or producing a linked data dataset), and reference works (collections of various resources, sometimes searchable like datasets, sometimes cataloguing editions or displaying digital exhibits). The other 25 per cent is dedicated to software tools that can be used to encode or analyse music or laboratories that allow for some type of data manipulation by end-users. This latter 25 per cent of directory entries set Music DH apart from the earlier projects that catalogue digital scholarship. Many of the software tools and laboratories are relatively new projects and reflect a major technological shift in digital scholarship.

Quite a few of the 200+ projects are dedicated to music performed or composed during the nineteenth-century, however, anyone using Music DH will need to think creatively to discover these projects in the directory, have a bit of patience with searching, and come with a willingness to read the descriptions of projects. The lack of controlled temporal, geographic, or nominal terms was intentional by the Music DH team for a couple of reasons:

First, we wanted this directory to be relatively simple to maintain, and we didn't want to overload our contributors or our users with lots of metadata. Second, while there are separate professional societies for ethnomusicology, historical musicology, music theory, music librarianship, music information retrieval, and so on, we notice that these distinctions are getting fuzzier, perhaps especially so in born-digital works. We'd like to plant a flag in the soil of interdisciplinarity rather than recreate disciplinary divides. To that end, unless disciplinary keywords show up in a blurb, adherence to any given musical field won't be so obvious.⁷

As a metadata professional, I understand completely the challenges of providing enough information to make content discoverable versus creating the perfect record with exactly the right fields at twice the time and effort to maintain. The realities of maintaining Music DH at present means that users will have to spend a bit of time assessing whether their area of interest has representation in Music DH.

For the purposes of this review, in order to highlight a few of the projects that showcase music of the nineteenth century, I am focusing on how to find some the gems in Music DH that engage with nineteenth-century music. My search strategies centred around temporal terms related to the nineteenth century, but other controlled vocabulary would potentially pertain as well. Some metadata records did not have any references to ‘nineteenth’ or ‘19th’ century, meaning that an end-user would need to look for a particular topic or composer to find these projects. The Richard Strauss: Works. Critical Edition – Online Platform is a good example. Richard Strauss was active as a composer as early as 1870 and had composed nearly 200 pieces by the turn of the twentieth century. The Ricard Strauss platform lacks any reference to the nineteenth century in its record (see Figure 1), which means that an end-user would need to search on Strauss's name to identify this platform as part of Music DH's directory.⁸ Similarly, Broadside Ballads Online, a collection of English ballads printed on broadsides in England ‘between the 16th and 20th centuries’⁹ lacks any direct use of the terms nineteenth or 19th in the description or elsewhere in the record (see Figure 2). More than 18,000 broadsides are catalogued in the online platform; a wealth of resources will be potentially missed because ‘19th’ is not included in the record for the Broadside Ballads.¹⁰

Fig. 1. Music DH's record for Richard Strauss: Works. Critical Edition – Online Platform.

Fig. 2. Music DH's record for Broadside Ballads Online

I conducted a couple of searches using ‘nineteenth’ and ‘19th’ as terms. The term ‘nineteenth’ returned two relevant results:¹¹

1. 1) Prices of Music at Breitkopf & Härtel: Publication Lists from the Intelligenz-blatt zur Allgemeine musikalische Zeitung has been created by Derek R. Strykowski as a dataset of Breitkop & Härtel's supplemental advertising published between 1798 to 1838.¹² This index of prices will be helpful for anyone studying the history or the economies of early nineteenth-century music publication.

2. 2) Another project is the Corpus Musicae Ottomanicae (CMO), which comprises digital editions and critical commentaries of Ottoman music.¹³ CMO is divided into two parts: part one consists of transcriptions, commentary, and texts from manuscripts in Hampartsum notation and part two contains the editions of manuscripts in staff notation. Only a fraction of the CMO has been published on the website since it launched in 2020, but the project team behind it – Ralf Martin Jäger, Thomas Bauer and Raoul Motika – seems to be well-supported and it is likely this project will continue for some time.

The term ‘19th’ returned five relevant and varied results:

1. 1) Harp Music of the 18th and 19th Centuries, which is stewarded by the Harold B. Lee Library at Brigham Young University, is a digitized collection of harp music collected by the International Harp Archives.¹⁴

2. 2) Binder's Collections, stewarded by the University of North Texas Library, is an archival and digitized collection of sheet music compilations from the nineteenth century.¹⁵

3. 3) Coleccionistas de Sonidos. Siglo XIX: El álbum musical de Ana y Cristina Echeverría, created by Juana Monsalve, Daniel Castro Pantoja, Juan Fernando Velásquez and Rondy Torres, is an in-depth look at a particular binder of music performed at Ana and Cristina Echeverría's salon in Bogotá during the middle part of the nineteenth century.¹⁶

4. 4) Beethovens Werkstatt is a major project, funded by the Akademie der Wissenschaften und Literarur Mainz, to study how Beethoven developed his compositions. It is built with Music Encoding Initiative software (MEI), which is used to trace compositional development, and on the Edirom tool to map the MEI data to digital scores.¹⁷

5. 5) Hispamus, a large, team-based software project stewarded by the University of Alicante in Spain, is an in-development software tool that will eventually take an image of a score and transcribe its music using pattern and optical music recognition technologies.¹⁸ While the project has its origins in making images of Spanish folk music readable as musical scores, the technology will likely be used more widely. This software is also known as Music Recognition, Encoding, and Transcription (MuRET).

User experience with Music DH

While I wholeheartedly commend the work of Music DH, especially with how the team has thoughtfully curated and tagged the content, I found the user experience of the platform mixed. Some pieces of the website worked very well and others were more difficult to use. The team invested well in providing metadata for each of the directory items. For anyone seeking to develop their own digital directory project, Music DH could serve as a model of how to describe individual items.¹⁹ Each project they catalogued is assigned a persistent identifier (a PID) that is unique to the Music DH directory,²⁰ a link to the project, its title, the status (in development, active, archived, defunct), several semi-controlled description terms (e.g., digital collections, network visualization, linked open data, digital curation, music encoding), a longer-form free-text description, the creators, the stewards (i.e., who maintains it, usually the hosting organization) and the tools used to create it (e.g., Omeka, Solr, SoundCloud, MusicXML). These pieces of information have been entered systematically in each of the records, which has ensured that each of the 200+ items are assigned a fairly robust set of metadata. And each of the fields that are visible in the record are searched in the ‘Search the Collection’ button.

Content organization and cataloguing in any digital project requires creators to think about how end-users will engage with the project once it lives on the internet. It is evident that Music DH has been designed by librarians and scholars with experience assigning metadata. Indeed, the descriptions for each of the projects are key to the success of MusicDH as a research tool. That said, however, end-users should be aware of a few things about the descriptive records in Music DH.

1. 1) The PID used in MusicDH is germane only to this project; provided that the structure of the directory does not change in the semi-near future (years perhaps), the directory links will be stable and shareable. Each item is assigned a unique and persistent directory number (e.g., mdh212) and this PID then becomes part of a persistent URL. Thus this link for The Workshop of Virtual Polyphonic Reconstructions, for example, should always be https://rutgersdh.github.io/musicdh/musicdh/mdh212/, which is important for anyone wanting to cite this entry or share it with fellow scholars.

2. 2) The Open Access status of the projects has been suppressed from the records in the website. That information is available at the CSV of the project (available in this link: https://rutgersdh.github.io/musicdh/reuse/), but it cannot be searched in the directory records. If it is important to you to use open access content in your research, it will be necessary to double check whether a given resource is in fact openly available.²¹

3. 3) The controlled vocabularies have been used only to describe a project's genre (collections and exhibitions, data, edition, laboratories, reference, software) and for its status (in development, archived, active, defunct). While the lack of controlled terms, or extra fields to contain them, is an intentional choice on the part of the Music DH team, ideally a small tag set would formally control the descriptions and to offer temporal, topical, or other subject descriptors for discovery purposes.

4. 4) Only one genre tag is assigned to a project and these categories generally make sense for the platform, but at least a few projects could be identified with more than one genre tag. Bach digital (www.bach-digital.net/content/index.xed) for example, is classified as a reference work; but it has many digital scores in its collection, suggesting that it could have been classified under collections and exhibits as well as under reference. For scholars using Music DH, it will be worth it to explore each browsable category thoroughly to find the composer, topic, or era relevant to their scholarship.

5. 5) The search box searches on the text visible in the records. It does not handle text strings particularly well, meaning that a term like ‘nineteenth harp’ will not return the one directory item that is a project about harp music. A search on ‘nineteenth harp’ will return records that have either nineteenth or harp in them (see Figure 3), but it will not restrict the search to records that have both ‘nineteenth’ and ‘harp’. In addition, a search of ‘nineteenth’ will also not return the record for Harp Music of the 18th and 19th centuries because the word ‘nineteenth’ is not present in the record, but instead the number ‘19th’ is used to convey the period covered in the database (see Figure 4 for an image of the record).

6. 6) End-users may use both the directory records native to the platform and the CSV of all directory items that can be downloaded in their research process. While the metadata records of each item are visually appealing and are screen-reader friendly (kudos to the Music DH team on constructing an accessible website), I have found the CSV to be easier to search and use than the Music DH platform. But the CSV is not inherently citeable or reusable in the manner of the persistent URLs attached each resource (with the PID built into the URL).

Fig. 3 Results for the search on ‘nineteenth harp’

Fig. 4 Music DH's record for Harp Music of the 18th and 19th centuries

The Realities of Digital Project Directories

One significant challenge of maintaining any directory of digital projects is overcoming link rot, the phenomenon where URLs gradually stop linking through to the content originally assigned to a page. Three projects in Music DH are already defined as being ‘defunct’ (Song of the Victorians, Augmented Notes, and Open Music Library). Songs of the Victorians and Augmented notes are dead as of this writing (early 2023). The WayBack Machine has not been able to grab shots of Songs of the Victorians since 6 August 2021²² and Augmented Notes was last captured on 28 July 2021.²³ The Open Music Library (OML), André Avorio's (previously of Alexander Street and now with Coherent Digital) project that worked in collaboration with Alexander Street's (now part of Clarivate) musical score repository, remains only as a page acknowledging the existence of the project.²⁴

Access to other projects in the Music DH collection will almost certainly be compromised as the links currently archived in the directory go dead. This begs the question of Music DH: will its mission expand to provide alternate access points to defunct content? The Music DH team does acknowledge the challenges of keeping abreast of the ever-changing digital landscape in the FAQs when they state ‘“In perpetuity” is not achievable’.²⁵ It takes time and care to maintain directory metadata, just as it takes time and care to maintain a library collection. Good metadata – which includes more than URLs – is an investment that can significantly improve useability and findability of resources and should never be considered one and done. It remains to be seen how much maintenance Music DH will require to keep URLs and other descriptive metadata current for the directory records.

For anyone considering developing a digital project, these already-defunct projects should be a cautionary tale. For scholars already producing born-digital projects, it is necessary to think about your projects in the context of a wider community of scholars. Even the sponsorship of well-endowed companies like ProQuest (now Clarivate) did not protect OML from being closed down. Smaller projects stewarded by one scholar, a small university centre, or a community non-profit organization exist in a more precarious state. Digital work requires its creators and stewards to think much longer term than the initial launch of a project.

The Bottom Line

No other platform or publication exists to date that has canvased this many subject-specific digital projects and provided detailed and accurate records for each of them.²⁶ For scholars and students of nineteenth-century music, Music DH places a number of resources at your fingertips. The searching and discovery experience inside of Music DH is not amazing, but it is functional. At present the directory is small enough that it is manageable to execute a discovery search and explore most of the items in-depth; meanwhile successful known-item searching works well. Moreover, the website and metadata design can serve as models for others hoping to produce and maintain a project without deep technical expertise. The future of Music DH remains to be seen, but as long as it has champions in the Music Library Association, a stable steward like Rutgers University, and an eager project manager like Francesca Giannetti, it will continue to grow and be useful to the scholars and students around the world.