Printer Friendly

MUSIC DISCOVERY: PAST, PRESENT, AND FUTURE.

ABSTRACT

Those designing library discovery systems are presented with numerous challenges, many of which relate to "non-book materials" such as music scores and recordings. Generally, cataloging rules and search or discovery systems are optimized for text written "about" a specific subject or works bearing distinctive, consistent titles. Music scores and recordings do not routinely have distinctive titles, and they exist in a multiplicity of formats. Two additional issues affect users' ability to retrieve music results: a bias toward Western art music and users' expectations based on familiarity with the "one search box approach" found in commercial online retailers' Web sites. This article explores the history of music librarians' efforts to facilitate discovery of music materials. By examining the problems inherent in music discovery and the attempts to solve them through cataloging and discovery systems, the history provided here ends with the release of the Music Discovery Requirements 2 (MDR2): recommendations for those who are guiding the development of discovery interfaces which include music materials, and those who are developing standards for recording and encoding data. Four music-specific issues along with MDR2's recommendations are explored in detail.

INTRODUCTION

But do you know that, although I have kept the [phonograph] diary for months past, it never once struck me how I was going to find any particular part of it in case I wanted to look it up?

--John Seward comes to a realization about recorded media metadata, in Bram Stoker's Dracula (1897), chapter 17

Information that cannot be found may as well not exist. This is true in a pervasive sense (when an item has not been cataloged, no one can find it save by happenstance) and from a singular viewpoint (an information seeker does not find it, so it does not exist for that person). Perhaps most frustrating of all is the situation hinted at in the quote above: the seeker suspects it exists, but (a) lacks the knowledge of how to search for it, or (b) the resource lacks the metadata that would make it findable.

Discovery presents numerous challenges for libraries or indeed any institution that organizes information. Poor or inconsistent metadata can impede efficient research, and data silos can obscure available resources, often necessitating duplicate searches. Others disproportionately affect what part III of the original Anglo-American Cataloging Rules (AACR) called (and which are still sometimes collectively referred to) as "nonbook materials," a motley mix of manuscripts, maps, moving-image materials, music, sound recordings, and two-dimensional images. To this list we could also add rare books, the primary interest in which may not lie in their conventional bibliographic elements. In the days of the card catalog these formats followed specialized descriptive rules and were sometimes accompanied by equally specialized finding aids (such as a publisher file for sound recordings).

For over a century, music librarians working to help users find music materials have contended with library search systems optimized for textual content. Such systems emanated from a card catalog structure and focused on books, thereby favoring author, title, and subject as primary access points. (1) As early as 1902, Library of Congress Music Division head Oscar Sonneck, however, classified music scores primarily by instrumentation and genre, not subject, the norm for books. (2)

Today keyword-based search tools in library online public access catalogs (OPACs) or Web-scale discovery layers (search interfaces presenting results from the library catalog as well as online databases) can expose a wider and often richer array of metadata toward search and discovery than card-based systems ever could. At the same time, however, this environment has eroded the guidance and targeted aid older methods once provided for "nonbooks." Both the second edition of Anglo American Cataloging Rules (AACR2) and its successor, Resource Description and Access (RDA), implemented by the Library of Congress and many US libraries in 2013, endeavored to reduce the book bias of cataloging. In the context of a book-heavy database, however, these nonbooks still can suffer from assumptions made for description of text-based resources: that they (1) contain text that can be indexed for keyword searching, (2) are "about" something in the conventional sense, or (3) have distinctive and consistent titles (indeed if any title at all).

Music librarians are not alone in working on discovery-related issues of nonbook materials. Map librarians have struggled with how the materials they work with fit into current cataloging models and how to squeeze the data they need into MARC (MAchine Readable Cataloging, a standard for computerized encoding of bibliographic data developed by the Library of Congress in 1965 and still widely used today), (3) or how to supplement the catalog with tools better suited to cartographic discovery. (4) Rare-materials catalogers worry about issues of provenance when who made or owned an object is the most important thing about it. (5) Film and image catalogers continue to grapple with questions about how to describe a work that represents another work. (6)

This article describes music librarians' efforts to facilitate the discovery of music materials by examining the problems inherent in searching for music and music-related materials. Whereas today the term "discovery" may be linked to discovery products that incorporate search results of an OPAC together with proprietary database content, in the course of this article the term will be used more genetically to mean any system, from printed catalogs to present day online systems, designed to allow people to find library materials. The article then explores four music-specific issues in detail, including recommendations taken from the Music Discovery Requirements 2 (MDR2), a document containing recommendations for those creating or guiding the development of discovery interfaces that will include music materials, including those developing standards for recording and encoding data.

UNIQUE NEEDS OF MUSIC MATERIALS

Music information-seeking in the library is heavily weighted toward known-item searching, which suggests that looking for music should be simple. But consider the gradations of knowledge that serve to make a piece of music "known": library users might refer to the same item as "the 1985 recording of Georg Solti and the Chamber Orchestra of Europe performing Mozart's Symphony no. 41 in C major, K. 551" or "the Jupiter symphony" or "I heard it on the radio last Thursday and it went kind of like...."

Knowing the composer and title of a work could lead to a long list of search results that might be entirely useless for the purpose at hand. Kiritin Dougan summarized some of the challenges thus:
   The multitude of formats, ... versions that have been orchestrated
   or arranged for different instruments/voices, the high occurrence
   of foreign languages, the frequent presence of both generic (e.g.
   sonata) and nickname titles (e.g. "Eroica" Symphony), and the
   resulting need for uniform titles in catalog records, and the
   frequent need to find a smaller item in a larger context, such as
   one song in an anthology ... make finding music scores and
   recordings a tricky subject to navigate for the average patron.
   This is why even a "known item" music search can be quite involved
   and why many music reference interactions involve a high level of
   instruction. (7)


Moreover, precision vs. recall quickly becomes a critical issue when searching for music. A prolific book author may produce dozens of works, but a prolific composer could produce hundreds, which may be multiplied when one considers formats and editions. Modern library search interfaces, whether traditional OPACs or Web-scale discovery services, have tended to add features such as automatic stemming of search terms designed to simplify searching. Such measures typically serve to cast a broader net and retrieve a greater number of results. Although this can be helpful, as the number of results scales up, it could easily become a hindrance for the user, who must hunt for the right needle in this haystack. A robust faceted browsing system (a search interface allowing users to refine search results by adding/excluding categories or "facets" for author, format, location, etc.), ideally one that takes advantage of music-specific data in catalog records, can help alleviate some but not all problems of this kind.

An additional issue develops when we consider "music" as a homogenous entity. Ethnomusicologists refer to "musics" of differing cultures, and those cultures are not necessarily defined geographically, nor do they necessarily recognize creatorship and responsibility as in Western art music. Current cataloging rules are largely constructed based on Western art music as a template for all music. Kevin Kishimoto and Tracey Snyder have explored the implications of this for popular music, where "the 'composed' songs can be quite skeletal, sometimes consisting of just melody, lyrics, and perhaps a general idea of the chord progression and tempo." They note that "one of the biggest differences from the classical music tradition is the expectation that popular musicians perform their own creative and individualized interpretations of previously existing songs of which they themselves can take ownership." (8)

Applying Western art music's standards of creatorship relegates these performers to a contributor role, (9) which some catalogs may not even display in brief search result lists. For example, in the context of Caribbean steel pan music, the arranger of a piece (also a contributor rather than a creator under RDA cataloging rules) is often considered to fill a more important role than the composer; and in yet other traditions the distinction between an "original" composition and an arrangement is effectively nonexistent. (10) Searching for music outside the Western art tradition presents other challenges, many outside the scope of the traditional library catalog. (11)

HISTORY OF MUSIC DISCOVERY: AUTOMATION AND THE MUSICAL WORK

In recent decades, two developments, automation and consideration of the "work," have had great effect on cataloging theory and therefore the issues associated with the discovery of music in libraries.

Automation

The earliest United States music librarians recognized that book-centric rules were insufficient. They often turned to local customizations and to classification. At the dawn of the twentieth century, Oscar Sonneck and Walter Whittlesey built the Library of Congress music classification schedule around medium of performance, rather than subject as was the norm for books. Music librarians struggled with the manual labor required to provide thorough access, and argued whether form, genre, or medium of performance was the more important access point.

As the automation era dawned in the 1950s to 1970s, librarians' excitement was palpable. James A. Tait wrote, "the computer offers the opportunity to list bibliographical items under numerous bibliographical tags, which would be quite impossibly uneconomic using manual records." (12) Music librarians envisioned finally providing access by form, genre, and medium of performance, with the MARC format ushering in this new era. (13) Soon after the MARC book format's 1968 publication, work began on what became the 1976 MARC music format for scores and recordings. (14)

Donald Seibert, one of the MARC music format's creators, wrote of the MARC format's "greatly enhanced capability for information retrieval" and identified "two factors [which] loom large in this advance: (1) much of the data stored here has never been available in a search capacity before, and (2) the computer can combine these data elements into multi-faceted searches which are better adapted to user needs than the simpler approaches available in the card catalog." (15)

The MARC music format aimed to achieve these goals through coded fixed-length fields. (16) The music-specific fixed-field codes provided the access points music librarians had long yearned for: form, genre, and medium of performance. Seibert highlighted the computer's capacity to provide multiple access points and combine them in a single search:
   In the traditional card catalog, a singer may survey her local
   library's holdings of opera scores by checking under the subject
   OPERAs--SCORES. With the help of the MARC Format, a more
   sophisticated search may be formulated. Availing herself of the
   codes for language, chronological coverage, form of composition,
   and format of score, the singer might search for SCORES of FRENCH
   OPERAS of the CLASSIC ERA. (17)


Even today, however, few library catalogs can execute such searches. How did the MARC music format falter? In two ways: first, creating the fields did not guarantee they would be coded in actual records; second, there was no mechanism for ensuring systems could adequately parse the data.

Music librarians were active advocates from automation's earliest days. After cooperatively developing the MARC music format, they organized the Music OCLC Users Group (MOUG) to ensure ongoing input. (18) In 1984, the same year the Library of Congress put the MARC music format into production, (19) the Music Library Association (MLA) convened a pre-conference workshop on automation, and by 1986, MLA's Subcommittee on Library Automation, led by Lenore Coral, published its "Automation Requirements for Music Information" in this very journal. (20) Over two hundred people attended another MLA preconference workshop on the online catalog in 1989. (21) Presenters noted, as they have ever since, that "music materials highlight the built-in shortcomings of computerized catalogs," and that music librarians must be involved in designing displays and searching functionality. (22) But the fact remained: when computers replaced card catalogs as a discovery mechanism, libraries may have gained possibilities, but catalogers lost direct control over the interface, usually to commercial vendors.

The story of MARC fields 045, 047, and 048 offers a microcosm of the MARC music format's implementation problems. Fields 045, 047, and 048 contain coded data for composition date, musical form, and medium of performance, respectively. In 1989, over a decade into MARC music format implementation, (23) a MOUG member survey demonstrated that libraries were regularly coding the fields, including at least 66 percent of libraries coding one or more fields in some situations, and 42 percent always coding medium of performance. (24) Simultaneously, only a handful of the responding libraries' online catalogs indexed the fields (045, 5 percent; 047, 8 percent; 048, 8 percent). (25) The problem from an administrator's perspective is obvious: coding unused fields seems to be wasted effort. Music librarians identified additional structural problems with the fields: the complexity of making codes searchable in natural language, false hits due to compilations, (26) and instrument codes being too few and biased toward Western art music. (27)

Together, the administrative and structural problems put the fields in jeopardy. Survey authors Jerry McBride and Joy Pile urged MLA and MOUG to:
   work quickly to make these fields and codes useful and easy to
   implement in an automated system.... Viable solutions need to be
   proposed now--solutions which can be implemented easily by current
   local online systems. If we push for the most sophisticated
   indexing application of these codes, there is a very real risk of
   losing this kind of access for years. (28)


The risk became reality. After several years of heated discussion in the music cataloging community, the Library of Congress ultimately ceased coding 045, 047, and 048 on 1 October 1991, citing "the current emphasis on cataloging modification and arrearage reduction." (29) Some libraries continued to code the fields well into the twenty-first century, however, even designing public-facing systems that took advantage of the data. Cooperative solutions waited decades for the Library of Congress Medium of Performance Thesaurus for Music (LCMPT) and the Library of Congress Genre/Form Terms for Library and Archival Materials (LCGFT) music terms, which were released only in 2014 and 2015, with full implementation still in progress today. (30)

In the 1990s through 2004, various music librarians updated the 1986 "Automation Requirements," but never formally released a revision. (31) Nonetheless, music librarians continued investigating musicians' discovery needs and enacting changes to discovery systems. (32) In the wider library world, North Carolina State University's 2006 faceted catalog launch spawned similar products (often called "next-generation catalogs") from numerous vendors. (33) Web-scale discovery (marked by Serials Solutions' 2009 launch of Summon, followed in short order by similar products from other vendors) added the additional wrinkle of a centralized index containing article and other database content, often in such volume as to dwarf local catalog holdings. (34) As always, music librarians promptly began work to configure these tools to meet music users' needs, as evidenced by Tracey Snyder's 2007 assessment of the faceted AquaBrowser catalog for music users and MOUG's interface-specific recommendations for OCLC's faceted WorldCat Local catalog in 2010. (35)

The Music Library Association released the first version of Music Discovery Requirements (MDR1) online in 2012 and published them in this journal in 2013. (36) Like the 1986 "Automation Requirements," MDR1 was not vendor-specific. Unlike the "Automation Requirements," MDR1 narrowed the focus to discovery and searching of musical works, excluding back-end functions, where fewer music-specific needs occur. (37)

In the years immediately following MDR1, extensive changes occurred in data recording and encoding standards. RDA implementation in 2013 ended AACR2's three-decade reign as the primary Anglo-American library cataloging standard. (38) In 2014, the United States cataloging community began implementing two new vocabularies for faceted access to music materials: the Library of Congress Medium of Performance Thesaurus for Music (LCMPT) and music terms in the Library of Congress Genre/Form Terms for Library and Archival Materials (LCGFT). (39) At last, catalogers could use vocabularies designed for music rather than shoehorning (albeit skillfully) form, genre, and medium of performance access into systems designed for subjects. Yet, it still remains for discovery systems to fully take advantage of these new standards. (40) In an isolated, but promising, example, McGrath and Lowery exposed genre, form, medium of performance, and other fields for searching in their consortial discovery interface. (41)

Given these developments, MLA's Emerging Technologies and Services Committee and its Cataloging and Metadata Committee jointly convened the Music Discovery Requirements Update Task Force, resulting in the second edition of the Music Discovery Requirements (MDR2), released January 2018. (42) MDR2 incorporated new vocabularies and standards as well as technological changes in the discovery landscape, but simultaneously recognized change is proceeding at a rapid pace and therefore formalized a plan for ongoing, annual review of the MDR. (43) In addition to the ever-changing functionality of discovery tools, changes on the near horizon include linked data, BIBFRAME, and fuller implementation of LCMPT and LCGFT, as these vocabularies currently coexist with LCSH in most bibliographic records. (44)

The Musical Work and Its Manifestations

One of the perennial issues that arises when discussing the discovery of music materials is the distinction between the "work" and its myriad instantiations, or what Donald Krummel refers to as "the dichotomy between intellectual content and physical form." (45) The musical work is largely considered to exist as an idea or concept, apart from any physical token, that is then manifested as a material product. The musical work can find physical existence as a printed score or a set of parts, as recorded sound, and even as a video recording of a live performance. Furthermore, musical works are known to have a great number of bibliographic relationships with one another, (46) because one work can be published, adapted, and arranged in a variety of ways. While the concept of the work is applicable to all library materials, music librarians are especially aware of the difference between the content of the work, and the physical carrier used to convey it. Understanding the complexities of music discovery involves understanding this distinction, which various cataloging codes have struggled to codify throughout the history of librarianship.

Moreover, cataloging works themselves can be difficult. Although it is easy to compile an alphabetized list of transcribed titles from physical instantiations, it is much more difficult to catalog every musical work held by a library in every form. Devising a system for collocating all instances of a given work in the era of alphabetical filing was done by assigning each musical work a unique heading, known throughout cataloging history as "common title," "conventional title," or "uniform title."

Although they have little to say about music, the very first cataloging codes developed in the nineteenth century implicitly acknowledge the distinction between the work as an idea and its manifestation as a physical object. Anthony Panizzi's 1841 Rules for the Compilation of the Catalogue allowed for the filing of various editions of works together, and his rules for assigning a singular uniform title for different editions of the Bible presage the use of uniform titles to collocate musical works in the twentieth century. (47) Charles Jewett's system of cataloging for the Smithsonian Museum, developed in the 1850s, instructed the librarian to file all translations of a work under the original title, ensuring they would be shelved together as a cohesive intellectual unit, and treating their translations as different editions of a singular work. (48) Charles Cutter's landmark Rules for a Dictionary Catalog (1876), perhaps the most influential cataloging code of the nineteenth century, allowed for the filing of all editions of a work together, regardless of title change. (49) But Cutter's guidelines have been criticized for not explicitly defining the work, instead referring to "what the library has by a given author," (50) and for implying a one-to-one correspondence between a work and a physical "document" that does not represent the complex bibliographic world of music. (51)

While these early cataloging codes may have recognized the idea that a single work could be manifested in multiple editions, none of them discuss music directly, and it was up to music librarians in the late nineteenth and early twentieth centuries to simply be pragmatic when it came to cataloging music materials. (52) Oscar Sonneck's three-page appendix in the 1904 edition of Cutter's Rules (53) briefly acknowledged the issue with devising titles for musical works, but offered no additional rules to resolve the problem. A 1920 ALA Bulletin includes a "Report of the Committee on Catalog Rules" in which a few more rules are offered for music, including the instruction that, for dramatic vocal works, one should "choose a common title for those scores where two or more entries are necessary for different words accompanying the same music either as title or text," (54) understanding that an opera in translation is still the same work as the opera in the original language, and allowing, for example, Die Walkure and The Valkyrie to file together. For the most part, however, cataloging rules in the pre-World War II era simply ignored the issues with representing musical works. The 1908 American Library Association cataloging code, for instance, allowed for uniform titles to represent the Bible and anonymous classical works, but contained nothing on music. (55)

Although it was clear that a comprehensive system of uniform titles was needed to aid in the discovery of musical works in a library catalog, the growth of sound recording collections forced the issue. Many sound recordings were issued as compilations of several works, necessitating new rules to guide a patron to the compilation that held the work sought. (56) By 1927, an ALA report on the practices of libraries with significant music collections indicated that many music librarians were using "common titles" in order to collocate musical works. (57) The growing discussion about how to handle music led to the formation of the Music Library Association in 1931, and the music cataloging community began to develop a set of rules largely divorced from the general cataloging community in the American Library Association. (58) It would not be until the late 1970s when the two communities would begin to operate in tandem again.

By 1941, the ALA cataloging rules finally recognized the uniform title officially, defining it as "The distinctive title by which a work which has appeared under varying titles and in various versions is most generally known," (59) while failing to define what was meant by "work." In the MLA rules published the same year, the guidelines went further, with paragraphs describing the necessity of "conventional titles" in music cataloging. Here, MLA states:
   Conventional titles are filing titles established according to rule
   and included in the catalog entry in order ... to identify and
   bring together in the catalog all editions and arrangements of a
   composition.... The precedent for this treatment comes from similar
   usages in the cataloging of the works of voluminous authors such as
   Goethe and Shakespeare and in the cataloging of editions of the
   Bible. The necessity for this device is particularly great in the
   field of music because of the widespread use by composers of all
   periods of titles consisting of names of musical forms and because
   of the fact that musical compositions are frequently issued in
   numerous editions with variations in the language and the wording
   of the title pages. (60)


For the first time, here was a clear call for the necessity of distinguishing a musical work from its printed manifestations, citing both precedent from Panizzi and Cutter, and pointing out the unique problem of musical titles.

The MLA rules represent a major step forward in terms of recognizing the problem of cataloging musical works. Yet both ALA's and MLA's 1941 rules fail to distinguish the "work" as a separate entity with its own set of attributes. Instead, the rules favor utilitarian filing techniques over any sort of real distinction, with uniform titles used as a "filing expeditor" for published manifestations, rather than representing the work itself. (61)

One of the most famous catalogers of the twentieth century, Seymour Lubetzky, can be credited for bringing the concept of the work as a separate entity into prominence in cataloging codes. Throughout the late 1950s and 1960s, Lubetzky spearheaded a movement to reform cataloging, simplifying some of the arcane filing principles that had been propagated and focusing on certain first principles upon which the discipline of cataloging could be built.

One of Lubetzky's central principles was that of the "work," and he maintained that the intellectual material of the "work" should not be confused with the physical manifestation of the "book." (62) "The book, it should be noted," he wrote, "comes into being as a dichotomic product--as a material object or medium used to convey the intellectual work of an author. Because the material book embodies and represents the intellectual work, the two have come to be confused.... It must be recognized that ... a book is not an independent entity but represents a particular edition of a particular work by a particular author" (63) Lubetzky's own draft of the Code of Cataloguing Rules proposed that one of the objectives of the catalog was to "relate and display together the editions which a library has of a given work and the works which it has of a given author," (64) while the functions of the catalog outlined in the 1961 "Paris Principles," the product of the International Conference on Cataloging Principles based on Lubetzky's work, state that one of the functions of the catalog is to ascertain "which works by a particular author and which editions of a particular work are in the library." (65)

Lubetzky himself wrote very little about music specifically, but his idea of the "work" was broad enough to apply to music cataloging, and he did briefly address the organization of music in several respects. He affirmed that a "work" could exist across multiple forms of media, including a printed score and a sound recording, writing that "any given communication or work may be represented in different formats and by different media," (66) a definition that stands in contrast with others who maintained that a change in medium meant a distinction in work. (67) Lubetzky also addressed the issue of compilations of several different musical compositions on a single sound recording, and proposed the idea, later codified, that performers could be considered the author of the "work" of a sound recording compilation. (68) While this idea of the performer as author was used to facilitate discovery and prevent onerous cataloging rules, rather than propose the idea of a "pure theory of performance as authorship," (69) the idea of "performer as author" would be included in several later cataloging codes.

Lubetzky's refraining of cataloging to focus on the "work" was not without pushback, as many librarians worried that the idea of cataloging works as well as physical books was too time-consuming and cost-prohibitive. (70) The idea soon became widely accepted, however, and the work principle became incorporated into the Anglo-American Cataloging Rules (AACR) published in 1967, and its second edition (AACR2), published in 1978. These rules implicitly acknowledge the work/manifestation distinction, but still stop short of a full definition of a work; the closest they come is a brief footnote in the AACR2 chapter on uniform titles that indicates that the definition of a "work" can in elude collections or compilations. (71) By treating a compilation as a "work," AACR2 also allowed the principal performer of a sound recording to be considered the "author" of that work. (72) It was not until 1983 that the ALA specifically defined the work as "a specific body of recorded information in the form of words, numerals, sounds, images or any other symbols, as distinct from the substance on which it is recorded." (73)

The rise of the Anglo-American Cataloging Rules also coincided with the shift of library catalogs from the classic card catalog to an OPAC. Creating a digital library catalog allowed for new relationships and connections that were not possible in the card catalog era, and new research began to reveal just how many bibliographical relationships there were between musical materials. (74) Music librarians had long hoped that digital catalogs would allow for "appropriate techniques for providing substantial access to musical works," (75) and the use of uniform titles was seen as a key toward that end. In 1990, Sherry Vellucci called for an end to linear catalogs, and instead for a more multidimensional structure in which uniform titles could exist as "effective linking devices to make known many of the bibliographic relationships that exist among various types and formats of material today." (76)

But while uniform titles existed to collocate different manifestations of the same work, they did not represent the works themselves, and by the 1990s, many librarians were beginning to look for ways to represent the musical work beyond a mere textual heading. Martha Yee lamented that OPACs had no specific format for cataloging works, and most did not even handle the authority records for uniform titles very well. (77) Gregory Leazer, meanwhile, worried that the sheer number of volumes cataloged in a worldwide digital catalog would make music discovery impossible if there were not some conceptual structure applied to organize works and "reduce the number of individual records the searcher is required to peruse." (78) David Thomas and Richard Smiraglia called for a schema "centered on the control of bibliographic works, [which] would in part replace current bibliographic structures that are centered on the control of bibliographic items." (79)

As OPACs continued to grow, music librarians were not the only ones to worry that the new digital world was unsustainable. The problem of scale and the lack of digital structure for the work/manifestation distinction led to the organization of a seminar on "universal bibliographic control" in Stockholm in 1990, which in turn led to the IFLA Standing Committee on Cataloging forming a study on the "functional requirements of bibliographic records." (80) The product of this study would be IFLA's 1997 Functional Requirements for Bibliographic Records (FRBR), (81) a new conceptual schema for cataloging that focused on an entity-relationship structure.

FRBR proposed a four-tier structure for cataloging, distinguishing between the work, the expression, the manifestation, and the item. Each one of these four tiers is defined as its own entity with its own set of attributes. Unlike previous theoretical structures of cataloging, FRBR defines a work entity as something completely separate from the physical manifestations, and one could see FRBR as finally fulfilling, a century later, the idea of separate definitions for intellectual and physical content.

The existence of the expression entity is a new concept, an entity that is neither a work nor a physical manifestation. Works can be different expressions based on what medium they are expressed in (e.g., printed score or recorded sound) or what language (e.g., a Wagner opera in the original German vs. an English translation). In this way, one could view the expression entity as a method for dealing with the long-running debates about whether translations or publications in different mediums were new works or merely different instantiations of the same work. According to FRBR, such entities are the same work, but different expressions. Likewise, a performer on a sound recording is now considered to be contributing to the expression of a work, rather than creating a work.

In response to FRBR, a new revision of the Anglo-American Cataloging Code, originally intended as "AACR3," morphed into a completely new cataloging code known as "Resource Description and Access," or RDA. (82) RDA's structure centers on the four FRBR entities, with separate rules for cataloging works, expressions, manifestations, and items.

After a lengthy period of drafts and revisions, RDA was officially adopted for use at the Library of Congress and many other American academic libraries on 31 March 2013. Librarians and knowledge organization professionals continue to tweak the details of FRBR and RDA. An object-oriented version of FRBR, known as FRBRoo, was released by IFLA in 2010, (83) largely for the museum and cultural artifact communities, and a large-scale update to FRBR, integrating work on authority records and subject cataloging, was released in August 2017 as the IFLA Library Reference Model (IFLA-LRM). (84) There is still work to be done-while the rules now call for a separation of the work and its instantiations, few discovery products allow for easy or intuitive searching for these separate entities, and music librarians continue to stress their importance. Only a small number of discovery systems have attempted to fully capitalize on the FRBR structure. (85) Richard Smiraglia's words from 2003 still ring true today: "The challenge for the postmodern catalog is to incorporate the richness of hyper-linkages to fully exploit the informative capability of works." (86)

CHANGES IN USER EXPECTATIONS AND THE FUTURE

As should be clear by now, music library patrons have unique needs due to the complexity of music searching. Moreover, user expectations for discovery systems have increased greatly, largely due to the rise of single-box Internet search engines such as Google, (87) which feature proprietary relevance ranking, and interfaces like Amazon's, that offer faceted browsing. (88)

In this new world of one-size-fits-all keyword searching, browsing an alphabetized list of controlled terms, or employing Boolean search phrases, may be far afield from what the average library patron expects when attempting to search for music. While such search strategies are better suited for discovering musical materials than using a single keyword search box, they may be unintuitive to the general user today.

Library interfaces have changed to accommodate these evolving patron expectations. Many public catalog interfaces now direct patrons to a single-box keyword search, with more detailed searches confined to an "advanced search" page. This is done to make searching easier, more intuitive, and less confusing. Yet, in the pursuit of simplicity, advanced search techniques can end up buried, or absent altogether--a problem for music searching.

Despite new features being added to every iteration of library discovery software, the fact remains that nearly all bibliographic data is encoded in MARC, and follows cataloging guidelines meant to facilitate left-anchored alphabetical browsing. This means that there are limitations to what can be done. Content standards including the Library of Congress Subject Headings (LCSH) and the Library of Congress/NACO authority file are meant to facilitate browsing an alphabetized list, but may be less than ideal for a keyword or faceted data search.

Moreover, the absence of unique titles and the need to recognize musical work titles in different languages require a system for searching music that is not as simple as recalling a single phrase. Recent developments in full-text keyword searching are likewise unhelpful when searching for nontextual materials such as music; library systems designed to search for specific melodies or incipits have, at best, been applied to a small set of materials for a niche audience. (89)

Current metadata developments aim to make bibliographic data more adaptable for these newer forms of searching. Newer controlled vocabularies, such as the LCGFT, were conceived to function in a faceted search, rather than as part of an alphabetized, browsable list. Name authority records encoded in RDA, based on the requirements laid out in the Fundamental Requirements for Authority Data (FRAD), now often include information on a creator's nationality, occupation, and gender, raising the possibility of a faceted search based on creators' attributes, for example. RDA and its entity-relationship model allow for a stronger distinction between an abstract work and its published iterations, which, as mentioned earlier, is very important for music. These developments could potentially make the discovery of music easier for the novice as well as the advanced searcher. The Music Discovery Requirements provide needed guidance, and the following section explores several major issues.

THE MUSIC DISCOVERY REQUIREMENTS

The Music Discovery Requirements serve two broad purposes: (1) to provide technical guidance for those responsible for developing, implementing, and modifying discovery interfaces (especially those without a background in music or in metadata creation), and (2) to document and explain the significance of data elements that describe musical materials (particularly when the use or significance of those elements differs from textual or other resource types). The 2018 Music Discovery Requirements 2 (MDR2) updates the 2012 MDR1 to reflect current standards. Four issues exemplify music discovery challenges: known item searching, titles, content and carrier, and medium of performance.

EXPLORING THE ISSUES: KNOWN ITEM SEARCHING

Known item searching is a frequent topic in discussions of catalog use in general and access to music materials in particular, but as Lee, Renear, and Smith have illustrated, the boundaries of the known item search are fuzzy and working definitions tend to vary from researcher to researcher. Definitions may be conceptual, based on philosophical distinctions from other types of searches, or operational, based on the actions undertaken by the searcher. (90)

One philosophical stumbling block is the word "item," which might conceivably be interpreted in its most restrictive sense--the sense in which FRBR uses it, as a singular and specific physical object--but is more likely to be a manifestation, expression, or work, depending on the level of specificity the searcher requires. Another complication, already noted, is the many ways a resource might be "known," not all of which are bibliographically useful (the "blue cover" problem). Many known item searches might more accurately be described as partially-known item searches: the searcher seeks any resource that fulfils a limited set of criteria. The searcher may have specific requirements of the expression (an audiobook, a Spanish translation) or manifestation (a large print edition, the original 1962 publication), but for most purposes one edition is effectively interchangeable with any other: the fundamental intellectual content alone--the work embodied--suffices to meet the searcher's needs.

For music the work embodied alone is seldom sufficient to meet the searcher's needs because "music" encompasses two generally non-interchangeable categories of expression: notated music and recorded performances. At a minimum, most known item searches for music must commit to one of those categories. Even narrowing a search down to notated music leaves the searcher with critical decisions to make based on presentation format, notation style, medium of performance, and other criteria.

Presentation format refers to the "form factor" of notated music: full score, vocal score or piano reduction, study score, score and parts, etc. Each of these formats serves different purposes that only sometimes overlap. For example, a full score contains all the parts of a musical composition arranged to be visible at once, which in the case of a work for large orchestra or multiple ensembles can completely fill the vertical space of a significantly oversized page. This provides the information necessary for conducting or for many kinds of analytical study but is often completely impractical to play from. A singer who presents a pianist with a full orchestral score rather than a vocal score to play from may soon require a new accompanist.

Medium of performance is another common selection criterion, whether the searcher considers it at the outset or not. The act of transcribing music written for one instrument to another often involves making alterations to account for the different limitations and capabilities of the new instrument. Even when this is not the case, a transcription for a transposing instrument is rarely trivially reversible. A transposing instrument (generally one of a family of different sizes and ranges, such as the B-flat clarinet, E-flat clarinet, bass clarinet, etc.) is notated higher or lower than the sounds it actually produces, so that the same fingerings may be used on all members of the family; a violinist playing from a clarinet transcription would therefore be playing in the wrong key.

The form of notation is easily overlooked as modern Western staff notation is generally the default assumption. Much popular music, however, is also published as guitar tablature; jazz might appear as a lead sheet consisting only of melody and chord symbols. Each of these notation forms is uncommunicative to those unfamiliar with them, and even a pianist who knows guitar tablature is unlikely to be satisfied with it: staff notation tells performers which pitches to produce; tablature tells them where to put their fingers on the guitar. Braille is of obvious utility to the sight impaired and of limited interest to others, but there are also more esoteric tactile applications such as Nelson Howe's Fur Music, in which the score doubles as the medium of sound production.

The frequency and non-interchangeability of multiple instantiations for a musical work present numerous challenges. Even in the event that a searcher would be satisfied with any expression of a work, identifying works in music is a minefield for the unwary. Known item problems permeate music discovery, so they are addressed throughout the entire MDR2, which provides guidance in how to present, index, and provide access to metadata to minimize the frustrations inherent to searching for music materials.

TITLES

Musical titles present several potential difficulties for description and access, one of which is linguistic. If a manifestation of a textual work bears a title in a foreign language, that is generally an indication that the content is in the same language, and thus only usable to speakers of that language. But musical notation, like mathematical symbols, transcends language barriers. From the music user's perspective, it is irrelevant whether the title of a manifestation is "String Quartet in B minor," "Streichquartett H-moll," "Quatuor a cordes en si mineur," [phrase omitted] or even all of the above arranged with varying degrees of typographical prominence; the notes are the same regardless. It is, however, relevant to how or if the user finds such a manifestation. In the example above the titles consist only of a type of composition (generally a genre, a medium of performance, a tempo indication, or a number of performers) and a key, but even more evocative titles such as Quartet for the End of Time or Die Zauberflote are just as likely to appear as Quatuor pour la fin du temps or The Magic Flute.

In music cataloging, titles consisting only of a type of composition are known as generic titles, as opposed to distinctive titles. Since many generic titles derive from shared musical terminology they often have many cognates, at least in European languages, which are obvious to human eyes even if some nuance is lost (e.g., "sonate" is clearly related to "sonata," though it is singular in French but plural in Italian), but opaque to a naive string-matching algorithm. An additional consequence of generic titles is that the better-known works that bear them often acquire well-known nicknames: Beethoven's Piano Sonata no. 14 in C-sharp minor is more commonly called the "Moonlight Sonata." Haydn's works are a goldmine of inventive nicknames, a small sampling of which includes Razor, How do you do?, Trauer, La passione, Tempora mutantur, Kaiser, Oxford, Der Philosoph, and Farewell--which may or may not be known in other languages and may or may not be used in published manifestations.

Often the difference between a generic and a distinctive title is obvious. Puccini's Madama Butterfly, Rameau's Le rappel des oiseaux, and Peter Schickele's (writing as P.D.Q. Bach) Oedipus Tex are all distinctive in the same sense as titles like Hamlet, One Hundred Years of Solitude, or The Butter Battle Book in that the title alone is in many cases sufficient to identify the work unambiguously. The same cannot be said of "symphony," which without additional qualifying information is no more bibliographically helpful than "cookbook." Other distinctive titles are more ambiguous; per RDA definitions a title like Petite sonate, no matter how generic sounding, still qualifies as distinctive thanks to the additional adjective. The Music Library Association's Cataloging and Metadata Committee maintains a document providing rulings as to whether common titles are generic types of composition, either always, conditionally, or never, and their cognates in other languages; printed on standard 814 x 11 paper it is fifty-seven pages long. (91)

It can be difficult for nonmusicians to fully grasp the level of complexity that can arise with musical titles because it is not a problem with which they have practical experience. A brief musicological excursion may help to illustrate and give context to these issues. Consider the keyboard works of Domenico Scarlatti (1685-1757), which nearly every serious piano student will encounter at one point or another.

Scarlatti was both a brilliant performer and music teacher to Maria Barbara de Braganca, infanta of Portugal and queen of Spain, and later to her children. In that capacity he composed approximately 555 one-movement keyboard sonatas, most of them quite short (many under six minutes long, including repeats). Of that prodigious number Scarlatti himself published only thirty, in a luxuriously oversized 1738 London edition entitled Essercizi per Gravicembalo, or "exercises for harpsichord." Thanks to copies of other sonatas circulating in manuscript, by the end of the eighteenth century audiences in England (which Scarlatti never visited) had access to printed editions of around eighty sonatas, with perhaps a further fifteen available in manuscript if one knew where to look. (92)

Although there was no equivalent to the London edition elsewhere, there were two significant manuscript collections, now in Venice and Parma, thought to have been copied under Scarlatti's own supervision. Over time more and different sonatas came to light, culminating in the publication of several collected editions that have done much to shape the way musicians perceive this body of work.

In 1838-40 Carl Czerny published an edition of 200 sonatas (including a few misattributions) in twenty-five volumes. That number blossomed to 545 in the complete edition by Alessandro Longo in ten volumes plus a supplement (Ricordi, 1906-8). Ralph Kirkpatrick listed 556 (accounting for a 204a and a 204b) in his 1953 biography, which formed the basis for a modern edition by Kenneth Gilbert (Heugel, 1971-84) and Kirkpatrick's own 1972 facsimile edition from early printed and manuscript sources. In 1978 Ricordi began issuing a new complete edition to replace the Longo, edited by Emilia Fadini. There are of course many other publications of substantial subsets of this corpus.

There are no known copies of any Scarlatti sonata in his own hand, a fact perhaps attributable to the 1755 Lisbon earthquake and the subsequent fires in which so much documentary evidence perished. The surviving manuscripts provide no insight into compositional chronology. It would therefore be deliriously optimistic to expect any of the above editions (Gilbert and Kirkpatrick excepted) to organize themselves in anything like the same way. Longo grouped the sonatas into "suites" of five according to his own aesthetic sensibilities. Kirkpatrick paired the sonatas by key in what he claimed was approximate chronological order, although this entailed a certain amount of sometimes fanciful extrapolation from scant evidence. Fadini follows Kirkpatrick's ordering within groups of sonatas from a common source; the order of sources is not the same as Kirkpatrick's. To this jumble of conflicting opinion we must also add the musicologist Giorgio Pestelli who in 1967, taking issue with Kirkpatrick's numbering, published his own list of 559 works chronologically ordered on the basis of stylistic analysis. (93)

The net result is that Scarlatti's keyboard sonatas enjoy no fewer than four purportedly complete and mutually incompatible numbering systems: Longo, Kirkpatrick, Pestelli, and Fadini. Even if Kirkpatrick's system has largely won out in modern publications while Fadini's is seldom spotted in the wild (perhaps due to her inclusion of K. and L. numbers alongside her own), Longo's numbers still appear with some regularity, and of course in older publications and recordings all bets are off. Only Czerny's relatively modest listing has slipped mercifully into obsolescence.

These quirks of transmission history have serious implications for how librarians catalog these works in a way that anyone could possibly hope to find them. To summarize the challenges for description and access:

* There are more than 550 works by the same composer with the same generic title.

* The title(s) by which these works are most commonly known ("sonatas," "sonate," "sonates," etc.) is not the title used for the only subset of them published under the composer's supervision ("essercizi").

* Out of twenty-four possible major and minor keys (ignoring enharmonic equivalents), the sonatas use twenty-one, of which any given key appears between two and seventy times; key is therefore not sufficient to uniquely identify even a single work.

* The numbering of the sonatas is unrelated to the composer's will, the chronology of composition (in any reliable and verifiable way), the publication history of the works, or any other inherent quality of the works.

* Any given manifestation of Scarlatti, whether printed or recorded, may use any numbering system or combination of numbering systems or even (cruelly) none at all.

Whereas Scarlatti presents an extreme example, we can see parallels elsewhere: Haydn's eighty-four string quartets (not all of which were originally labeled as such), or C. P. E. Bach's dueling catalogs of different vintages (Wotquenne or Helm?). Moreover, it is not necessary for a composer's works to be numerous to cause confusion. Anton Bruckner, an inveterate reviser, wrote only eleven symphonies, sensibly numbered 1-9. The first one didn't count; the original no. 2 was found wanting and withdrawn and is now known by a variety of monikers in several languages, including "Die Nullte" and "Symphony no. 0" (despite actually being the third chronologically). Others were not outright disclaimed but were significantly revised or had whole movements excised and replaced; his fourth symphony exists in three different versions.

It is no wonder then that music librarians have been enthusiastically adopting standardized forms of titles and drawing clear lines between the works they denote and their haphazardly identified instantiations for decades. Although changing technology has meant many library practices of ninety years ago are no longer necessary or helpful, the ever growing number of musical publications, many in formats undreamed of in the 1920s, has only increased both the potential for musical titles (and generic titles in particular) to bewilder and overwhelm patron and librarian alike and the need for systems that responsibly handle the authority records painstakingly created to disambiguate and collocate them.

CONTENT AND CARRIER

Musical titles' vagaries are not the only challenge facing musical experts and novices. The music discipline depends on using information in different formats, including scores, audio recordings, videos, and texts. IFLA-LRM terminology calls these formats "content types." These content types, particularly recordings, exist in many different carriers: physical carriers such as compact disc, LP, audiocassette, videocassette and videodisc; and digital files encoded using formats such as MP3, MPEG, and WAV. IFLA-LRM terminology calls these physical and digital carriers "carrier types."

Multiple carrier types regularly nest within each content type, especially for audio and video recordings. On an intellectual level, carrier types are more interchangeable than content types: an LP, CD, and audio file all provide a recording of the piece. On a practical level, carrier types are rarely so interchangeable. From the 1877 invention of recorded sound to the present, numerous recording formats have been employed, and users generally have playback capabilities for only a select few. Contrast this to the situation for printed matter, where one can read a well-preserved 1877 book with no special accommodation. Not so an Edison cylinder.

In the card catalog era, librarians debated whether to file score and recording cards together in the same drawers or organize them in separate cabinets. Filing together provided convenience when musicians needed multiple formats at once--a score and a recording as they learned or studied a piece. On the other hand, users easily missed the content type distinction while searching and wasted time obtaining the wrong item. One solution was different colored cards for different content types, which worked for the educated user, but was meaningless to the beginning searcher. Cazeaux suggested wholly separate card files were best, sending the user to the correct place from the beginning. (94)

Today, searching happens online, and properly-configured systems can provide both approaches proposed in the card catalog era. Systems must allow users to select a particular content or carrier type at the beginning of their search or via post-search filters (the equivalent of going to a different card catalog cabinet) and easily distinguish between content and carrier types in a results list or detailed record (the equivalent of different colored cards.) But no longer are colored cards needed to distinguish content and carrier types; icons are now available for this task and can carry intellectual meaning: even a novice can distinguish between well-designed icons for "score" and "audio recording."

Content type and carrier type must be separately actionable. Allow a single record to be assigned multiple content and carrier types, or group content and carrier types together for public interfaces. For example, an item must be able to be both an "audio recording" and a "CD." The proliferation of e-books has helped bring this need to the forefront. Both are the content type "book" and users need to be able to select the "book" content type, and to further filter by carrier type to include only e-books or only print books. The situation for audio and video is similar and can generally be met with like mechanisms, except that libraries will usually have more--perhaps many more--than two carrier types for audio and video.

MEDIUM OF PERFORMANCE

Medium of performance is defined by Richard Smiraglia in his 1986 guide to music cataloging as "the instruments, voices, etc., used in the realization of a musical work." (95) The ability to search for musical compositions by medium of performance has been a problem in music cataloging because some forms of music, such as symphonies or operas, imply a specific medium of performance. Other forms of music, such as sonatas or concertos, can be composed for a variety of different instruments or performers.

Traditionally, discovery of music by medium of performance has been tackled through classification. The Library of Congress classification system's section on music (the "M class") is largely organized by medium of performance, separated into instrumental and vocal works, and subdivided further from there. While robust and continually growing, the classification system cannot tackle all possible mediums of performance, and due to its genesis as a musicological tool developed in the early twentieth century, the Library of Congress system is poorly suited for classifying music other than common practice era Western art music. (96)

Additionally, medium of performance is more complex than simply a one-to-one correlation between instrument and work. A budding violin student, for example, could browse the LC classification section for solo violin music (M40-M44), but would also want to look at music for violin and piano (M217-M223), or music for violin and orchestra (M1012), necessitating a visit to multiple locations in the library stacks.

Subject headings describe the medium of performance of musical works as well. A piano sonata, after all, is meant to be performed on a piano, however, it does not make sense to describe "piano music" as the subject of a piano sonata. Subject headings for music have described what the piece is rather than what the piece is about. (97) Even allowing for subject headings that describe "is-ness" rather than "about-ness," traditional Library of Congress Subject Headings (LCSH) still fail to collocate all pieces containing a specific medium. Again, a violin student would have to look in the catalog under different subject headings such as:

Violin music

Violin and piano music

Violin music (Violins (2))

LCSH headings have included form as well as medium, however, which means that there is no good way to alphabetically file all violin music together when taking into account headings such as:

Concertos (Violin with chamber orchestra)

Sonatas (Violin)

String trios (Violins (2), cello)

The medium of performance becomes even more confusing when one considers the multitude of arrangements published that are then collected by libraries. Works written for one medium of performance can be arranged for another, such as Maurice Ravel's arrangement of Mussorgsky's Pictures at an Exhibition for full orchestra. Dramatic vocal works also are published as "vocal scores," in which the orchestral accompaniment is reduced to a solo piano line. In these cases, subject headings can indicate that a piece is an arrangement or a vocal score, but the original medium of performance would be relegated to a note in the catalog record if stated at all.

MARC records contained some other methods for recording medium of performance, such as the 048 field, but these were never widely implemented, nor were discovery systems developed that made these fields searchable. It was not until the publication of the Library of Congress Medium of Performance Thesaurus (LCMPT) that libraries began to rethink the ability to search for medium of performance on a large scale.

The LCMPT is developed as a faceted vocabulary, making it easier to search for any one term in a medium of performance statement. While the subject heading

String trios (Violins (2), cello)

makes it difficult to browse a list of headings to find pieces for violin or cello, the LCMPT would allow one to encode a record as such

$a violin $n 2 $a cello $n 1 $s 3 $2 1cmpt

indicating the work was for two violins and one cello for a total of three performers. Not only does LCMPT allow for one to encode variant instruments (e.g., a piece for piano or harpsichord) and doubled instruments (e.g., a flutist who plays the piccolo during one section of a piece), the thesaurus vocabulary allows for a robust hierarchy. A "soprano saxophone" is defined, in increasing generality, as a "saxophone," a "single-reed instrument," a "reed instrument," a "woodwind instrument," a "wind instrument," an "aerophone," and finally simply an "instrument." The hope is that this hierarchy will eventually allow patrons to search for broader or narrower terms as appropriate.

MDR2 contains specific recommendations to take advantage of the new LCMPT vocabulary and ensure the most robust discovery of medium of performance statements in older records. This includes indexing and displaying the MARC 382 field, a field that may require additional display recommendations, because it was not developed to display directly to patrons. (98)

The four issues of known item searching, musical titles, content and carrier types, and medium of performance exemplify music discovery challenges. Medium of performance stems from a unique but vital attribute of music. The other three are prevalent in, but not unique to, music. Necessity has forced music catalogers to address these issues with more depth and detail, and fixing these problems for music will trickle down to optimize discovery for other formats where these problems occur less frequently.

CONCLUSION

It is understandable that library discovery systems remain very much in flux at this time. For over a century librarians based searching on a left-anchored, alphabetically-sorted browsing system derived from card catalogs, whereas today we have the opportunity to design discovery systems employing digital strategies unbound by previous limitations. As these systems continue to be refined, it is imperative that librarians, programmers, and vendors developing the next generation of search products consider the unique needs of music users. To that end, the Music Discovery Requirements provide a map for those creating discovery environments and enable them to keep music users' needs at the forefront.

Chris Holden is a cataloger at the Music Division of the Library of Congress. Keith Knop is head, Music Cataloging at the University of Georgia. Nara L. Newcomer is head, Music/Media Library at the University of Missouri-Kansas City. All were members of MLA's Music Discovery Requirements Update Task Force. URLs cited herein accessed 8 March 2019.

(1.) Cutter defined the first object of a dictionary catalog as "to enable a person to find a book of which either (A) the author; (B) the title; or (C) the subject is known." Charles A. Cutter, Rules For a Printed Dictionary Catalogue (Washington, DC: Government Printing Office, 1876), 10.

(2.) Sonneck created the music ("M") schedule for the Library of Congress classification system, assigning subclass M to scores. The primacy of subject access elsewhere in Library of Congress Classification is apparent not only by examining the schedules themselves, but also through works on the history of the system, such as Francis Miksa's The Development of Classification at the Library of Congress ([Champaign-Urbana, IL]: University of Illinois, Graduate School of Library and Information Science, 1984). Miksa's discussion of "subject collocation" clearly assumes that "subject collocation" is a primary feature of classification.

(3.) Stacie Traill, "Exploring the Terra Incognita of Access and Discovery: The Evolution of Cartographic Cataloging in the Twenty-First Century "Journal of Map & Geography Libraries 10, no. 1 (2014): 48-61, Digital Object Identifier (DOI): https://doi.org/10.1080/15420353.2014.884035.

(4.) Marcy M. Bidney, "Can Geographic Coordinates in the Catalog Record Be Useful?" Journal of Map & Geography Libraries 6, no. 2 (2010): 140-50, https://doi.org/10.1080/15420353.2010.492304.

(5.) Allison Jai O'Dell, "Maker Metadata: Problems and Possibilities," Cataloging & Classification Quarterly 53, no. 7 (2015): 785-800, https://doi.org/10.1080/01639374.2015.1020584.

(6.) Martha Yee, Moving Image Cataloging: How lo Create and How to Use a Moving Image Catalog (Westport, CT: Libraries Unlimited, 2007), 59-60.

(7.) Kirstin Dongan, "Information Seeking Behaviors of Music Students," Reference Services Review 40, no. 4 (2012): 558-59, https://doi.org/10.1108/00907321211277S69.

(8.) Kevin Kishimoto and Tracey Snyder, "Popular Music in FRBR and RDA: Toward User-Friendly and Cataloger-Friendly Identification of Works," Cataloging & Classification Quarterly 54, no. 1 (2016): 64, https://doi.org/10.1080/01639374.2015.1105898.

(9.) At the time of writing it remains to be seen how the incorporation of IFLA's Library Reference Model, which includes a standard of "intellectual equivalence" for distinguishing between expressions and derivative works, will be reflected in RDA rules for music cataloging.

(10.) Lynnsey K. Weissenberger, "Investigating Music Information Objects" (PhD diss., Florida State University, 2016), 60-62, http://purl.flvc.org/fsn/fd/FSU_2016SP_Weissenberger_fsu_0071E_13143.

(11.) For a more general summary of research on music information seeking over the past six decades, see Kirstin Dougan, "Music Information Seeking Opportunities Then and Now," in Trends in Music Information Seeking, Behavior, and Retrieval for Creativity, ed. Kostagiolas Petros, Martzokou Konstantina, and Lavranos Charilaos (Hershey, PA: Information Science Reference, 2016), 42-57. Music Information Retrieval research has devoted considerable time to studying popular music discovery, particularly in commercial environments. For more on popular music information behaviors see, e.g.: Jin Ha Lee, "Analysis of User Needs and Information Features in Natural Language Queries Seeking Music Information," Journal of the Association for Information Science and Technology 61 (2010): 1025-45, https://doi.org/10.1002/asi.21302; and Jin Ha Lee, J. Stephen Downie, and M. Cameron Jones, "Preliminary" Analyses of Information Features Provided by Users for Identifying Music," in Proceedings of the International Society for Music Information Retrieval, 2007, http://ismir2007.ismir.net/proceedings /ISMIR2007_p325_lee.pdf. For an example of discovery challenges in the context of a specific folk music tradition, see Lynnsey K. Weissenberger, "When 'Everything' Is Information: Irish Traditional Music and Information Retrieval," in iConference 2014 Proceedings, 496-507, 2014, https://doi.org/10.9776/14131.

(12.) James A. Tait, Authors and Titles; an Analytical Study of the Author Concept in Codes of Cataloguing Rules in the English Language, from That of the British Museum in 1841, to the Anglo-American Cataloguing Rules, 1961 (Hamden, CT: Archon Books, 1970), 140.

(13.) The Library of Congress developed the MARC (MAchine Readable Cataloging) format to automate cataloging for itself and other libraries. The MARC format has served as the primary data standard for United States Library cataloging ever since. For a detailed contemporaneous history, see: Henriette D. Avram, MARC; Its History and Implications (Washington, DC: Library of Congress MARC Development Office, 1975), https://eric.ed.gov/?id=ED127954.

(14.) Library of Congress, MARC Development Office, Music, a MARC Format: Specifications for Magnetic Tapes Containing Catalog Records for Music Scores and Musical and Nonmusical Sound Recordings (Washington, DC: Library of Congress, 1976).

(15.) Donald Seibert, The MARC Music Format: From Inception to Publication, MLA Technical Report, 13 (Philadelphia: Music Library Association, 1982), 3-4.

(16.) Seibert's definition of "fixed fields" included both those fields in the MARC format that require fixed-length data elements, and variable fields that contain fixed-length coded data. Seibert explained that these variable fields came to contain fixed-length data because the coded fields were inadequate, and so "spilled over into nearby portions of the Variable Field area." Seibert, 5.

(17.) Seibert, 5.

(18.) Jay Weitz and Neil Hughes, "MOUG Time Line," MOUG Newsletter 70 (September 1998): 10, http://musicoclcusers.org/wp-content/uploads/70Sep1998.pdf. The inaugural MOUG meeting, in 1978, boasted approximately 250 participants, a number that has yet to be surpassed as of 2019.

(19.) Nancy Bren, "Preconference: Computers and New Technology in the Music Library," MLA Newsletter 56 (April 1984): 5.

(20.) Lenore Coral, "Automation Requirements for Music Information," Notes 43, no.1 (September 1986): 14-18. It is sobering to read those requirements and note how often current-day systems do not fully meet them; for example, facilitating accurate, comprehensive author-title searching by indexing "dependent uniform titles (240. 700 $t and all following title subfields)" and by providing "combined author/title searching ... for 100/240, 100/245, and 700 St fields" (p. 17).

(21.) Michael Rogan, "Preconference Workshop: Music in at? Online Environment," MIA Newsletter 76 (April 198!)): 5.

(22.) Rogan, 4.

(23.) Music librarians eagerly implemented the MARC music format following its 1976 publication, though the Library of Congress delayed until 1985. Richard P. Smiraglia, "Foreword," in Music Coding and Tagging: MARC Content Designation for Scores and Sound Recordings, by Jay Weitz, 1st ed., Soldier Creek Music Series, no. 2 (Lake Crystal, MN: Soldier Creek Press, 1990), xv.

(24.) Of note: 511 questionnaires were mailed but only 38 (7.4%) were returned, so response bias may have skewed these results. Even so, there existed a body of libraries regularly coding the fields.

(25.) Jerry McBride and Joy Pile, "Survey on the Coding of 04x Fields," Newsletter: Music OC.LC Users Group 42 (November 1989): 6. The numbers increase slightly when including those with current plans to index the fields: 045, 8 percent; 047, 13 percent; 048, 16 percent.

(26.) Compilations, also known as aggregates, are found in all disciplines but especially common in music, where scores and recordings containing multiple musical works in a single physical item are the norm.

(27.) McBride and Pile, 8.

(28.) McBride and Pile, 8.

(29.) "045, 047, and 048 Fields Discontinued in LC Music Records," Cataloging Service Bulletin 55 (Winter 1992): 38. See also: Announcement of Library of Congress's proposal: David Sommerfield, "LC Report," Newsletter: Music OCLC Users Group 39 (May 1989): 22-24. A heated series of letters to the editor, including from Garrett Bowles, a chief architect of the MARC Music Format: "Letters to the Editor," Newsletter: Music OCLC Users Group 40 (August 1989): 11, 12, 17-19. MOUG's resolution desperately urging LC and other libraries to continue coding the fields: Linda Barnhart, "Minutes from the MOIJG Business Meeting February 20, 1990," Newsletter: Music OCI.C Users Group43 (May 1990): 13-14.

(30.) LCMPT terms were first available for assignment on 25 February 2014. LCGFT terms for music were approved February 2015, though genre/form terms had existed for other disciplines since 2007. See most current versions of the vocabularies at: https://www.loc.gov/aba/publications/FreeLCMPT/freelcmpt.html, and https://www.loc.gov/aba/publications/FreeLCGFT/freelcgft.html. The MLA Vocabularies Subcommittee released best practices for using the vocabularies, available at: http://www.musiclibraryassoc.org/mpage/cinc_mlabestpractices.

(31.) Music Library Association Automation Subcommittee, "Automation Requirements for Music Materials, Final Subcommittee Draft," 1997, never formally released. Further revisions were made through at least 2004 and made available on a now-defunct Web server at Indiana University.

(32.) David M. King summarized the published research to 2007: "Catalog User Search Strategies in Finding Music Materials," Music Reference Services Quarterly 9, no. 4 (2007): 1-24. Many music librarians also worked tirelessly for changes without leaving evidence in the published literature.

(33.) "NSCU Libraries Release Endeca-Powered Facetted [sic] Catalog," Library Hi Tech News 23, no. 2 (March 2006): 27. The Endeca software also powered popular e-commerce sites at that time.

(34.) Josh Hadro, "Summon Aims at One-Box Discovery," Library Journal 134, no. 3 (2009): 17-18.

(35.) Music OCLC Users' Group Reference Services Subcommittee, "WorldCat Local Enhancement Recommendations for Music" (April 2010), http://www.musicoclcusers.org/WorldCatLocal20100412.pdf.

(36.) Nara L. Newcomer, et al., "Music Discovery Requirements: A Guide to Optimizing Interfaces," Notes 69, no. 3 (March 2013): 494-524. The version published in this journal was slightly different, lacking appendices, etc.

(37.) Newcomer, et al., 495.

(38.) RDA was published in 2010, but most US libraries delayed implementation until the Library of Congress fully implemented RDA on 31 March 2013, using the intervening time for testing. Program for Cooperative Cataloging libraries were required to switch to RDA for all new BIBCO- and CONSER-standard bibliographic records by the end of 2014. Library of Congress, "Final Date for BIBCO and CONSF.R Bibliographic Contributions Formulated According to AACR2: December 31, 2014," http://www/loc/gov/aba/pcc/documents/Final-date-AACR2-PCC-Bib.html.

(39.) Beth Iseminger, et al., "Faceted Vocabularies for Music: A New Era in Resource Discovery," Notes 73, no. 3 (March 2017): 410.

(40.) The process will be helped by retrospective conversion, which passed a significant milestone with the 2018 public release of the Music Toolkit for generating faceted terms based on Library of Congress Subject Headings. Casey Mullin, Retrospective Implementation of Faceted Vocabularies for Music: Efforts Led try the Music Library Association and Recommendations for Future Directions: A Technical Report (19 April 2018), http://c.ymcdn.com/sites/www. musiclibraryassoc.org/resource/resmgr/BCC_Resources/RetrospectiveImplementation.pdf.

(41.) Kelley McGrath and Lesley Lowery, "Getting More out of MARC with Primo: Strategies for Display, Search and Faceting," The Code4Lib Journal 41 (9 August 2018), https://journal.code41ib.org/articles/13600.

(42.) Music Discovery Requirements Update Task Force members were: Nara Newcomer, University of Missouri-Kansas City (chair); Anna Alfeld LoPrete, Indiana University; Rebecca Belford, Brown University; Chris Holden, Library of Congress; Keith Knop, University of Georgia; Nancy Lorimer, Stanford University; Karen Peters, Bates College; and Patricia Sasser, Furman University (Web master). MDR2 released in January 2018, approved by the MLA board of directors in October 2017, and essentially complete in summer 2017, with most work accomplished in the preceding twelve months.

(43.) The maintenance plan is available on the "Music Discovery Requirements" pages of the MLA Web site, http://www.musiclibraryassoc.org/BlankCustom.aspPpage-mdr_es.

(44.) See the OCLC Music Toolkit, as announced at http://cmc.blog.musiclibraryassoc.org/2018/04/20/new-oclc-music- toolkit-for-generating-faceted-music-data.

(45.) Donald Krummel, "Musical Functions and Bibliographical Forms," The Library s5-XXXI, no. 4 (December 1976): 348-49.

(46.) Sherry L. Vellucci, Bibliographic Relationships in Music Catalogs (Lanham, MD: Scarecrow Press, 1997), 231.

(47.) Martha M. Yee, "What is a Work? Part 1, The User and the Objects of the Catalogue," Cataloging & Classification Quarterly 19, no. 1 (1994): 11.

(48.) Yee, "What is a Work?," 12.

(49.) Yee, "What is a Work?," 12-13

(50.) William Denton, "FRBR and the History of Cataloging," in Understanding FRBR: What It Is and How It mil Affect Our Retrieval Took, ed. Arlene G. Taylor (Westport, CT: Libraries Unlimited, 2007), 42.

(51.) Richard P. Smiraglia, Bibliographic Control of Music, 1897-2000 (Lanham, MD: Scarecrow Press, 2006), 6.

(52.) Smiraglia, Bibliographic Control of Music, 1897-2000, 6.

(53.) O. G. Sonneck, "Music," in Rules for a Dictionary Catalog, by Charles A. Cutter, 4th ed., US Bureau of Education. Special Report on Public Libraries; Pt. 2 (Washington, DC: Government Printing Office, 1904), 138-40.

(54.) Bulletin of the American Library Association, vol. XIV (Chicago: American Library Association, 1920), 295.

(55.) Sherry L. Velucci, "Uniform Titles as Linking Devices," Cataloging & Classification Quarterly 12, no. 1 (1990): 50.

(56.) Smiraglia, Bibliographic Control of Music, 1897-2000, 8.

(57.) Ruth Wallace, The Care and Treatment of Music in a Library (Chicago: American Library Association, 1927), 20.

(58.) Wallace, The Care and Treatment of Music in a Library, 20.

(59.) A.L.A. Catalog Rules: Author and Title Entries (Chicago: American Library Association, 1941), xxxi.

(60.) Code for Cataloging Music and Phonorecords (Chicago: American Library Association, 1958), 14-15.

(61.) Velucci, "Uniform titles," 53.

(62.) Seymour Lubetzky, "The Objectives of the Catalog," in Foundations of Cataloging: A Sourcebook, ed. Michael Carpenter and Elaine Svenonius (Littleton, CO: Libraries Unlimited. 1985). 189-90.

(63.) Lubetzky, "The Objectives of the Catalog," 190.

(64.) Lubetzky, "The Objectives of the Catalog," 191.

(65.) "Statement of Principles: International Conference on Cataloguing Principles, Paris, October 1961," in Foundations of Cataloging: A Sourcebook, ed. Michael Carpenter and Elaine Svenonius (Littleton, CO: Libraries Unlimited, 1985), 179.

(66.) Seymour Lubetzky, Code of Cataloging Rules: Author and Title Entry. Additions, Revisions, and Changes Prepared in Light of Discussions of the March 1960 Draft for the Consideration of the Catalog Code Revision Committee (Chicago: American Library Association, 1961), 1.

(67.) C. P. Ravilious, A Survey of Existing Systems and Current Proposals for the Cataloguing and Description of Non-Book Materials Collected by Libraries, with Preliminary Suggestions for Their International Co-ordination (Paris: United Nations Educational, Scientific and Cultural Organization, 1975), 58.

(68.) Martha M. Yee, "What is a Work? Part 2: The Anglo-American Cataloging Codes," Cataloging & Classification Quarterly 19, no. 2 (1994): 12.

(69.) Martha M. Yee, "What is a Work? Part 4: Cataloging Theorists and a Definition," Cataloging & Classification Quarterly 20, no. 2 (1995): 13.

(70.) Steven A. Knowlton, "Criticism of Cataloging Code Reform, As Seen In the Pages of Library Resources and Technical Services (1957-66)," Library Resources and Technical Services 53, no. 1 (2009): 15-24.

(71.) Yee, "What is a Work? Part 4," 4.

(72.) Martha M. Yee, "Integration of Nonbook Materials in AACR2." Cataloging & Classification Quarterly 3, no. 4 (1983): 5.

(73.) Yee, "Integration of Nonbook Materials," 10.

(74.) Sherry L. Vellucci, Bibliographic Relationships in Music Catalogs (Lanham, MD: Scarecrow Press, 1997). Richard P. Smiraglia, "Authority Control and The Extent of Derivative Bibliographic Relationships" (PhD diss., University of Chicago, 1992).

(75.) Richard P. Smiraglia, "Theoretical Considerations in the Bibliographic Control of Music Materials in Libraries," Cataloging & Classification Quarterly 5, no. 3 (1985): 11.

(76.) Vellucci, "Uniform Titles as Linking Devices," 55.

(77.) Martha M. Yee, "Lubetzky's Work Principle," in The Future of Cataloging: Insights from the Lubetzky Symposium, ed. Tschera Harkness Connell and Robert L. Maxwell (Chicago: American Library Association, 2000), 97.

(78.) Gregory H. Leazer, "Applying the Concept of the Work to New Environments," in The Future of Cataloging: Insights from the Lubetzky Symposium, ed. Tschera Harkness Connell and Robert L. Maxwell (Chicago: American Library Association, 2000), 119.

(79.) David H. Thomas and Richard P. Smiraglia, "Beyond the Score," Notes54, no. 3 (March 1998): 665.

(80.) Olivia M. A. Madison, "The Origins of the IFLA Study on Functional Requirements for Bibliographic Records," Cataloging & Classification Quarterly 39, no. 3-4 (2005): 18-19.

(81.) IFLA, Functional Requirements for Bibliographic Records (Munich: K. G. Saur, 1998).

(82.) Kathryn P. Glennan, "The Development of Resource Description & Access and Its Impact on Music Materials," Notes, 68, no. 3 (March 2012): 526-27.

(83.) IFLA. Definition of FRBRoo: A Conceptual Model for Bibliographic Information in Object-Oriented Formalism (The Hague, Netherlands: IFLA, 2016).

(84.) IFLA, III A Library Reference Model: A Conceptual Model for Bibliographic Information (The Hague, Netherlands: IFLA, 2017).

(85.) For one example, see: Mark Notess, Jon W. Dunn, and Juliet L. Hardesty, "Scherzo: A FRBR-based Music Discovery System," in Proceedings of the 2011 International Conference on Dublin Core and Metadata Applications (The Hague: Dublin Core Metadata Initiative, 2011), 182-83.

(86.) Richard P. Smiraglia, "The History of 'The Work' in the Modern Catalog," Cataloging & Classification Quarterly 35, no. 3-4 (2003): 565.

(87.) Joan Lippincott, "Net Generation Students and Libraries," in Educating the Net Generation, ed. Diana Oblinger and James L. Oblinger (Boulder, CO: EDUCAUSE, 2005): Bennett Claire Ponsford and Wyoma vanDuinkerken, "User Expectations in the Time of Google: Usability Testing of Federated Searching," Internet Reference Services Quarterly 12, no. 1-2 (2007): 176-77.

(88.) Karen Calhoun, et al. "Online Catalogs: What Users and Librarians Want: An OCLC Report," 2009, https://wwTV.oclc.org/content/dam/oclc/reports/onlinecatalogs/fullreport.pdf.

(89.) Of these, RISM's Themefinder (http://risin.themefinder.org/) is probably the most well-known.

(90.) Jin Ha Lee, Allen Renear, and Linda C. Smith, "Known-Item Search: Variations on a Concept," Proceedings of the American Society for Information Science and Technology 43, no. 1 (2006), https://doi.org/10.1002/meet.14504301126.

(91.) Music Library Association Cataloging and Metadata Committee, "Types of Composition for Use in Authorized Access Points for Music: A Manual for Use with RDA," https://www.musiclibraryassoc.org/BlankCustom.asp?page=cmc_accpointsrda.

(92.) Todd Decker, "'Scarlattino, the Wonder of His Time': Domenico Scarlatti's Absent Presence in Eighteenth-Century England," Eighteenth-Century Music 2, no. 2 (2005): 274-75, https://doi.org/10.1017/S1478570605000382.

(93.) For a more detailed, though hardly impartial, account, see Joel Sheveloff, "Domenico Scarlatti: Tercentenary Frustrations," Musical Quarterly 71, no. 4 (1985): 399-436. Sheveloff has also been critical of Kirkpatrick.

(94.) Isabelle Cazeaux, "Classification and Cataloging," in Manual of Music Librarianship, ed. Carol June Bradley (Ann Arbor, MI: Music Library Association, 1966), 47.

(95.) Richard P. Smiraglia, Cataloging Music (Lake Crystal, MN: Soldier Creek Press, 1986), 162.

(96.) Richard P. Smiraglia, Music Cataloging: The Bibliographic Control of Printed and Recorded Music in Libraries (Englewood, CO: Libraries Unlimited, 1989), 99-100.

(97.) Geraldine E. Ostrove, "Music Subject Cataloging and Form/Genre Implementation at the Library of Congress," Cataloging at Classification Quarterly 32, no. 2 (2001) : 92.

(98.) For one example of how the 382 is prepared for patron use, see: Kelley McGrath and Lesley Lowrey, "Getting More out of MARC with Primo: Strategies for Display, Search and Faceting," Codr4Lib Journal 41 (2018), https://journal.code41ib.org/articles/13600.
COPYRIGHT 2019 Music Library Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Holden, Chris; Knop, Keith; Newcomer, Nara
Publication:Notes
Geographic Code:1USA
Date:Jun 1, 2019
Words:13097
Previous Article:MUSIC RECEIVED.
Next Article:REPRINTS REVISITED: RECOLLECTIONS OF A PUBLISHER (AND A USER).
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters