UDC Consortium About UDC Outline Master Reference File UDC News UDC Publications Bibliography Consortium Members UDC Users UDC Blog

UDC Consortium, PO Box 90407, 2509 LK The Hague, The Netherlands
Tel.: (+31) 70 314 0509    Fax: (+31) 70 314 0450 attn UDC    E-mail: mail[@t]udcc.orgClick on the link but please remove the
part of the email address

Notes & Queries

We welcome replies or comments from users. Please send a message to mail[@t]udcc.orgClick on the link but please remove the
part of the email address


More about the meaning of common auxiliaries  
Persons and roles  
Persons and animals  
Past civilization and cultures  
Relocation of Medicine - your views are sought  
UDC - users figures  
Why German libraries will NEVER use the UDC? (Response)
Inner and outer form: a note on citation order  
Truncation of dates - and what does the date refer to?  
Should the asterisk * be abolished? (Response)
Should slash / be abolished? (Response)
A biterminal symbol for non-UDC notation (Response)
East African states  
More about slash  
Place and institution are not coextensive (Response)
Future of class 65 (Response)
Punctuation and intercalation (Response)
(3) versus (4/9) — is the distinction needed? (Response)

More about the meaning of common auxiliaries

In this section last year (1), I considered the meanings that common auxiliaries can acquire from context1, and remarked that 'in UDC an auxiliary qualifies the number that it is juxtaposed with'. In the course of the Revision Group's work in cancelling and relocating parts of classes 06 and 65 that are rendered redundant by the new management table at 005, an example occurred that illustrates the point even more emphatically, showing that the same auxiliary table can acquire contextual meaning from different parts of a single built number. A place was needed for temporary exhibitions (now at 069.9), and examples were agreed on:
069.9(100) World fairs, expositions (expos)
069.9(100)"1851"(410.111) The Great Exhibition (London, 1851).
The second of these contains two place auxiliaries, and their meaning is different: the first expresses the scope or inclusiveness of the exhibition (world-wide), while the second expresses the location (London). And the order is logical; (100) qualifies the subject of the main number, and is juxtaposed with it. The other elements follow in conventional sequence - time and place, in the reverse of filing order. If there were an inflexible rule for citation order, it would separate similar entities:
069.9(100) World fairs in general
069.9"1851"(100)(410.111) An 1851 world fair in London -
first a sequence divided by place, then a sequence divided (primarily) by time. Unhelpful!

1 The meaning of common auxiliaries. In: Notes & Queries. Extensions and Corrections to the UDC, 23, November 2001, page 10.


Persons and roles

A query on how to classify the concept 'child pornography' has drawn attention to a gap in the expression of roles in the common auxiliaries of persons (Table 1k-05). Certainly the concept, and term, are likely to be sought, and so notation has been agreed on - 343.542.1-053.2. This should be adequate, though it is unspecific as to role, meaning simply 'pornography in connection with children'. There are two role indicators available in -05… :

-051 Persons as agents, doers, practitioners (studying, making, serving etc.)
616-051 Medical staff: doctors, nurses etc.
929-051 Biographers

-052 Persons as targets, clients, users (studied, served etc.)
616-052 Medical patients
929-052 Biographees

The impression given is that the terms listed (e.g. 'targets, clients, users') are strings of synonyms, though synonymy is never an exact match. Roughly, -051 = active; -052 = passive. If these indicators were added to the number cited above, we would have

343.542.1-053.2-051 Pornography - children - active (As creators? As organizers?)
343.542.1-053.2-052 Pornography - children - passive (As targets? As recipients or consumers?)

both of which seem absurdly inept. There is no way in -05… of indicating persons as subjects, and only an unsatisfactory way of indicating victims, by equating them with targets - but targeting could be beneficial. Clearly, roles in human activities are more various and more subtle than is allowed for in UDC. There is no vacant 3-digit notation available, but it might be useful to subdivide -051/-052; for instance, active roles [-051] include makers or creators as well as organizers or facilitators, while passive roles [-052] include beneficiaries, victims and consumers. The matter is complicated by the fact that the nature of the role is already implied in some subjects, e.g. the recipient of cash benefits can only be a beneficiary, while the recipient of assault or murder can only be a victim, so to specify the role by still more detail would be a tautology. Yet there are cases, as we have seen, where ambiguity can occur.


Persons and animals

Among the many human attributes listed in the -05 auxiliaries (Table 1k) are a few which are equally applicable to non-human animals, the obvious ones being:

-053.1 … Born, unborn, living or dead (embryos etc.)
-053.2 Young…
-053.8 Adult…
-055.1 Male…
-055.2 Female…
-055.5/.7 …Kinship relation [sire, dam etc.]

In classes 59 and 63 in the MRF, the concepts 'male animals' and 'female animals' do not occur at all, though related concepts do, and are denoted either enumeratively, as at 591.35 'Young animals', or by special auxiliaries, such as the subdivisions of 636.082.31 'Structure of herds and flocks'. However, the old 'full edition' of 636 does include a single instance of the -05 auxiliaries:

636.06 Characteristics, constitution and special biological features of domestic animals
(Example) 636.06-053 Domestic animals according to age etc.

There seems to be no reason why the relevant numbers from Table 1k should not be used for animal characteristics, while the irrelevant ones would of course be ignored. There cannot be a simple substitution for the existing notation, because it would alter the filing order and disrupt the structure of the classes, but the option could be kept in mind for future revisions. A simple addition to the scope note at -05 would suffice: 'Where relevant, the subdivisions of -05 may also be applied to non-human animals'.

Past civilizations and cultures

A user asked how to denote the history of the Incas in UDC. At the time, the term did not occur in the MRF except as an example under the rather emotive heading 'Primitive art' at 7.031.2, but the common auxiliaries of race, ethnic grouping and nationality (Table 1f) allow you to derive numbers from the language auxiliaries (Table 1c) for linguistic-cultural groupings. However, you can only do this if you know what was the language of the people concerned. The two best-known pre-Columbian American nations, the Aztecs and the Incas, are contrasted in this respect: the language of the Aztecs is actually called Aztec (though the native name Nahuatl also appears), and can easily be traced. Inca, however, has never been the name of a language, and you need to know that the language of the Incas was Quechua; then you can trace the language auxiliary =873.122 and hence the ethnic auxiliary (=873.122). It seemed worthwhile to add the two examples, Aztecs and Incas, under 94.

Another problem arises from this. The common auxiliaries of place (Table 1e) include a number for pre-Columbian America, (399.7), but no further subdivision is given, so specification of particular civilizations can only be done by ethnic auxiliaries. This raises the possibility of cross-classification. It would be a very careless classifier who used (399.7) for pre-Columbian peoples in general and (=8...) for various specific pre-Columbian peoples. But it would be an easier mistake to make, to class some cultures, e.g. ancient Greek and Roman, in (3...) and others, especially non-European, in (=1/=8). To some extent, UDC invites this, and it shows its Eurocentric origins in (3...), which is based mainly on places known to Graeco-Roman antiquity, with a few other places offered as an afterthought. The resultant mixture would be very unhelpful:

94(3) History of ancient cultures
94(37) History of ancient Rome
94(38) History of ancient Greece
94(4/9) History of all the countries of the modern world
94(=...) History of other ancient cultures
94(=822.1) History of the Aztecs
94(=873.122) History of the Incas

The separation of similar entities is bad enough in history alone, but is worse when searching only on the auxiliary, seeking material on all subjects related to a given culture (pre-Columbian science, arts, folklore etc.). There are three ways to avoid this:

[1] Choose either place or ethnic auxiliaries for all pre-modern cultures - never mix them in a given system; however, this involves a loss of precision, since some detailed concepts occur only in one or other of the tables.

[2] Avoid (3...) altogether. Derive all auxiliaries for pre-modern cultures from the language auxiliaries, e.g.
=124'02 Classical Latin; hence (=124'02) Ancient Romans
=14'02 Classical Greek; hence (=14'02) Ancient Greeks

[3] Never use ethnic auxiliaries on their own for pre-modern cultures - cite them after (3...), so that the general immediately precedes the particular:
94(399.7) History of pre-Columbian cultures
94(399.7=822.1) History of the Aztecs

The last option seems to be the best; but all of these possibilities are available within the rules of UDC, which does lend a little weight to the criticism sometimes made, that UDC allows too many options.



Relocation of Medicine - your views are sought

The draft restructuring of Medicine (class 61) continues in E&C, and the question has arisen in the Revision Group as to whether it should remain in its present location when the proposal is finalized. The long-standing gap at class 4 in UDC is an obvious candidate as host location, solving several problems at once: providing space for the redeveloped medical classification, leaving class 6 purely for technology, and filling that annoying hole in UDC, which has been there for nearly four decades.

The views of users and editors on the pros and cons of relocation would be welcome. Please email the Editor in Chief at:
i.mcilwaine@ucl.ac.uk .


UDC - user figures

The problem
We are sometimes asked for figures of the total use of UDC.
There are no authoritative up-to-date figures for worldwide use of UDC, because it is published in many different languages by different organizations, so even the sales figures are scattered.

The world
The most recent research currently known to us (as at June 2002) is from Estonia, by Sirje Nilbe, dated 1997 and entitled The worldwide use of the Universal Decimal Classification. A summary can be seen at
Its conclusion at that time was:
UDC has 100 000 institutional and individual users in 60 countries…
The tables at abridged, medium or full level are published in 25 languages.

The UK
The last census of users in the UK was done in 1980, so the figures are very dated, but it was an active survey (not a postal questionnaire) so they are exhaustive. There were 640 users in the UK at that time, making up 22% of libraries and information agencies in the country.
Around 2500 copies of the standard English version BS 1000M have been sold since 1993, though that includes sales abroad, and the picture is complex, as a large organization may buy more than one copy.


Should the asterisk * be abolished?

The asterisk * is used in UDC as an indicator of non-UDC notation, but is commonly used in computer searches as a wild card. This dual role could cause confusion, and the asterisk is not really essential in UDC notation; so should it be abolished? Where alpha characters occur in a class mark, they are themselves a facet indicator, and no other symbol is needed. For example,

66–97*C100     100° Celsius (boiling point of water)
could easily become:
66–97C100     100° Celsius...

Where alpha characters do not occur, a symbol is needed to replace the asterisk, and there is an obvious candidate — the hash #. It is already commonly used (especially in north American usage) to mean ‘number’, and would be easily and intuitively understood, e.g. as the atomic number of an element:

546.791.027*238 Uranium 238
could become:

546.791.027#238 Uranium 238

Return to top of page

Should the asterisk * be abolished? Responses to Notes & Queries 1999, from D Strachan

I take these two items together, as they are both about Table 1h signs.


The hash sign (#) has two special uses in CDS/ISIS:

  • It is used as the default field tag (single hash) and record tag (double hash) when exporting or importing records using the ISO 2709 interchange format.
  • In the CDS/ISIS search language it is used to specify a search by number in refining a previous search or in combining previous searches.

Some tests with the hash sign suggest that the CDS/ISIS system recognises the difference between the hash as part of a text string and as a tag, and the example at -023.3 (in E&C 21 and the 1999 MRF), which seems to have jumped the gun by using the hash, appears to have avoided any problems. (Note that there is a data entry error at this example in the 1999 MRF: -023.3 is wrongly entered as -023).

The note about a biterminal symbol, if accepted, supersedes everything in the note suggesting the hash, for the reason given in the first sentence. However:

  • The pointed brackets are used in CDS/ISIS for 'Indexing Technique 2'. In the MRF database this technique is used at present in four different 'notes' fields or subfields to mark those UDC numbers within notes which should be in the inverted file. This means that < and > cannot be used as ordinary characters in these fields unless or until a different solution for including these numbers in the inverted file is agreed upon and implemented. Implementation could be done fairly simply by a specially written program.
  • As the note mentions, the square brackets already have a subgrouping function, which would conflict with their suggested use for isolating non-UDC notations. It would be bad to have two different uses of the square brackets - one enclosing UDC notations and the other enclosing non-UDC notations - especially in computer-based systems.
  • The main problem with the brace brackets is that on-screen they can be difficult to distinguish from ordinary brackets.


1. In my view these two notes are trying to find hard-and-fast solutions to chimerical problems. Table 1h should be made less prescriptive, not more. I suggest that an introductory note should be included in Table 1h along the following lines and nothing else need be changed.

In some applications using the UDC, it may be useful to include non-UDC elements (codes, names, abbreviations etc.) in the notational representation of the subject. The notes which follow describe the long-established recommended method that is used in examples throughout the UDC tables. It is recognised, however, that this may not suit all applications, including those using software packages in which the asterisk is a control character. In such cases more appropriate methods must be devised by the user.

2. If the UDC Consortium agrees that a new sign can be added to the UDC's notational character set, then:

  • This is a very important decision and should only be made after widespread consultation.
  • The sign should be used in the most effective way for improving the UDC. To use it for tagging non-UDC codes would be to squander it on what is probably the least important part of all UDC notation.
  • In my view the top priority for use of a new symbol is to replace .0 as a facet indicator for special auxiliaries.

Return to top of page

A biterminal symbol for non-UDC notation

There is no reason why non-UDC notation should not occur in the middle of a built number. If it consists of alpha characters, its status is self-evident, e.g.

16WIT:81     Work on logic by Wittgenstein, in relation to linguistics.

But if the introduced notation is numerical, how can one tell where the non-UDC part ends and the UDC part resumes? For example, in

66–97C100–936.35     Reactions between liquids at 100° Celsius,
how much of the class mark is UDC? It might be better to enclose the introduced notation between biterminal signs. As quotation marks “...” and parentheses (...) are reserved for different meanings, the possibilities are square brackets [...], already used as a grouping device, pointed brackets <...> and brace brackets {...}. For example,

66–97[C100]–936.35     Reactions between liquids at 100° Celsius.


See D, Strachan response

Return to top of page


Place and institution are not coextensive

The current schedule for education in the MRF contains this entry:

378.4         Universities
Class here traditional universities with all or most of the usual faculties

378.4(410)OU     The Open University
378.4(410.117)     University of Oxford
378.4(410.126)     University of Sussex
378.4(443.611)     Université de Paris à la Sorbonne
implying that the place auxiliary for a county or a town will serve to specify a particular university. This is misleading. Firstly, one town may contain more than one university. Oxford (410.117) has three: the University of Oxford, Oxford Brookes University and Westminster College Oxford (with university status). Secondly, a university may not be limited to one place even if it is named after it. The University of London has colleges at Egham, Surrey (Royal Holloway), Wye, Kent (Wye College) and elsewhere. The place and the institution are not coextensive. Alphabetic extension is needed to specify an institution, even if the place auxiliary is also used:
378.4(410.117)OXF     University of Oxford.

The place auxiliary could be kept at a more general level if preferred, e.g. a country:

378.4(410)OXF     Universities in the UK — University of Oxford.

The alphabetic extension is outside the parentheses as is not part of the place facet.

For a map showing all recognised Universities, University Colleges and Higher Education Colleges in the United Kingdom, see


Return to top of page

Place and institution are not coextensive - Response from D. Strachan

There is no significant change to the UDC proposed in this 'Note' in E&C 21 - only the updating or removal of a few old examples at one entry to reflect current preferred practice in applying Table 1h within the MRF. It can also be said that four examples is at least two too many, especially when three are of the same country.

The choice mentioned in the note either to keep the place auxiliaries at country level or to use more detailed subdivisions is one which would normally be taken for the system as a whole, and need not be spelt out at individual notations.

The wisdom in keeping examples of non-UDC extensions to a minimum in the MRF is shown at 378.635.5. The example (443.62St. Cyr) is given for the institution 'Military Academy of Saint Cyr'. In this case changing it to (443.62)St. Cyr would not be enough, as the Academy has long since relocated from Saint Cyr in Yvelines to Brittany but has kept the same name.

As with the first two 'notes and queries' in E&C 21 (treated above), the MRF should provide the possibilities for non-UDC extension at Table 1h, and leave the application to the user. It should of course be applied in a consistent way throughout the MRF itself, wherever examples utilising non-UDC notation appear. The principle behind the 'Note' is sound - that alphabetical extension of the place facet should be inside the bracket, but names of institutions outside.

David Strachan, Dublin


Return to top of page

Future of class 65

The present content of class 65 is an odd mixture, particularly as management is split between 65/651 and 658, with unrelated material in between. Some aspects of management are also classed under 06. As management, like computer science, is all-pervasive (any activity may be managed), should management be entirely relocated in class 0? Of the remaining parts of 65, accounting, publicity and PR are also partly covered under 06 and could be developed there. Telecommunication, postal and transport services could move to 339. Printing could be more logically placed with printing technology at 681.6.

Future of class 65 -Response from D. Strachan

The ideas in this note have been around for several decades. There has been general agreement that organization/management should be in class 0. The most appropriate location is probably 005, beside Computing, rather than 06. The problem is that it would be futile to relocate the existing 65 'Management' breakdown to a different place without a radical overhaul.

If Management (Business studies, Organisation) could be redeveloped in 00 as 'all-pervasive', it would be worth considering devoting 65 to important service industries, preceding the production industries at 66/69. It could include the already present Accountancy, Advertising/PR, Transport services, Postal & Telecommunications services, Publishing.

It is sensible to merge 655 (printing) with 681.6 (printing machines), but it would not be simple. Throughout most of 67/68 manufacturing industries are developed consistently with the end products as main numbers, and machinery, tools etc. as .05... special auxiliaries. A new table for the printing industry should follow that pattern. The present 681.6 enumerates many complex concepts (special machines for printing particular kinds of product) which could be expressed syntactically if the machines were at .05....

David Strachan, Dublin

Return to top of page

Should the slash / be abolished?

The oblique stroke (or forward slash) / is merely a device for saving space. Is it still needed? It replaces a series of pluses +, thereby making a class mark shorter at the cost of obscuring information. Even the official note on its use (Table 1a, Section 2 ‘Consecutive extension’) advises against it, recommending separate class entries instead, i.e. not:

546.32/.35     The more important alkali metals
546.32     Potassium
546.33     Sodium
546.34     Lithium
546.35     Rubidium

In information retrieval systems, this is vital, as items classed at the invisible numbers in the middle of a range would be missed; class marks with the oblique stroke are adequate only for linear sequences such as shelf order. And there is an overlap of function: where the component numbers in a range are non-consecutive, the plus is obligatory, but when they are consecutive, the stroke may be used. Thus, when there are only two consecutive numbers, the meaning of the two signs is identical: (7/8) ‘The Americas’ means the same as (7+8). When there are more than two numbers in the range, the extra length caused by using pluses is unimportant in the era of computer files, copying, cutting and pasting.

Should the slash / be abolished? - Response from D. Strachan

The answer to this query must be 'no'. First, the opening premise of its argument, that the slash is "merely a device for saving space" cannot be reconciled with the description of the slash in Table 1a. Next, the query's quoting of Table 1a is selective, e.g.

"Even the official note on its use (Table 1a, Section 2 'Consecutive extension') advises against it, recommending separate class entries instead."

which ignores the important qualifying clause in the note:

"in cases where there is a need to retrieve information from all the component numbers"

There are two different uses of the slash. One is where the range is a class number appearing as an entry in the UDC tables, usually denoting a simple class at a level directly above the level of the components of the range. This type of usage is most often found at the high levels of the classification scheme (broad subject areas), as that is where hierarchical divisions are less clearly perceived, and there is a need to distribute the short class numbers efficiently e.g.

23/28 Christianity
343.2/.7 Criminal law
355/359 Military affairs
592/599 Systematic zoology
623.8/.9 Naval engineering
625.1/.5 Railway engineering, and
625.7/.8 Highway engineering
745/749 Applied arts

It would be interesting to see what the notations for these simple subjects would be if the slash were "abolished".

The other use of the slash is similar to the use of the plus sign. If a document treats more than one subject separately, and the subjects are consecutive in an array, the user is presented with two possible approaches and his choice will probably depend on whether the system is used mainly (a) for linear ordering (shelving, filing, browsing) or (b) for search and retrieval. For example a work containing sections on the physics of heat, light and sound might be classed as (a) 534/536 or (b) 534 and 535 and 536.

The choice between these options should remain with the user.


Return to top of page

Punctuation and intercalation

Intercalation of auxiliary notation in main numbers is a well-established UDC technique. Some sources illustrate it while omitting some of the punctuation, e.g. in the old Abridged English Edition (BS 1000A), last updated in 1963, is the example (page 9):

338(42)5     Prices in England
where the number is built up from 338... ‘Economic situation’, (42) ‘England’ and .5 ‘Prices’, but the point before the 5 has been omitted. This is possibly because the intercalated auxiliary is regarded as punctuation enough, but this is misleading. When a punctuation mark is omitted, how can you tell what kind of punctuation it would have been? The number might be interpreted as a colon combination, 338(42):5 ‘Economic situation in England, in relation to the sciences’. More recent English editions avoid this ambiguity, as does Prof. McIlwaine's Guide to the use of UDC (FID 703, 1993), which gives the example (page 34):
622(410).33     ‘Mining — Great Britain — coal,

not 622(410)33, which might mean ‘Mining in Britain in relation to economics’.

Punctuation and intercalation - Response from D. Strachan

The argument in this 'Note' in E&C 21 assumes that the point and the colon are the same kind of notational element - called "punctuation".

But the colon and the other UDC signs (plus, oblique, hyphen etc.) are syntactical devices. The point, on the other hand (except in a few special usages dealt with below), is only a visual aid to break up long strings of digits - in particular any string of four or more digits - making the notation easier for humans to read and to transcribe without error. It brings nothing meaningful to the notation, either as a linking device (like the colon) or a facet indicator (like the hyphen), and has no sorting value.

Thus to say that the example 338(42)5 "omits some of the punctuation" is to misunderstand the function of the point. As there is no need for a visual separator, nothing is omitted, and there is no basis for the suggestion that perhaps there is "ambiguity", i.e. a missing colon or other UDC sign.

To demonstrate this, one need only look at the recently revised special auxiliaries for Class 7 (E&C 18, 1996), where there are numerous examples of intercalation of a place auxiliary, followed directly by a digit, e.g. 7.036(492)81 and 7.038(410.1)54.

Furthermore, in both the examples contained in the 'Note' the intercalation is after the third digit, where if there were no intercalation a point as a visual separator would appear. But how would the "ambiguity" thesis deal with intercalations after the second, or fourth, or fifth digit? Suppose the example were not

622(410)33 Mining - Great Britain - coal


63(410)6 Agriculture - Great Britain - animal husbandry

or (from E&C 18)

7.033.4(430)1 Romanesque style - Germany - Merovingian.

David Strachan, Dublin


Return to top of page

(3) versus (4/9) — is the distinction needed?

It is well known that the distinction in the place auxiliaries (Table 1e) between the ‘ancient’ world (i.e. Graeco-Roman) and the ‘mediaeval-to-modern’ one (i.e. since the end of the western Roman empire, late 5th century AD) is Eurocentric and quite meaningless for many ancient nations such as India and China, and of course the civilizations of pre-Columbian America. Some UDC editions offer the user the option of ignoring (3) and using the appropriate divisions of (4/9) instead, and dealing analogously with the history and geography sections when these were divided enumeratively — before the analytical principle was restored in EC 19 (1997). But another question arises: do the (3) auxiliaries even serve their Eurocentric purpose satisfactorily? The current schedule for history allows for peoples not confined to one geographical area, using the auxiliaries of race and ethnic grouping (Table 1f), for example,

94(=214.58)     History of the Romany peoples...

Many of the peoples mentioned under (3) were similarly mobile, e.g. the Germanic tribes at the subdivisions of (363) and at (368). Should the Angles, Saxons, Jutes and Nordic peoples be thought of as inhabiting continental Europe or parts of the British Isles? If the former, which part of Europe? Normans were Scandinavians who migrated to northern France, and from there some of them colonized England; where does one class a complete history of the Normans? In fact, the very concept ‘the ancient world’, like the auxiliaries that express it, offends against logic by confusing two incompatible principles of division — it is partly spatial, partly chronological. Cultures not limited in time or space would be better denoted by ethnic subdivisions. For the history of a given area, the place and time facets should be clearly distinguished.


(3) versus (4/9) — is the distinction needed? - Response from D. Strachan

This query, although it seems to be in favour of cancelling (3) and its subdivisions, gives examples which lead one to the opposite view - that the cancellation of (3) would not solve the difficulties of denoting certain kinds of historical entity. The example given most space in the query is the Normans, and it is notable that, while asking "where does one class a complete history of the Normans?" the author of the query offers no answer. Nor is the use of Table 1f (race and ethnic grouping) much help, as it is based on language, and the Normans, for example, did not have their own language - they arrived speaking Norse and after a century or two were speaking French. And if the Normans are seen as a problem, what about the Vandals - a Germanic tribe which made important contributions to the history of Spain and then of North Africa?

The fundamental problem is that (4/9) is divided according to present-day, or at least 20th century, political divisions, and the further back in time one goes the less do these correspond to the historical political divisions. The notational change in E&C 19 (1997) from 94/99 to 94(4/9) has not affected this problem, as both the old and the new notations have the same content on which to draw.

Like the author of the query I have no solution to propose, only a plea that things should not be changed unless it is first demonstrated that the new is an improvement on the old. Probably the best way of denoting the migratory tribes of the old world (for those with sufficient literary warrant to be in the MRF) is by enumeration.

Textual notes:

The 'query' refers to "the distinction between the 'ancient' world and the 'mediaeval-to-modern' one" as being 'Eurocentric'. The MRF entry for (3) does have a note about the late 5th century, but this note is in conflict with the subdivisions of (3) in two ways. Firstly, many of the entries have their own terminal dates which are not late 5th century - e.g. China (ca 600 AD), Egypt (640 AD), India (647 AD), Greece (323 AD). Secondly, some of the (3) subdivisions flourished centuries after the end of the 5th century.

So it is interesting to see that

a) in the Cumulated UDC Supplement 1965 - 1975 there is a modified entry at 931 "Ancient history in general. History of ancient peoples. (deleting limiting date 476)". Unfortunately this was not also shown at (31).

b) while the Normans appeared in many editions at (368), including the English Medium Edition of 1985 (whence it reached the MRF) the text of the 1988 English Full Edition has removed them, I suspect because the original had been checked.

c) In E&C 20, (1998) the text of (3) has been changed to 'Places of the ancient and mediaeval world' without any explanation and without any corresponding change at other records where (3) appears.

David Strachan, Dublin


Return to top of page

East African states

A user of the Pocket Edition of UDC (PD 1000) asked why there was a place auxiliary for West African states, (66), but none for East African states. There is a mixture of reasons. In the MRF, there was an anomaly in the notation as well as some historical factors that influenced the selection. There used to be a number for East African states, (676), but it was a 3-digit number, unlike (66), and of course it appeared to be a subdivision of (67) 'Equatorial and central African states and territories' - not a very logical arrangement. Moreover, (676) was used as the generic placing for British East Africa, an ambiguous and now obsolete name for the former British colonies of Kenya, Uganda, Tanganyika and Zanzibar, but also sometimes used to mean Kenya alone. That entry has now vanished from the MRF, though I cannot at the moment trace when it was cancelled. It is not much needed, as Table 1e includes special auxiliary notations for the concepts 'East' (1-11), 'Colonies' (1-52) and 'Historical status' (1-89), so any of the individual countries involved could be indexed in great detail; and the generic grouping 'East African states' can be classed either at (6-11), or - if that is too distant from the individual states - at (67-11), which would imply the eastern part of the group of states listed under (67). There is no free 2-digit number that could be assigned to the east African countries, but the description at (67) should perhaps be enlarged to reflect its true content: 'Equatorial, central and eastern African states and territories'.

Return to top of page

More about the slash /

The above item brings to mind an oddity in UDC: though I said that UDC numbers can be truncated from the right, NOT the left, there is one exception, and that is the second element in a range number, i.e. the one following the slash. If it is more than 3 digits long, it can be abbreviated by omitting digits (in groups of three) that it has in common with the first element, resulting in a point (or other sign) as the first thing after the slash. The example in the MRF (Table 1a, Section 2),

546.32/.35 Heavier-than-air craft. Aeroplanes

conceals not only the numbers in the middle, 546.33 and 546.34, but also maims the final number, 546.35 - another reason for disliking it. It is fair to add, however, that Dr Riesthuis has shown (in his article 'Searching with words' [E&C 21, November 1999]) that an algorithm for deconstructing range numbers can be devised - but it is 'rather complex'. He also mentions that range numbers are of two sorts, 'those that are listed in the tables and those that are formed by the classifiers', and adds:

'It is necessary to look up each contraction to determine whether it is listed in the tables before splitting it up.' He gives an example of a range number that would not be correctly deconstructed, observes that 'It is almost always possible to express the same concept by simpler notation',

and concludes that

'Classifiers should be economical in their use of range numbers formed with the slash'.

This is a convincing argument; but it would be interesting to know whether UDC users out there make much use of the slash in user-built numbers.

Return to top of page

Why German libraries will never use UDC?

This is an unofficial explanation by German libraries, as to why the UDC will never be used by them. In course of time, the Austrian and Swiss libraries will also prefer to use the DDC. The source of the information was some librarians whom I met at a meeting in Germany.

Arguments against UDC according to the German libraries:

1) It took the German libraries several months and repeated tries to get any answer from the UDC Consortium.

2) No regular updates of the notation are available. Therefore, some libraries (e.g. ETH-Zürich) had to draw up their own updates. That damages consistency between different institutions.

3) There is no consistency in terms of time. Previous UDC versions and updates are not comparable, and (to take a fictitious example) a notation that once represented 'inorganic chemistry' might, twenty years later, represent 'life sciences'. This inconsistency over time is unacceptable for librarians (who count in centuries). Maybe this method is acceptable for a documentalist, but not for a librarian. This criticism has also been made by some UDC users in the past.

4) Licences are too expensive.

5) The data format is not clear and transparent to librarians. Sorry for that bad news. But just take a look at the DDC website and their transparent and easy software tool:

UDC is not yet dead, but if major changes do not take place, it will definitely be dead in the long run. In Austria, we still are in favour of the UDC, but if the above-mentioned problems are not remedied within the next few years, we may have to change our mind.

Libraries are some of the most influential agents in the information industry. Their decisions have a spill-over effect on other institutions and on the industry (publishers, software companies). In that particular field, UDC has lost the battle. I would be pleased if you could respond briefly to those five arguments.

Gerhard K. Wagner,
Secretary General Verband für Informationswirtschaft in Österreich - VIW
(Federation for the Information Industry in Austria) gkwagner@via.at

Why German libraries will never use UDC? - Response from the Director of the UDC Consortium C. Apers

Re 1): I would be very pleased to learn which German and Swiss libraries are dissatisfied with any aspect of UDC business. The two German and Swiss libraries that I have been in contact with recently now have licence agreements with the Consortium and I am not aware of any other issues they wish to raise.

Re 2): Your observation is correct where it concerns UDC in the German language. The Consortium has not been able, so far, to find a successor to DIN as a principal publisher of German UDC, although a few alternatives have been considered. We recognize that Germany is a country where there is a wide variety of classification schemes in use. At present, our French-language publisher (in Belgium) is preparing under licence a German abridged UDC edition based on the successful French abridged edition. This will be an authoritative, substantially up to date edition, but, obviously, it will not be as extensive as the "old" medium edition.

The core version of the UDC is embodied in the Master Reference File and is maintained in the English language. It is constantly updated with priority given to keeping track of new developments. All publisher members of the Consortium and licence holders receive copies of the new editions of the MRF issued annually so that they can maintain their products. In view of the fact that UDC covers all subjects, it is almost inevitable that some backlog exists but I daresay that the other classification schemes experience the same problem.

As to consistency, it is not uncommon for UDC (and other classification) users to adapt their copy of the classification scheme to suit their own, specific needs. We regard that as one of the strong points of UDC. We do not set out, and are not able, to ensure that all such adaptations are compatible, nor is the Consortium able act as 'policeman' over such variations. We can help, if asked, to ensure that the fundamental principles of the UDC are followed, so that proper use is being made of the well-established and well?proven protocols and practice. Where any major differences between two editions are brought to our attention, we will contact any two such parties to see whether a common solution can be found or, at least, to understand why differences occur.

Re 3): You have pointed out contradictory interests here. On the one hand, advances in knowledge call for major revisions of any classification or retrieval system; on the other hand specifically some librarians complain of disruption. The UDC Editorial team (as with any other scheme) is inevitably always in the middle. But to our knowledge, there has never been a change of meaning as drastic as the example given, except for one case: the cancellation of class 4 in 1963 and the possible re-use of it for medicine. Otherwise, I am afraid that we do not agree with you that the successive versions would not be comparable. In the MRF we do track all changes and rationale for change. It is possible to give an audit trail for every entry that has been changed since the MRF was first created in 1992.

Re 4): We cannot agree. Even some of our own Executive Members have been pressing for licence rates to be increased to become more in line with commercial rates. Our policy is to set rates that are common to all applicants, that are reasonable, and that allow recovery of administrative and production costs.

Re 5): The data format is a major concern to the Consortium and we agree with you that ISIS is not a user-friendly package, however the formats (ASCII and ISO 2709) in which the UDC can be output from the MRF via ISIS are compatible with many applications, and our many publishers seem to be able to produce their various editions from the ISIS files without too much difficulty. One Executive Member, the British Standards Institution, is at the moment developing an online version of English UDC that will be very searchable and user-friendly (to be updated annually in line with the MRF). It should be available in 2001. The Consortium has been offered the opportunity to collaborate in a subsequent multilingual version of this product, which would have a direct and positive spin-off on the products of the other Consortium members and licensees as well. We are confident, however, that there always will be UDC users, particularly in Eastern Europe and Africa, who will continue to use ISIS, since it is a very powerful package, it's free, and I have found that the librarians in the geographical areas I mentioned are quite fluent in ISIS. They will not be disenfrachised by any online UDC, as there are no current plans to alter the way the Master Reference File is maintained via ISIS. Over the last two years, the Consortium has undergone several drastic changes within its organization, involving new people, computer systems, offices etc., and the prospect of new services (such as online, mentioned above). We are still in a transitional period, and would welcome discussion and cooperation to improve the lot of users of German UDC. It is clear from your message how concerned you are about the state of UDC in the German language, and I can assure you that such concern is appreciated by the Consortium.

Caren Apers
Director UDCC (resigned 2004)



Return to top of page

Inner and outer form: a note on citation order

Form characteristics of a document are of two kinds: inner form, where the form influences the subject (e.g. historical presentation) and outer form, which expresses only the physical characteristics of the information carrier (e.g. a sound recording).

In UDC, Table 1d (the Common auxiliaries of form) does not distinguish between the two kinds, and they are intermingled in the table. This means that when several form auxiliaries are cited, ascending numerical order may not be appropriate; inner form should occur next to the subject before outer form is expressed,

A spoken-word history of the theatre
(subject - historical form - sound recording)

A note to this effect should be added at the beginning of Table 1d.

I.C. McIlwaine
Editor in Chief


Return to top of page

Truncation of dates - and what does the date refer to?

A user has inquired how calendar dates are truncated in UDC ('Would the current century be "20" or "00"?'). The answer is that calendar dates, though quoted directly from an external source, behave in this respect like all UDC numbers: they can be progressively truncated from the right - thus denoting more and more general concepts - but NOT from the left. So "2000" is a single year; "200" is a decade (= "2000/2009"); "20" is a century (= "2000/2099"); and "2" is a milennium (= "2000/2999"). "00" would mean the first century AD (= "0001/0099"). Incidentally, if one pedantically observes the Gregorian calendar, which does not recognise a year 0, the Christian era begins with AD 1, i.e. "0001", and all centuries begin with a number ending in 1. So the 21st century, strictly speaking, is "2001/2100"; but this is a fine distinction that is not needed for most purposes. "20" is more economical.

The same inquirer asks about the exact significance of a time auxiliary in classifying literature: 'Does the time indication pertain to the century in which the author was born, or to the century in which his/her works were published?' The answer is that any auxiliary qualifies the number to which it is attached, so if literature is the subject being classed, the time auxiliary indicates the date of origin of the literature. Very often, the date of first publication follows soon after completion of composition, but not invariably: a previously unpublished work may receive its first publication centuries after it was written. Where there is any significant time difference, the date of composition is more helpful, especially since the publication date is in any case likely to be given elsewhere in a catalogue or list (among the bibliographic details). However, if a biography of the writer is being considered, then the time auxiliary indicates the dates of his/her life. Thus:

821.112.5-1"19" 20th century Dutch poetry
(according to when it was written)


929:821.112.5-1"19" Lives of 20th century Dutch-speaking poets
(according to when they lived).

Last updated: 15 December 2006    ^^Top