10.4.1 Boggs and Lewis and American Geographical Society Classification Systems

These two classification systems, S.W. Boggs and D.C. Lewis (B&L), and the AGS, are quite specific to cartographic materials and not part of an overall classification system. “The Classification and Cataloging of Maps and Atlases,” more widely known as the B&L classification, was developed to satisfy needs of the U.S. State Department's Map Library and as such did not have a North American bias (Romero & Romero, 1999). B&L was the first system devoted to maps, atlases, relief maps, and globes to be formalized in a publication (Boggs & Lewis, 1945). The disadvantage of this classification was that in spite of re-printings, there were no updates after its creation in 1945.

B&L used the 1941 ALA cataloging rules and emphasized the importance of order for descriptive elements. First was a three-digit number representing area, second was a letter representing the subject and location symbols, third was the date of situation, and finally the type of map, author, and title (Abresch, Hanson, Heron, & Reehling, 2008). While not the oldest classification scheme, it was the first specific classification for maps and was popular in Canada and Australia (Larsgaard, 1998).

The American Geographical Society of New York was a 19th century professional group of geographers who devised the map classification for their collection. When the group disbanded, the 1.3 million items went to the University of Wisconsin at Milwaukee (University of Wisconsin Milwaukee Libraries, 2016b). The library began a digitization project in 2001, and the impressive Digital Collections can be viewed online (University of Wisconsin Milwaukee Libraries, n.d.).

The AGS classification was exclusive for maps, atlases, and reference materials. It used a three-digit numeric notation to represent geographic area and alphabetic notation for subject, followed by the date (University of Wisconsin Milwaukee Libraries, 2016a). The system's limitation was that it did not accommodate thematic maps (Romero & Romero, 1999). In both B&L and AGS, the date of situation was considered vital information. This date was not the date of publication or reprinting, but rather the date of the data represented. This was important because it qualified usefulness, which was likely related to the main users at the time (e.g., the Department of State). The call number begins with a three-digit number representing area or world regions. These systems progress from general to specific; brief examples of the classification are shown in Tables 10.1–10.3. In the tables, the B&G listing examples are from ANZMapsS (n.d.); the AGS listing examples are from University of Wisconsin Milwaukee Libraries (2016a).

10.4.2 Dewey Decimal Classification

The DDC was created by Melvil Dewey in 1873 and is a proprietary system first published in 1876 as a four-page pamphlet (OCLC, 2015). The latest edition is from 2011, revised and expanded through 23 major editions in a four-volume set (OCLC, 2016c). It has an abridged version for smaller libraries and is currently maintained by the Online Computer Library Center (OCLC). OCLC licenses access to an online version called WebDewey (OCLC, 2016e), which may be downloaded for a 30-day free trial (OCLC, 2016c).

In general, the DDC allows concepts of relative location and relative index for new materials added to libraries in the appropriate location. There are main classes by subject and fractional decimals beyond the three-digit Arabic numerals. For example, the 900 class is history and geography; maps could be classed in 911 for historic geography or 912 for graphic representations of specific subjects. Map types could be expanded in linear fashion. The DDC is the oldest and most widely used in the U.S. and many other countries (Taylor, 2004). Specifically, the DDC is used in 200,000 libraries and in at least 135 countries (OCLC, 2016c).

In spite of this being a popular classification system, Davis and Chervinko (1999) found fewer than 6% of map-cataloging libraries used DDC. In DDC, most cartographic materials are classified under 912 and added to this base number is the more specific subject. Romero and Romero (1999) remarked that the main drawback for map librarians was classifying subject first and making the geographic location a secondary aspect, given that most reference questions requested a map of a given geographic area. DDC also has a U.S. bias, and cartographic resources are global. Larsgaard (1998) called the DDC an inappropriate classification for maps and cartographic resources.

10.4.3 Universal Decimal Classification

In 1885, Paul Otlet and Henry LaFontaine were working on a classified index to published information. Otlet was aware of Melvil Dewey's work, and in 1895, Otlet gained permission to translate the DDC into French (UDC Consortium, 2016b). The DDC formed the basis for Otlet and Lafontaine's system, and an English language version was published in the 1930s. It was initially managed by the International Federation for Information and Documentation until 1992 when the UDC became affiliated with the UDC Consortium (UDC Consortium, 2016c).

The UDC is an indexing and information retrieval tool, made up of 10 classes, each divided into 10 divisions, each in turn having 10 sections. It uses Arabic number notation, three whole numbers representing the main classes, subclasses, and decimals for further divisions. The structure is hierarchical and 900 is the general class for history, maps, and geography. Unlike DDC, the UDC does not have a U.S. bias and cartographic materials may be classified first by area and then by subject; if deemed more important, materials are then classified by subject first (Romero & Romero, 1999).

The UDC is a system widely used by libraries and information services in more than 130 countries and translated into 50 languages (UDC Consortium, 2016a). The UDC Consortium is a nonprofit group, headquartered in The Hague, Netherlands, and made up of publishers with an editorial team and advisory board who maintain, develop, and distribute this classification system.

Larsgaard mentioned the importance of the UDC outside the United States, and while “the first one thousand classes (000/999) has been maintained, constant revision has produced increasingly serious deviation in details” (Larsgaard, 1998, p. 143). Again, maps are primarily in 912 and are designated by country, and parentheses enclosing place or country or place and form.

The UDC system is flexible in that if the subject of the map is more important than country, then the number of the place may appear at the end of the entry. This classification system is widely accepted outside the U.S., and fits the way clients conduct a search, which is often by geographic area first. For more detailed examples and explanation, the following references are recommended: UDC Consortium (n.d., 2016d) and Allington-Smith (2015, May 31).

10.4.4 Federal Depository Superintendent of Documents Classification

The Superintendent of Documents (SuDocs) system for library classification was developed in the office of the Superintendent of Documents of the U.S. Government Publishing Office (GPO) between 1895 and 1903 (Federal Depository Library Program, 2015). The Superintendent of Documents was tasked with storing, cataloging, indexing, and distributing government publications, but the person who devised the classification scheme to organize government publications was Adelaide R. Hasse. She worked in the Los Angeles Public Library in the 1890s, but Hasse moved to the GPO Public Documents Library from 1895 to 1897 (GPO Access, 2004). This library no longer exists.

What distinguishes this scheme from other library classification systems is a reliance on the origin of the document or provenance, rather than an arbitrary subject. Provenance has proved to be a flexible, expansive, and descriptive system for collections. The origin or authorship is not usually a personal author, but the agency, bureau, or office where the document was created. This alphanumeric scheme is arranged alphabetically by the leading letter of the agency that originated the document. This is followed by a number, period, whole number, and colon; the colon is a break between the SuDoc stem and its suffix, which consists of a sorting hierarchy including dates, letters, numbers, words (Federal Depository Library Program, 2015). The documentation for SuDocs cataloging was last printed in 1993 and is available for download online (Federal Depository Library Program, 1993). An example for a topographic map follows.

The SuDocs map number for Kittitas, Washington, is I 19.81:46120-H 4-TF-024/978, each element is explained later. This SuDocs classification example is from a map in the Federal Depository. SuDocs is called a provenance system because it organizes publications by issuing agency, which in this case is “I” for the Interior Department (U.S. Department of the Interior, n.d.). The “I 19.81” is the class stem and the “19” is the designation for the USGS (U.S. Geological Survey, n.d.), one of the agencies under the umbrella of the Interior Department. The “81” is the designation for 7.5-minute topographic series quadrangles. Following the colon is “46” and “120” or the coordinates in degrees latitude and longitude. The “H 4” is a map reference number based on the north and west coordinate directions (latitude and longitude) and the North American Datum of 1927. For information on datum, see Chapter 3 and MapTools (2016). “TF” represents the type of map, topographic, and “024” is the scale, 1:24,000. Finally, the last three numbers “978” represent the edition date 1978; with dates, always drop the first number in a date prior to the 21st century, and if the map is from 2000 and beyond, the record would end in four numbers.

Davis and Chervinko (1999) found that of the map-cataloging libraries 16% reported using SuDocs classification. However, this is a bit misleading because many libraries used several systems for cartographic resources. Most government documents would be classified using SuDocs, yet the library would classify other cartographic resources using LCC. Interestingly, many libraries created their own local system, greater than 20%, yet nearly 30% reported a local system based on the LCC (Davis & Chervinko, 1999). While these statistics are dated, it is clear that LCC is the preferred classification system for cartographic resources.

10.4.5 The U.S. Library of Congress Classification, Schedule G

This LCC scheme was devised by Herbert Putnam (Minneapolis Public Library, 1889). Putnam developed the system in 1897 at the Minneapolis Public Library and later became the 8th Librarian of Congress, serving from 1899 to 1939 (Library of Congress, n.d.c). The LCC was designed and developed specifically for the LOC collection, replacing Thomas Jefferson's fixed location system. When Putnam left the LOC in 1939, all the classes except K (Law) and B (Philosophy and Religion) were fully developed.

The LCC is used by most research and academic libraries in the U.S. and several other countries. The LCC system overall is organized according to 21 basic classes, which then follows a logical order based on a discipline's domain divisions with numbers that are assigned creating a detailed item call number (Library of Congress, 2014). The call number was used to locate or physically call for the resource during times of closed stacks in libraries, which may still exist today in the U.S. if the maps are outsourced to a storage-only location. LCC is a subject-oriented classification with specific numbers called cutter numbers, introduced by C. A. Cutter; they are a coded representation of the author, organization, map publisher, and the like.

Specifically, Davis and Chervinko (1999) report 83% of the map-cataloging libraries reported using the LCC. In a 2004 survey, Thiry and Cobb (2006) discovered this trend among unclassified to fully classified map collections and institutions that reported classification systems as well; for example, the University of Illinois at Chicago reported 99% of the maps were classified and the systems were LCC and SuDocs, whereas the University of Chicago had only 60% of the maps classified and it was using only LCC. Larsgaard (1998) affirmed this LCC endorsement when she stated, “Schedule G of the LC class system contains the best classification scheme for cartographic materials” (p. 120). The first edition of Schedule G was introduced in 1910, but this classification continued to develop and was completed for atlases in 1928 and maps by 1946. The basic atlas call number structure is area, subject, author cutter, and date of publication; map call number structure is area, subject, date of situation, and author cutter. Larsgaard noted this difference in the order of structures for atlases, and maps are also one of those unexplained anomalies. A brief, generalized introduction to LC call numbers follows and an easy to read general explanation is provided by University of Illinois at Urbana-Champaign (2015).

The LCC groups, divided by major classes of information, are signified by one or two letters that are not mnemonic. The geographic portion of the LCC happens to be indicated by a G—Geography, Anthropology, Recreation. Class G is divided into subclasses from G-GV; maps are primarily under G, geography atlases and maps; GA, mathematical geography and cartography; and GB, physical geography, and so on. Furthermore, the subclasses have subsets of those groups, which are numerals up to four digits. For specific examples, Schedule G atlases are classed at G1000.3-3122, globes G3160-3182, and maps G3190-9999. Beyond the four digits, alphanumeric codes follow for subjects after a decimal point. These are cutter numbers, and “each major cultural or political unit in the world or universe has been assigned a block of numbers” (Larsgaard, 1998, p. 123). The current version of geographic cutter numbers has over 100,000 categories in 2016. This file may be downloaded as a pdf, but it is 6.5 MB in size with more than 3000 pages.

Also, there is a more thematic classification accomplished by decimal and subject code system. It is alphanumeric from A to Z, except I, O, W, X, and Y; it is not mnemonic and letters are followed by numbers representing subtopics. Within the maps class, subject code categories include bird's-eye views, plans, cross sections, diagrams, remote-sensing images, relief models, digital maps, and more. For example, C is for Physical sciences and .C2 is physiography, .C22 is relief features, and .C225 is shaded relief. The A indicates special categories in maps and atlases. It should be noted though that these subject letter/number combinations are not cutters.

The entire classification is not reproduced here but is available for Class G, Tables G1–G16, and Geographic Cutter Numbers (Tables G1548–G9804), throughout the subclass G (Library of Congress, n.d.d). Online access is available for libraries by subscription to Classification Web (Library of Congress, n.d.a; Library of Congress, n.d.b).

In addition to classifying cartographic resources, libraries use descriptive standards to organize knowledge resources and enhance access and retrieval. This final section briefly summarizes and contrasts two cataloging content standards, which are still based on some form of Cutter's principles of organization used in classification. Also, for remote access, a structure framework was developed by the Library of Congress in the 1960s, known as MARC or MAchine-Readable Cataloging, which is still in use today (Library of Congress, 2016b). The historic progression in cataloging and its future follows.

10.5 Cataloging Cartographic Resources

10.5.1 Progress in Cataloging

At the end of the 19th century, librarianship was being formalized as a career. Librarians organized collections of materials on shelves and in storage cases within library facilities. To make these resources accessible, catalogs were created. Catalogs were essentially an inventory and listing of resources as well as providing locations for each resource. Making the catalog an effective retrieval tool meant identifying the most important access points in a bibliographic record, or today, using the relationship model in works, expression, manifestation, and item.

As noted at the beginning of the chapter, isolated cases for cataloging cartographic collections began in the 18th and 19th centuries. Geographic area and subject were the main entry or access points used to organize collections within catalogs. Unfortunately, the 1908 code book entitled, “Catalog Rules, Author and Title Entries,” worked effectively for books but did not extend these early cataloging lessons for maps (Hanson, 1908). The assumption was that including a description under the cartographer or publisher name would suffice for finding maps. However, recording the USGS as author on hundreds or even thousands of separate topographic map sheet entries and adding map titles such as World, Texas, or Blue Lake, Colorado would do little to help locate a specific map with the needed scale effectively. It took several decades to design a system that worked for both book and cartographic resource.

In 1947, the new code book, “Rules for Descriptive Cataloging in the Library of Congress” was widely accepted. There was a section devoted to maps, relief models, globes, and atlases, and two years later, a second edition had a new section on maps and atlases. Still, librarians were not adding cartographic collections to catalogs (Morsch, 1949).

In the second half of the 20th century, the first edition of the Anglo-American Cataloguing Rules (AACR) was issued in 1967. In the 1970s the AACR benefited greatly with the addition of MARC as the encoding standard. This moved the catalog into a digital format where records could be read by computers and easily shared among libraries. Cataloging was increasingly complex, more items were digital, and preserving metadata with the record was problematic. Descriptive and subject cataloging evolved along with classification systems such as the DDC and LCC.

Also in 1967, the Ohio College Library Center (OCLC) was founded and WorldCat was launched; the initial cataloging records were added in 1971 to the OCLC database, which was the first online cataloging done by any library (Bryant & Mason, 2016; OCLC, 2016a). The plan at that time was to merge Ohio library catalogs electronically with a computer network and database; the purpose was to increase library efficiency, better serve researchers, and lower complexity and cost.

Today, OCLC is a nonprofit computer library service and research organization still known by the same abbreviation, which now stands for Online Computer Library Center. WorldCat.org is a global library catalog, or a union catalog, that describes collections in many member libraries (OCLC, 2016a). Creating a crowdsourced catalog would not have been possible without forward thinking, a desire to create a cooperative regional and later global catalog for information and data, and digital encoding standards, namely MARC.

MARC is a digital format for describing bibliographic items developed in the 1960s to facilitate computerized cataloging from library to library in regional or international situations. In 1971, MARC format was the national standard for dissemination of bibliographic data, and by 1973 was also the international standard. Reitz (2004–2014b) defined the purpose of MARC standard format for libraries as a way to have predictable, reliable cataloging data and to act as a bridge between libraries and library automation systems; MARC assists libraries in sharing bibliographic resources, avoiding duplication of records, and ensuring bibliographic data is compatible when changing automation systems. The MARC record itself has three components: record structure, a content designation, and data content. The data content is defined by the external standards of AACR2, LC Subject Headings, and the like.

In the past, Cutter's principles of organization enabled patrons to find a book if author, title, or subject was known. The methods for doing this provided access points such as an author entry, title entry, subject headings, and cross references. Handwritten or typed cards were created and placed in a card catalog inside wooden cabinets. These cabinets and card catalogs were moving out of library reference areas, and by 1983 the content cataloging was updated and AACR2 adopted. Catalogers recognized that cartographic and monograph cataloging had much in common. A new field, MARC 255, was added along with other improvements for maps.

The MARC 21 family of standards was created in 1999 to herald the 21st century. It was a result of efforts to make the United States compatible with Canadian and European standards. MARC 21 has formats for five types of data including bibliographic data, holdings records, authority records, classification schedules, and community information. AACR2 continued to improve with some of the last revisions and updates in 2005 (AACR, 2006). It was at this time that many libraries were adding significant numbers of maps into catalogs.

A posting on the popular listserv, MAPS-L, documented the incredible amount of cartographic-materials records that have been added each year to OCLC from 2005 to 2015 (C. Winters, personal communication, July 17, 2016). Overall, these statistics were gathered for 18 of the biggest map library collections at private and public universities and agencies. One public library was represented along with 15 universities, the USGS, the LOC, and OCLC. Every group increased the number of records added to the OCLC catalog over the years; the one public library cataloged the fewest of all the yearly reports, at over 21,000 in 2005 and nearly 35,000 in 2015. Excluding OCLC and LC, one university had the highest number entered in 2005 at nearly 66,500, and a different university was the highest in 2015 at nearly 86,500 entries. In 2005, LOC and OCLC added nearly 243,000 and over 857,000, respectively; in 2015, they added over 312,500 and nearly 4,695,500. The pace of cartographic entries may slow as one librarian added that in his library “there are now only a few pockets of uncatalogued materials in the collection” (C. Winters, personal communication, July 17, 2016).

10.5.2 Tools of Cataloging

The Library of Congress (2016a) offers many Cataloger's Desktop services online for RDA, AACR2, and Web Dewey. In addition, many librarians benefit from participating in a shared catalog. Even though cataloging is “generally based on early 20th century information organization principles of Charles Ammi Cutter” and is an “aid for awareness and access to a local collection for librarian and client alike,” cataloging is “a common structural framework that bridges global collections” (B. Hanschu, personal communication, August, 2009). Cartographic resources vary in type, and there are two methods to add records to a catalog, original, or copy cataloging. Hanschu added sound advice from a cataloger's experience: when performing cataloging, verify everything, it is best to never assume anything, and never make anything up. The recommended tools include AACR2 manual, OCLC bibliographic formats and standards, OCLC Code list, and LC Free-Floating Subdivisions, and the Cartographic Materials (2nd ed.): A Manual of Interpretation for AACR2 (Mangan, 2003). Using AACR2 guidelines, Hanschu provided a quick tour of the process and procedure for copy cataloging maps with the map in hand, which can be seen in Appendix C.

10.5.3 Change is Inevitable

The longevity of AACR2 combined with advantages of sharing catalog records with OCLC and WorldCat.org have taken librarians into the 21st century. However, updates over the nearly 40 years of this content cataloging standard were needed and formal discussions began regarding change in 1997.

In the late 1990s it was becoming obvious that the World Wide Web was the primary means to connect library users to the library catalog. However, Coyle and Hillmann (2007) criticized the continued use of MARC, developed in the 1960s, as the “middleware between the cataloging function and library systems development.” Questions without easy answers were asked. Is the library's signature service, the catalog, proving to be an equal to Amazon and Google in the search for information as perceived by the public? Are the rules and instructions for cataloging meeting goals or just remnants of a long departed technology, the card catalog?

Coyle and Hillmann (2007) answered both questions above with no and argued that a simple “rearrangement of the cataloging rules is not the right starting point for libraries.” Coyle and Hillmann suggested the question in 2007 was not whether Amazon and Google had created a generation that no longer needed the library, it was how to change a mind-set from catalogs as inventory of the holdings in one library to recognizing information and data users' needs may include resources in libraries and nonlibrary communities.

The main disadvantage of records created using AACR2 rules was these records are not interoperable with other data records and metadata schema. “Crosswalks” and related tools must be applied to enable search engines to operate across databases with dissimilar record formats. “Other interoperable issues deal with various problems such as different records not having exact field-level matches or fields of importance in one standard not necessarily having a related field even similar to them in another standard” (Andrew et al., 2015, pp. 106–107).

As such, the U.S. Federal Geographic Data Committee's Content Standard for Digital Geospatial Metadata and other committees increasingly identified cataloging concerns such as the complexity and plethora of scanned-digital and native-digital resources being generated.

AACR began in 1967, and now the question was should there be AACR3 or something new. “New” was the choice and RDA was presented as the cataloging standard, designed to replace AACR2. It was published in 2010 and implemented into the current cataloging workflow by LC and others in 2013 and beyond.

Although RDA was launched several years ago, it appears MARC is still the middleware of choice for many institutions. New changes to MARC 21 were announced in August, an OCLC-MARC Update 2016 (OCLC, 2016d). The changes are in Bibliographic and Holdings formats; while Authority Format changes were discussed but not implemented until they can be coordinated with the “Library of Congress and the Name Authority Cooperative (NACO) of the Program for Cooperative Cataloging (PCC)” (J. Weitz, personal communication, August 23, 2016). The discussion and how it relates to cartographic resources continues.

Coyle and Hillmann (2007) took part in the discussions and opposed RDA. In an opinion article they summarized the historical perspective on talks regarding the future of AACR2. They argued that RDA was just more complex than any cataloging schema in the past without providing any fundamental improvements. Coyle and Hillmann suggested that adopting RDA would move libraries back into the 19th or 20th century, not forward into the 21st.

For three decades, multiple generations of catalogers have perfected and been comfortable with the AACR2 content cataloging standard. With numerous revisions, the system was well documented, used, and known, according to Andrew et al. (2015). They suggested the main advantages were that AACR2 focused on the resource in hand and its complete description as a bibliographic record. The AACR2 had easy-to-follow organization and principles, and the third chapter provided a one- or two-step process to speed up cataloging cartographic resources. Rules were grouped into eight different formats. This design made it easy to comprehend, and there were separate sections for rules such as covering access points.

Andrew et al. (2015) pointed out the main changes for cartographic catalogers using RDA. The arrangement is completely different, and the instructions are applied to a work, expression, manifestation, or item. The advantage was that with RDA focused on relationships, and the Functional Requirements for Bibliographic Records (FRBR) model could make cataloging relationships better able to accommodate the multiplicity of standards to preserve metadata. The disadvantages are that learning the theoretical foundations of RDA-FRBR and how inherent relationships are expressed would be a steep learning curve, given that it is described in a 1000-page document. Other disadvantages are a lack of format-specific documentation for guidance and some more subtle differences. RDA allows no Latin terms or abbreviations, yet there are some exceptions, which is problematic.

The debate will continue on the merits and pitfalls of any new cataloging system. Some libraries have adopted RDA, and others will continue using AACR2. Coyle and Hillmann had advocated for a “unified vision allowing us to harness our collective strength as we go forward” but whether or not unification happens, change is inevitable.