Tillbaka

Additional information needed by libraries and users

Deliverable 1.3.2

1 Introduction
1.1 What is this project about?
1.2 How was the project carried out?
1.3 Presentation of the results

2 Factors affecting the information management of music
2.1 Introduction
2.2 Subject analysis
        2.2.1 Facets of subject information
        2.2.2 Additional facets needed for music
        2.2.3 Performance information
2.3 Other factors affecting catalogue access
        2.3.1 Language
        2.3.2 Collections and anthologies
        2.3.3 Music users
        2.3.4 Summary

3 Classification schemes in use in music libraries in the EU countries and USA
3.1 Introduction
3.2   History and nature of music classification schemes
3.3 Outline of the three major classification schemes
        3.3.1 Dewey Decimal Classification (DDC)
        3.3.2 Revisions of Dewey
        3.3.3 Universal Decimal Classification
        3.3.4 Library of Congress
        3.3.5 Conclusion
3.4 Classification schemes developed for particular libraries or library sectors
        3.4.1 Dickinson classification
        3.4.2 Systematik der Musikliteratur and Musikschriftums für Öffentliche Musikbibliotheken (SMM)
3.5 Other schemes
3.6 Classification schemes for sound recordings
3.7 Index or classification

The aim of this project was to establish

• What additional information beyond composer and title is required for access to music information
• How the classification schemes used by EU and US and Canadian libraries cope with music
• What music thesauri and subject heading lists are available and how these cope with music information
• What recommendations can be given for the use of classification schemes, subject heading lists and music thesauri

The project involved

• A review of available recent literature on these subjects
• An examination of the principal classification schemes used by music libraries, archives and documentation centres in the EU, USA and Canada
• An examination of the available subject heading lists and music thesauri and of the developments in online subject searching currently being developed and how these might affect music information retrieval

The results of this research are presented in two deliverables: this deliverable (1.3.2), and deliverable 1.3.3. Deliverable 1.3.2 examines the additional information required for music information retrieval and looks at the principal classification schemes used. It examines how they deal with music materials and assesses their value for music. It looks at what other music classification schemes are in use for music in the EU, USA and Canada and assesses the reasons for the development of such schemes and the potential for further development. It makes recommendations on further action on classification

Deliverable 1.3.3 examines the available subject heading lists and music thesauri and how current online searching developments in areas other than music information retrieval may require increased use of such tools and makes recommendations for further research in that area.

Back to table of contents

In his seventeenth-century treatise Advis pour dresser une bibliothèque, Gabriel Naudé describes the purpose of a catalogue in elegantly simple terms:

‘On peut voir et sçavoir en un clin d’oeil tous les Autheurs qui s’y rencontrent sur le premier sujet qui viendra en fantasie; ... et ... toutes les oeuvres de certains Autheurs’ (Naudé 1876, p.107)

The English diarist John Evelyn rendered this advice

‘One may see and know in the twinkling of an eye, all the authors which do meet there upon the first subject that shall come into one’s head ... [and] know ... all the works of certain authors.’ (Evelyn 1903, p.150)

Though simple to describe, achieving this aim is not easy. Even where a given composer and title are known, there are many complexities in forms of name and title, and in the prolific production of many composers, that make the task of cataloguing a complex one. The international nature of music can make more problems for the cataloguer who must deal with the same work in many languages. The interposing of performer between creator and user adds an additional layer of responsibility to be reflected in catalogue access. This report, however, is not concerned with the author/title approach but with the information required to help music users find what is usually referred to as subject information.

Subject analysis is a way of describing the topicality of an item. It is about ‘aboutness’. Traditional subject cataloguing theory identifies seven main facets of subject information., including.

• Intellectual Content
• Time
• Place
• Intended audience
• Form
• Genre

For example, a school orchestra edition of Vaughan Williams Sea symphony might be described as

• Sea music (Intellectual content)
• 20^th century(Time)
• British(Place)
• For schools(Intended audience)
• Symphony(Form)
• Orchestral music(Genre)

These standard facets would cope with a request from, say, a music producer for sea music to accompany a documentary.

In this example, the inclusion of the term ‘sea’ in the title makes it easy to categorise this item as ‘sea music’. It is possible to argue that this is the ‘subject’ of this item. But there are many other pieces which might reasonably be considered aural descriptions of the sea - for example, Debussy’s En bateau - which would not have ‘sea music’ as a description of their intellectual content.

This use of music to describe an emotional mood is extremely common - especially in the film, television and multimedia world - and might be called ‘ambience’.

In a similar way, the requirement for music for a specific medium of performance, or a specific duration, is not covered by the traditional facets of subject cataloguing. For example, a musician seeking music for 2 violins, viola, cello, flute, oboe, bassoon and french horn, lasting about 10 minutes and appropriate for a 19^th-century period drama will not find this information covered by author/title or subject cataloguing.

Imagine that the work described in 2.2.2 is simply called ‘Oktet’ and published in Leipzig with a German title page. If catalogued in a library in an English-speaking country the uniform title constructed by the cataloguer to describe the work would not list the instruments individually, but as instrumental families - ie strings, winds, brasses. The names of the instruments may be transcribed from the title-page, but will be in German, and not helpful to the English-speaking user. The English names of individual instruments - if included at all - will have to be included as a note.

Some works have a distinctive title. In this case, there may be no details of instrumentation given. For example, imagine a French cantata with the title ‘La Bibliothèque’ for chorus with accompaniment for piano and string quartet, published in Rome in an Italian translation. A cataloguer in any country will provide a uniform title in French to ensure that all editions of this work - in whatever language - are located together in the catalogue. But this title will give no details of the instrumentation. The title page, if it includes details of instrumentation, will include them in Italian and, perhaps, one or two other languages. But this will not satisfy the needs of users in every country in which the work might be used. With works of this type, there may not even be a note detailing the performing forces required.

It may be argued that such information is beyond the scope of the catalogue. The problem with this argument is that performing needs are included in some catalogue entries and the user, therefore, comes to expect that such information will be available for all catalogue entries. It is unreasonable to expect the user to distinguish between works whose catalogue entries may include performing requirements and those which may not. It is, therefore, reasonable to regard such information as ‘subject’ information.

It should be obvious from the preceding discussion that part of the problem was created by language differences. The international nature of music means that music cataloguers (and users) will find themselves dealing with far more ‘translations’ than is the case with most other areas of library cataloguing.

Frequently, the translation will only apply to the title-page. The musical content of the same work published in different countries with title-pages in different languages will be identical, and will be equally comprehensible to a user in either country. This is not the case with many literary translations, where the textual content will also be in a different language, If international cataloguing co-operation is to be achieved, multilingual thesauri - offering machine translation of certain areas of a catalogue record (especially subject data) are needed.

The tendency in the music publishing and recording business to produce individual units for sale which contain several individual - often unconnected - musical works is another factor which requires careful consideration when providing subject access to music materials.

Traditional catalogue description requires a single entry to represent a single physical unit. A recording of chamber music may contain a work such as the ‘Oktet’ described earlier (for 2 violins, viola, cello, flute, oboe, bassoon and french horn) and a work for violin, viola, cello, and clarinet. If this recording is catalogued using a single bibliographical record - as provided for in most cataloguing rules - notes must be included in the catalogue record to indicate the instruments used in each work Because these notes form part of one bibliographical record, a keyword search for music for flute, oboe, clarinet, bassoon and french horn will return this catalogue record - which contains all these instruments - even though neither work matches the instrumentation sought.

It is possible to devise strategies which will surmount this problem, but they require a considerable level of user education. Though this might be possible within a small institution, it is much less feasible in a large public library, and with increasing remote access to library catalogues across the World Wide Web, simply not possible for such users. The trend in catalogue interface design is towards systems which are intuitive and require minimal pre-knowledge.

The range of music users is very broad. This is a product of the entertainment function of music. Many music users are interested amateurs who exist alongside the range of professional musicians. Whilst it is true that there are general readers in many subject areas, the proportion of non-professional music users is high. Moreover, these amateur music users will make use of many of the same documents as the professional musician - though their approach to the material will not necessarily be informed by the same level of knowledge or, most importantly, the same vocabulary.

To put this another way, there are many medical professionals who are enthusiastic and knowledgeable musicians and music users and who may well use the same recordings or miniature score or sets of parts as the professional musician. In contrast, there are comparatively few musicians who could claim to be knowledgeable about medical matters at anything like the level of the medical professionals - and even those with a lively interest are more likely to approach the subject through material aimed at a general audience and not through the same documentary sources used by the medical professional.

The implication of this for the music information community is a wider range of approaches to searching for music materials, and often using common language terms rather than a specialised professional vocabulary.

• An academic may seek
  Russian chamber music of the early 19^th century
• A performer may want
  Music for violin, cello, bassoon, horn and harp
• A teacher may want
  Advanced grade pieces between 1750 and 1900
• A producer may be looking for
  19^th century salon music

Despite the wide range of topics sought in these requests, all might be satisfied by Glinka’s Serenade on themes from Donizetti’s ‘Anna Bolena’. Providing access by all these routes will require a detailed level of subject indexing.

Access to music information by means other than composer and title requires not only the use of subject categories common to most cataloguing, but also details of

• Performance medium
• Duration
• General style or ambience (as opposed to specific genre)

Additionally,

• The nature of music and record publishing frequently results in documents containing several individual musical items. The tradition of bibliographical description on which library cataloguing is based is not suited to this publishing tradition and may create false results when searching for subject or performer information
• The international nature of music will result in a varied range of languages on music title-pages even in the smallest music libraries. This requires the provision of information consistent in language and form.
• The variety of users (and uses) of music will also demand a wide range of subject access points for any particular musical item.

Back to table of contents

3.1 Introduction

Harmonica deliverable 1.2.3 gives the results of a questionnaire investigating the cataloguing and classification schemes used in the EU and USA. Three schemes predominate:

• Dewey Decimal Classification - used in ten countries in one form or another
• Library of Congress Classification - used in five countries
• Universal Decimal Classification - used in four countries

Another nine schemes are used in only one country.

• Classification des discothèques publiques
• Dickinson Classification
• Eppelsheimer Systematik*
• Klassifikation für Allgemeinbibliotheken: Teil Tonträger Musik*
• Klassifikation für Allgemeinbibliotheken: Musikalien*
• Klassifikationssystem för Svenska Bibliotek
• Schema voor de indeling van de systematische catalogus in openbare
  bibliotheken
• Systematik für Öffentliche Musikbibliotheken: Systematik des Musikschriftums und der Musikalien*
• Tonträger-Systematik Musik*

Five of these nine schemes, those marked with an asterisk, are used only in Germany, which further emphasises the dominance of the three schemes listed. Furthermore, six of these schemes are designed either for sound recordings - not well provided for in the major schemes - or for public libraries.

The needs of a public library may often be met by a simplified scheme. In a provocative article published some 15 years ago, Brian Redfern has argued that a simple arrangement by composer might well satisfy the needs of many public library users. (Redfern 1985). Whilst this might be oversimplified, it is true that some of the schemes discussed below designed principally for public libraries use a simple division by medium of performance then arrangement by composer. This system is adapted for sound recordings to subdivide material by genre (pop, jazz, classical and so on) and then arrange by composer or artist.

The earliest attempts at systematic classification of music materials tended to be music publishers’ catalogues . Krummel 1984 reviews a number of these early catalogues or bibliographies and finds that the problem of organization of material in these items is generally solved in one of two ways.

• Subdivision by form or genre, further arranged by medium of performance
• Subdivision by medium of performance, further arranged by form or genre.

Thus a catalogue of vocal music might be arranged

• Madrigals
  Madrigals for 2 voices

  Madrigals for 3 voices
  Madrigals for 4 voices
          Madrigals for 4 equal voices
          Madrigals for 4 mixed voices

. . .

• Masses
  Masses for 4 voices
  Masses for 5 voices
          Masses for 5 voices unaccompanied
          Masses for 5 voices accompanied

etc.

Or it might be arranged

• Music for 2 voices
  Madrigals
  Masses
  Motets

. . .

• Music for 3 voices
  Madrigals
  Masses
  Motets

etc.

In practice, because musical forms often presuppose a medium of performance - at least at the fundamental level - it might be argued that division by medium of performance is the principal arrangement used in these catalogues. For example, madrigals and masses presume a performance by voices in most cases. And even a general catalogue containing music of all kinds, superficially arranged by form (Symphonies, Cantatas, Oratorios, Concertos) may have an underlying division between instrumental and vocal music.

As we discussed earlier, there are more facets than medium of performance and form, but these two predominate in classified arrangements. A failure to acknowledge the primacy of medium and form is often at the root of the failure of general classification schemes.

Before looking at how the three main schemes deal with music materials, two general points must be made.

• None of the major schemes copes adequately with sound recordings, though all have been used to classify recordings. This matter is considered later.
• The nature of music publishing - with its varying formats (printed and recorded) and large number of collected editions and miniature scores - has led many libraries to make a principal division by size. Books, music and recordings are separated from one another; large size books and scores are separated from standard size items; and miniature scores are often separated into a third size category. The major schemes acknowledge this in their provision.

3.3.1 Dewey Decimal Classification (DDC)

DDC remains one of the most popular classification schemes in the world. It has maintained this dominant position for many years. Twenty years ago, Dahlberg 1977 recorded that DDC was in use in over 100 countries.

Using an entirely numeric notation - perhaps one of the reasons for its popularity, given the international use of Arabic numerals - DDC divides knowledge into ten broad categories, each of which is further subdivided. In this arrangement, music is classified at 780-789, an artificial restriction necessitated by the enumerative nature of Dewey, but which created many problems for the scheme’s editors.

In editions up to the 19^th, the basic structure of the music scheme was confused. This fact was recognized early on. Even in the 1920s, the British music librarian Lionel McColvin found it necessary to revise the music schedule for use by public libraries. (McColvin 1937).

The initial divisions at 780 and 781 include physical format - with miniature scores and collected works separated out. At 782 and 783, the principal division is by genre - with dramatic music followed by sacred music, and within these groups a division by form (opera, oratorio, mass etc.). At 784-789 division is by medium of performance and then by form. Elliker 1994 argues that one might consider the genre of this section to be secular music. This allows him to propose a basic structure

• Genre S Medium (if applicable) S Form

The separation of dramatic music from secular or sacred is an artificial division, and there is no attempt to separate scores from music literature. In practice, most libraries make such a division and use some additional notation, such as ‘Mus.’ to indicate printed music.

The inadequacies of DDC for music were recognised early. One of the first attempts to revise the scheme was made by Lionel McColvin in 1924. With Harold Reeves, he published a revised version in 1937, which was further revised by Jack Dove in 1965.

This scheme separates books from scores and essentially subdivides scores by medium of performance. Works are them arranged in alphabetical without further subdivision. The scheme remains in use in a number of music libraries in the United Kingdom and Ireland.

In 1980, a revision of DDC by Sweeney and Clews was released for consultation, and eventually incorporated , with some modifications, into the 20^th edition (1989). This revision incorporates principles of faceted classification, as expounded by Ranganathan, and owes much to the British Catalogue of Music Classification. (This latter scheme was developed in 1957 for the music section of a national bibliography, but never used for practical library classification.)

The fundamental structure of this revision acknowledges the historical importance of medium as a principal subdivision. The scheme recognises three main facets: executant (medium), forms and character, with additional facets for time, place and format. The citation order used leads to a basic structure

• Medium S Form S Genre

This revision is fundamental in many ways, not least in its move from an enumerative to a faceted approach.

UDC grew out of, and was intended to be as compatible as possible, with DDC. Its principal difference lies in the ability to link together more than one number to express relationships between aspects of the classification. Thus, a quartet for piano, violin, viola and cello would combine the notation for quartets (785.74) with those for piano (786.2), violin (787.1), viola (787.2) and cello (787.3) to give a classification of 785.74 : 786.2 : 787.1 : 787.2 : 787.3.

As UDC is based on DDC, it suffers from some of the same problems, especially the placing of dramatic and sacred music at the same hierarchical level as the structurally higher vocal music, and the fundamental division by genre rather than by medium.

LC recognises the division of music from music literature. It uses an alphanumeric notation, using the prefix M for printed music, ML for books on music and MT for music instruction materials. These prefixes are mnemonic in English (M = Music; ML = Music literature; MT = Music Teaching), but would not be so in other languages (which may militate against a wider adoption of LC). Divisions within each letter group are numeric, with up to 4 digits available (with occasional further subdivision into decimal places). This allows a far greater enumerative range than is possible in DDC.

The principal subdivision is by medium of performance. Instrumental music is divided by size of ensemble, moving from smaller to larger, and then by instrumental family. Beyond that, division is by form, and then by format. Vocal music is subdivided by genre into sacred and secular and then further subdivided by medium, form and format. This structure might be represented as

• Medium S (Genre) S Form S Format

LC was not designed to accommodate sound recordings. It is used to classify recordings, but somewhat artificially, since there is no classification number provided for recording as format. The approach usually taken is to use the number allocated for full score. This is seen as the principal class number for a particular type of work. Thus, a recording of an opera will be placed at M 1500, as will the full score. The vocal score for the same item will be placed at M 1503. There is a fundamental flaw in adopting a number specified for one format for use by another, very different format. Although the fact that, in most music libraries, recordings will be physically separated from the scores means that this duplication of numbers will not create serious problems when shelving or browsing, it makes the classified catalogue inherently inaccurate and misleading.

LC was developed for one very substantial library collection, and has extremely detailed subdivision necessary to control the enormous volume of material in such a collection. In many smaller collections, such extensive subdivision can actually confuse the process of organising material.

Of these three schemes, LC and the 20^th edition of DDC both adopt the traditional approach of dividing music by medium of performance and then by form. Both serve the needs of large, general libraries well. They are less well adapted to specialist collections, and they fail to deal adequately with sound recordings - even though recording is an element of the format facet in DDC20. Both were developed in the US and have an English-language bias (although DDC does exist in several other language versions).

UDC suffers from the problems of earlier versions of DDC. Its ability to combine separate elements of the classification by using more than one classification number helps to provide more specific classification but at the cost of sometimes lengthy and not-easily memorable classmarks.

These factors undoubtedly contributed to the development of other schemes, as listed in paragraph 4.1.1. It has already been pointed out that most of these schemes attempt to deal with the problems of sound recordings or of providing for the needs of a particular library sector.

Elliker 1994, in reviewing the classification of music in some 24 classification schemes, identifies 10 schemes devised specifically to deal with music (as opposed to general schemes which allocate a section to music). Most of these were developed for an individual library.

The Dickinson Classification is one such, having been designed for Vassar College in the USA. It is restricted to music scores and almost entirely to Western art music. Interestingly, Dickinson does not prescribe a citation order, but allows a library to choose an order appropriate to its needs. This is an unusual approach that recognises that specialist libraries may have very different needs.

Elliker 1994 reduces the five systems Dickinson suggests to two fundamental structures, both based on medium as a principal subdivision, and then further dividing by form, then composer, or by composer then form. In principle, this differs little from McColvin’s simplification of Dewey.

Systematik der Musikliteratur and Musikschriftums für Öffentliche Musikbibliotheken (SMM) was developed by Dr. Alfons Ott, Director of the Städtische Musikbibliothek, München. Like McColvin, it was designed for public libraries.

It divides music into 22 main classes, each represented by an upper-case Roman letter. Certain material is separated out by format (eg. General collections, Collected editions, miniature scores, performance sets, and piano reduction scores). It will noticed that this is the same kind of material separated out by the main schemes discussed earlier. All other material is divided in SMM by medium. Pedagogical material is separated out at the beginning of each class, but apart from that the arrangement within each class is by composer. A further subdivision can be provided by separating out arrangements. This is achieved by adding the letter O (Original) or B (Bearbeitung) to the end of the classmark.

This simplification of arrangement is followed in other schemes developed for public libraries in most of the countries surveyed. Indeed, schemes are devised for individual libraries - perhaps indicating the inadequacy of the major schemes. In France, two schemes in particular have emerged in the past 15-20 years. One of these is designed for sound recordings in public libraries, and is discussed in section 3.6. The other system was developed by the National and Regional Conservatoires - whose specific requirement was for detailed access by medium of performance. This was provided by means of an indexing system which could then be incorporated into the classmark. The basic indexing system devised in the mid-1980s is

100 Solo instruments, duos (also indexable in 200) and single instrument ensembles (eg. flute choirs)
200 Chamber music
300 Music arranged chronologically
400 Orchestral music
500 Vocal music

To this basic scheme have been added

600 Music theory and composition

And at least one library has extended further to include

700 Dance

Within each broad category subdivisions cover a wide range of instrumental and vocal combinations. For example,

110 Keyboard music
111 Piano
112 Harpsichord
113 Organ

...

150 Percussion

Combinations are often mnemonic. Thus, from the above notation can be deduced

111.12 Piano and harpsichord
111.13 Piano and organ
111.5 Piano and percussion

By the addition of a three or four letter code for the name of a composer, and possibly a prefix to indicate type of material, a flexible classification can be achieved. For example, at the Conservatoire National Supérieur de Musique at Lyon, a set of parts for a Mozart string quartet would be classed

M 246 MOZ

M = Music intended for performance
246 = Index code for string quartets
MOZ = first three letters of composer’s name

By changing the order of elements, and the format abbreviation, a CD of the same work could be placed in an alphabetical sequence of CDs with the classmark

MOZ 246 CD

Thus, though devised for printed music, the flexibility of the indexing system has allowed simple adaptation for sound recording - or indeed videos or other multimedia formats.

Some of the systems listed in paragraph 3.1 have been developed specifically to cope with sound recordings. It has already been noted that LC, DDC and UDC do not cope well with sound recordings. The nature of the recorded product means that many items are compilations comprising several different items of music. Even where these items are all by the same composer, there is no certainty that they will be for the same medium or in the same form. In many cases, the linking factor in a recording is not the composer or the musical form but the performer. This produces many items which simply do not fit into a neat classification and which require multiple classmarks.

In addition, the security considerations posed by the recorded medium - the ease with which it could be stolen and the difficulty of marking the physical item - led many libraries to store recorded material on closed access long after this had ceased to be the case for books and music.

Where recorded music was displayed, it was often in a style borrowed from the record store, with broad categories such as Symphonies, Concertos, Jazz, Pop, Rock etc., and almost always a category called ‘Recitals’ or something similar, which would contain compilations by well-known artists. Subject access to recorded collections was often impossible.

This was clearly not an acceptable situation, and the development of classification schemes for recordings is to be welcomed. One such scheme well described in the literature (Smiraglia 1989, pp.114-9) is the Alpha-Numeric Scheme for Classifying Recordings (ANSCR). This divides recorded music into a number of broad categories each of which is identified by an uppercase Roman letter (in a similar way to SMM). Further subdivisions can be made. Genre is the principal subdivision here. Although this scheme is not mentioned in the replies to the questionnaire considered in Deliverable 1.1.3, it is commonly used in the United States, and by a number of UK libraries. [Two examples:
http://www.calstatela.edu/library/guides/anscr_class.htm, http://www.skokie.lib.il.us/s_audiovisual/av_music/anscr.html added by AL]

As was noted in section 3.5, the French libraries have successfully developed a scheme for use in french public libraries. The ‘Principes de classement des documents musicaux dans les bibliothèques publiques’, as it is called in Sineux 1993, was originally developed for classifying recordings at the Discothèque de France in the mid 1970s. After refinement and further experimentation, the version published in 1993 also provides for using the scheme for classifying printed music. In describing the scheme, Dominique Bertrand comments: ‘Cette classification est conçue en fonction du public non spécialisé des médiathèques. Elle se rapporte donc de la clientèle des disquaires et n’a pas de prétension scientifiques. Elle se veut avant tout empirique’. (Cousin 1995, p.29)

Nine principle divisions are identified

0 National and traditional musics
1 Jazz and blues
2 Rock
3 Classical music
4 Contemporary (art) music - from 1945
5 Functional music - Miscellaneous
6 Non-music recordings
7 Recordings for children
8 Science of music; musical techniques

Subdivisions are used to indicate musical form or medium; to indicate countries or regions and chronology; or, in class 5, to indicate function or intended audience. Thus, Beethoven string quartets would be classed at

3.BEE.14.40

3 = Classical music
BEE = first three letters of the composer’s surname
14 = quartets (from the form subdivisions)
40 = strings (from the instrument subdivisions)

A recording of the soundtrack of Mort de Venise would simply be classed at

520 MOR

5 = Functional music
20 = Film
MOR = first three letters of the title

It will be seen from these two examples that the scheme is extremely flexible and can accommodate either a simple alphabetical approach, or a systematic approach by form or medium.

Bertrand comments: ‘Dans l’ensemble, la classification privilégie les formes musicales ... ou les styles ... plûtot que les formations instrumentales. Elle se rapproche en cela du classement adopté par les disquaires que celui des professionels practiciens de musique. (Cousin 1995, p.28)

Despite this, and although designed principally for sound recordings, the flexibility of the scheme has encouraged adaptations for use with printed music and books. In Cousin 1995, Clément Riot describes his adaptation of the scheme for the Perpignan Conservatoire library. Standard repertory or ‘well-known’ composers are arranged alphabetically (like the Beethoven example above). In the case of lesser known music, an arrangement by form or medium is used.

One of the more interesting aspects of the French systems considered in section 3.6 is the use of indexes to define the more specific elements of the musical subject material whilst allowing the use of locally defined notation for more general elements (such as format) and, thereby, creating a flexible citation order which can encompass a simple alphabetical arrangement of material as well as a more complex, systematic arrangement. A similar use of indexing is seen in a slightly different context in the German Verzeichnis lieferbaren Musikalien (VLB), which has devised an exhaustive - and exhausting! - index of all possible instrumental combinations for use in cataloguing the output of German music publishers.

Whilst such indexes are not easily memorable for the user, there is the possibility to mount indexes as part of the catalogue, enabling the user to search using a browse facility. A similar approach to music thesauri is discussed in Deliverable 1.3.3. It may be that we shall see a further blurring of the distinction between classification and indexing in coming years.

Back to table of contents

4.1 Music classification is dominated by the three major schemes which also dominate book classification. A number of factors contribute to this.

• Many music libraries are part of larger institutions for whom the use of a single classification scheme makes much sense.

• A number of central cataloguing agencies provide LC or DDC classification in their catalogue records or in Cataloguing-in-publication data and the use of such predetermined classification is clearly a significant saving.

• The re-classification of a significant sector of a library is an expensive and time-consuming operation not lightly undertaken. There is, therefore, an inevitable inertia which maintains the large systems in place.

4.2 Of the three main schemes, LC has historically coped most satisfactorily with music. DDC20 has made significant and welcome changes to what was a very unsatisfactory classification scheme for music, and now looks to have a scheme capable of matching LC. It must be open to question, however, how many libraries will adopt DDC20 wholeheartedly, including the substantial reclassifying necessary, because of the financial implications. On the other hand, the drive towards retrospective conversion of catalogue records for inclusion in computer catalogues may have the advantage of simultaneously enabling re-classification. UDC is, in some respects the least satisfactory of the major schemes for music, suffering the disadvantages of the earlier editions of Dewey from which it originates.

4.3 None of the major schemes copes well with sound recordings. These have traditionally remained unclassified in many libraries, or been the subject of in-house schemes. The development of schemes for recorded music is to be welcomed. It has been noted that DDC20 makes some provision for recordings.

4.4 It is recommended that publishers of classification schemes be encouraged to consider how their schemes can be developed to include sound recordings, in the same way that new subject areas are incorporated into their schemes.

4.5 The extensive and detailed classification found in the major schemes is often too great for smaller or more general collections. This has led to the development of in-house or regional schemes which simplify the classification process, whilst, in general, adopting a similar approach to organising information.

4.6 Publishers of major schemes should be encouraged to authorise development of simpler schemes, which would then be consistent with the major schemes but more easily applicable in general libraries. This would be analogous to the varying levels of descriptive cataloguing authorised in some descriptive cataloguing rules.

4.7 Amongst the schemes developed specifically for music, those developed in France in the 1980s have taken particular cognisance of the importance of sound recordings in many music libraries - especially in the public sector. The empirical and flexible approach that they have adopted offers potential for development elsewhere. The simple, and fairly intuitive, nature of the schemes is a further attraction.

4.8 No one classification scheme can or should be recommended as being ideal for music. The economic factors involved in re-classifying substantial collections create a practical barrier to substantial changes in direction. However, the needs of music should continue to be brought to the attention of the editors of major schemes and further improvements sought.

4.9 Whilst the major schemes may satisfy the requirements of many public and academic libraries, the specialised, and sometimes unique, requirements of some music libraries, archives and documentation centres will continue to lead to the development of in-house systems to satisfy their particular requirements.

Back to table of contents

AGOSTINI 1997
Agostini, F.                ‘Optimisation d’accès dans un catalogue plurimedia’. In: Bulletin d’informations de
                                   l’Association des Bibliothècaires Français, 174, pp.160-4

COUSIN 1995
Cousin, N. (ed.) Guide pratique des classifications musicales à l’usage des bibliothèques. Paris: Association
                                des bibliothèques, archives et centres de documentations musicaux. Groupe français.

DAHLBERG 1977
Dahlberg, I.               ‘Major developments in classification’. In: Advances in music librarianship 7, pp.41-103

ELLIKER 1994
Elliker, C.                  ‘Classification schemes for scores: analysis of structural levels’. In; Notes, 50, pp.1269-                                    1320

EVELYN 1903
Evelyn, J.                     Instructions concerning erecting of a library ... interpreted by Jo. Evelyn, Esquire.
                                   [London, 1661.] Cambridge: Hoghton, Mifflin

GOLDTHWAITE 1948

Goldthwaite, S.         ‘Classification problems in bibliographies of the literature about music’.
                                In: Library Quarterly, 18, pp.255-63

KRUMMEL 1984
Krummel, D.            ‘The origins of modern music classification’. In: Elvers, R. (ed.) Festschrift Albi Rosenthal.
                                Tutzing: H. Schneider, pp.181-98

NAUDÉ 1876
Naudé, G.                 Advis pour dresser une bibliothèque ... Réimprimé sur la deuxième ed. (Paris, 1644).                         Paris: Liseux

REDFERN 1978
Redfern, B.                 Organizing music in libraries. Vol.1: Arrangement and classification. London: Clive Bingley

REDFERN 1985
Redfern, B                 ‘Dinosaurs to crush flies: computer catalogues, classification and other barriers to catalogue .                                   use’ In: Brio 21(1), pp.4-8.

SMIRAGLIA 1989
Smiraglia, R.             Music Cataloguing. Englewood, Col.: Libraries Unlimited

VLB                          Verzeichnis lieferbaren Musikalien. Frankfurt-am-Main: Buchhändler-Vereinigung                                    (2 issues per annum).

Back to table of contents

DEWEY 1979
Dewey, M.             Dewey decimal classification and relative index. 19^th ed.. / ed. By Benjamin A. Custer. Albany, N.Y.:                                  Forest Press, 1979.

DEWEY 1980
Sweeney, R & J. Clews DDC: Dewey decimal classification proposed revision of 780 Music based on Dewey decimal                                          classification and relative index. Albany, N.Y.: Forest Press, 1980

DEWEY 1989
Dewey. M.            Dewey decimal classification and relative index. 20^th ed.. / ed.
                              by John P. Comaromi. Albany, N.Y.: Forest Press, 1989.

DICKINSON
Dickinson, G.S.      Classification of musical composition: a decimal-symbol system. Poughkeepsie, N.Y.:  Vassar College, 1938

LC                          Library of Congress classification. Class M: Music and books on music. 3^rd ed. Washington, D.C.:                                 Library of Congress, 1978.

MCCOLVIN 1937
McColvin   L. R. & H. Reeves.    Music libraries: their organization and contents, with a bibliography of music and                                                      musical literature. London: Grafton

PCDMPB              ‘Principes de classement des documents musicaux dans les bibliothèques publiques’. In: Sineux, M. (ed.)  
                              Musique en bibliothèques. Paris: Editions du cercle de la libraire, 1993. Pp.227-56

UDC             Universal decimal classification. English full edition: UDC 78 Music.                   
                            London: British Standards Institute, 1978

Back to table of contents

Tillbaka