You have requested a page that has been archived. It has been left at this URL to ensure the persistence of links that have been made to it from other resources. It may contain broken links and out-of-date information.
You have requested a page that has been archived. It has been left at this URL to ensure the persistence of links that have been made to it from other resources. It may contain broken links and out-of-date information.
There is no simple delivery mechanism like interlibrary loan for pictorial materials in Australian libraries. Indeed, the only ways in which Australians can now see pictorial material outside the institution which actually holds it are:
(a) where a reproduction has been published somewhere;
(b) when it
travels on an exhibition;
(c) in the five public libraries participating
in the SLNSW's ISDN telecommunications trial; and
(d) in the few cases
where a library holds a copy of another library's videodisc.
It is not just a question about seeing the material, it is also about finding out about what material exists. Libraries have made very notable resource-sharing progress with their book and journal collections and the Australian Bibliographic Network (ABN) ensures that no more than one library in the country ever need catalogue the same book and ensures that the resources of all libraries are available via loan to all other libraries.
This survey is designed to explore some of the technical issues needed to ensure that a similar outcome can be achieved with pictorial materials.
The survey covered 23 major collecting institutions in three categories:
1. National and State Libraries
2. National collecting
institutions, eg National Film and Sound Archive
3. State-level Museums
Terms of reference for the survey were to:
1. Describe the existing state of bibliographic access to pictorial collections, including the following aspects:
2. Identify major shortcomings or gaps in bibliographic access.
3. Collect sufficient information to support the option of producing a directory of bibliographic access to documentary pictorial collections.
With the agreement of the Steering Committee and partly because of the existence of Jane Symons "Photographic resources of Australia 1990" (which provides an excellent guide to the content of photographic collections), the third term of reference was subtly changed in a manner reflected in this report. The report is in two parts:
(A) a synthesis of findings (Terms 1 and 2); and
(B) descriptions
of present practices and future plans in the surveyed institutions (Term 3).
It is hoped that Part B of the report will be of interest on two counts:
(a) in supporting and detailing the observations made in Part A; and
(b) as a "census" of bibliographic practices in the major institutions. A
notable finding of the survey was the very high proportion of institutions in
the process of evaluating and implementing new imaging technology and it is to
be expected/hoped that the findings of Part B will soon be overtaken by events.
Pictorial curators should find some interest in the experiences of other
collections described in Part B. Part B does not, however, purport to be a
document that would be of any particular interest to researchers or users of
the collections.
The survey methodology was influenced from the outset by perceptions that a mailed questionnaire would be unilluminating and that, in the constant state of bombardment by questionnaire experienced by the major collecting institutions, it would probably go largely unanswered. The other perception, which proved to be justified in many cases, was that pictorial sections were often under-staffed and that anything the consultants could do to help in the information gathering exercise would be welcomed.
The survey consequently involved personal visits by the consultants to each institution during early 1994 supported by a questionnaire which, because of the time constraints on pictorial curators described in the previous paragraph, was sometimes completed by the consultants themselves and sometimes only partially completed.
As part of the questionnaire, state and national libraries were asked to create specimen database records for six photographic items, some individual photographs and others albums or large collections. This was aimed at shedding light on specific differences in cataloguing practice. The state libraries of NSW, NT and SA complied with this but others who had not yet automated said that since they had not yet determined a database format, it would be misleading to comply. Part B of the report under State Library of New South Wales and State Library of South Australia shows examples of the specimen database records.
The survey attempted to cover all pictorial materials - photographs, art works, posters, architectural and engineering plans and drawings and pictorial ephemera such as postcards and illuminated addresses.
The covering instructions to the questionnaire defined "documentary pictorial material" as follows:
The survey covers libraries, museums and archives and is not directed at art galleries. We adopt the simplifying definition that pictorial material in such institutions will almost by definition be "documentary", whether the pictorial material exists as a primary record in its own right or whether it exists as a secondary record to another collection (eg "photographs associated with the collection of fossil molluscs.")
Another simplifying definition of "pictorial" might be as follows: If a user enquiry is unlikely to be satisfied except by looking at an illustration, then the enquiry is probably pictorial in nature. Collections of photographs, art works, posters, drawings, sketches, plans and diagrams are consequently likely to be "pictorial".
A specific exclusion, however, for this particular survey is that cartographic material is not to be included. This includes maps as well as aerial photographs taken for mapping purposes.
The survey is primarily concerned with Australiana. It is assumed, however, that the pictorial collections of the 23 participating institutions are overwhelmingly Australian. Significant exceptions should be discussed with the consultants.
In practice, the survey took a somewhat utilitarian turn and concentrated on materials in the control of pictorial curators or other interviewees. In some cases, unknown and inestimable quantities of relevant material were held in manuscripts collections.
The assumption that collections comprise overwhelmingly Australiana appears to have been correct if the definition is stretched to include some ethnographic and social history collections relating to Papua New Guinea and the South Pacific. Some of State Library of South Australia's maritime photograph collections are not limited geographically and National Film and Sound Archives holds some albums of tourist postcards collected by the singer Peter Dawson on his overseas tours. These are examples of minor exceptions to the general rule that collections are to all intents and purposes collections of Australiana.
There is a strong emphasis in this survey on still photographs because they constitute the bulk of pictorial items held by the libraries and museums surveyed and because their bibliographic control presents unique challenges. Moving images were not included in the survey.
Posters and art works are dealt with separately in Section 7.
The best indication of the size of Australia's photograph collections can probably be gleaned from Jane Symons "Photographic resources of Australia 1990". This 1990 directory of photographic collections quantifies the holdings of 647 private and public institutions and estimates that, between them, they hold about 45 million photographs. The directory is also a particularly good source of information about the subject matter of the collections.
A statistical breakdown of the data in Symons' directory shows that a large proportion of collections are in private for-profit hands, notably 35 newspapers with a total 18 million photographs (or 40% of the total) and 82 commercial stock libraries with a total of 9.7 million items (or 21% of the total). Another large contributor to Symons' totals are government lands and survey departments - primarily large state-level collections of aerial photographs (7 organisations listed with 6.6 million items, or 14% of the total.)
It may, however, be illuminating, in contemplating the extent of photographic collections in Australia's libraries and museums to give a complete tabular breakdown of Symons' figures. This is shown overleaf.
These figures show a total of about 11-12 million photographs held in public institutions (excluding the aerial photograph collections). Assume a collection growth rate of 3% a year since 1990 and add in the much smaller quantity of non-photographic pictorial material and one suspects that as Federation 2001 approaches, we will possibly be dealing with about 15 million pictorial items held by organisations of the sort that normally would be expected to contribute monograph cataloguing data to ABN.
The survey only covered publicly-funded institutions and no attempt was made to survey the large collections held in newspaper morgues (According to Symons, 18 million photographs in 35 institutions). While some curators reported good working relationships with their local newspapers, others were dismissive of suggestions that the newspaper firms would ever generally do anything useful or profitable with their older images. I was told, and this is hearsay only, of one metropolitan newspaper moving from a soon-to-be-demolished building that solved the problem of moving its glass negatives by throwing them down a lift shaft.
Speculating in a probably naive way, since we have a federal department for communications and the arts, is there not some corporate citizenship concept that could get newspapers involved in the Towards Federation 2001 program?
Click here for Table 1: Size of Photographic Collections
Image digitising and compression techniques give a new meaning to the adage "A picture is worth a thousand words". Again it might be illuminating to contemplate the size of the database that would be needed if all 15 million images were digitised and held on a central database. Assuming TIFF files of an average 1 megabyte per image, this would be a database of 15,000 gigabytes. Assuming JPEG compression of images with an average 30,000 bytes, the database would be 450 gigabytes. Assuming fractal compression with an average 10,000 bytes (as used in Microsoft's CD-ROM Encarta encyclopaedia), the database would be 150 gigabytes. Whatever way, or whatever other image digitising technologies were used, the database would be extremely large by anyone's standards. None of these estimates allows for the text component which would be comparatively small in size but would still weigh in at something approaching the size of the existing ABN.
Another way of looking at this is that the data would fit on 150 videodiscs, or on about 7,000 CD-ROMs (30Kb/image compression).
This survey is not a technological survey but it is impossible, given the fluidity of imaging technology, not to refer to the technology since nearly all institutions surveyed are experimenting in some form or another with imaging (eg, evaluating Kodak Photo-CD) or have established in-house committees to look at the future technological and commercial aspects of imaging. With photography, with which much of this survey is concerned, the survey period saw the launch of Apple Computer's $US795 digital camera. Some forecasters see this as presaging a new generation of lower-cost digital cameras which (given whatever marketing push) could well become popular consumer "necessities" and, before too long, spell the end of photographic processing as we know it now. Gadgets launched at this price tend to be available 3-5 years later at $195 so that the forecast is not entirely far-fetched.
The gee-whizzery is not confined by any means to technology directly aimed at imaging. The PC world is in an unprecedented phase of innovation: Pentium 60MHz PCs which have only been on sale in Australia for a matter of months are about to be superseded by Pentium 100MHz machines and, if these fail to compete with the PowerPC of Apple/IBM/Motorola, Intel will no doubt bring forward the launch of its "686" PC range which will probably be another 30% faster. The technical barrier that restricted the IDE standard of hard drives (which are cheaper to manufacture than the present SCSI standard for larger hard drives) to a maximum 520Mb has recently been overcome and I would take bets that 8 gigabyte hard drives will be available within the next five years for under $3,000 and possibly significantly less than that. This sort of development will give medium-sized libraries and museums immense capabilities for handling collections of 50,000 to 200,000 images.
Videodiscing projects have been undertaken by:
as well as other institutions not covered by this survey, eg Australian Stockmen's Hall of Fame at Longreach and Queen Victoria Art Gallery and Museum at Launceston.
Videodiscs generally create a strong positive or negative reaction among people: There is no question in my mind that users love them and that many find them almost addictive. This is a technologically-neutral observation, however, since users would have the same reaction to any image system that delivered images fairly rapidly. One can be confident that systems and networks in the not-too-distant future will be able to deliver images rapidly, highly jukeboxed CD-ROM setups being a probable exception.
Those organisations that have undertaken videodiscing projects seem to be generally delighted with the results, particularly those staff within the organisations who come into contact with users or those who are in a position to contrast the efficiency of the systems with previous frustrations of having to handle and retrieve items before users could view them. The other major plus has come, not from the videodiscs themselves, but from the associated text searching capabilities that overcome the previous limitations of linear filing systems.
Generally, the attitude seems to be "It was the only affordable technology and, if we hadn't taken the plunge, we would have been deprived of years of providing a useful service. The technology will be there to convert the videodiscs to other digitised formats, so altogether we think it has been the right approach."
I personally find it difficult to challenge this view particularly since the technology is now here to make the conversion (and has been successfully employed at the State Libraries of Victoria and New South Wales).
The opposition comes from those who decided to take a wait-and-see approach and there appears to be some opposition to videodiscs within some organisational IT departments, born partly of the fact that videodisc implementations have been seen by them as a PC technology and one which is difficult to network. I also encountered some photographers during the survey who disliked videodiscs on account of their less than ideal photographic quality.
It is possible, since the costs of videodiscing and direct digital imaging are now converging, that the issue is no longer a burning one.
The "information superhighway" and the Internet have spawned a whole new dialect within the space of a few years and words like "gophers" "Veronica" and "agents" are now, if not on everyone's lips, certainly part of the everyday lexicon of many people (including a friend of mine who has just become engaged on the Internet to a fiancee he hasn't yet met). Whether Australia gets a superhighway or an "information Hume Highway" (ie, that gets you to Liverpool or Broadmeadows but not quite into your own home) is one of the burning questions of the day.
Even here, unpredictable innovation may change the scene almost overnight: The May 1994 issue of the American Computer Shopper (p. 61) reports that a small New York firm called DCT (Digital Compression Technologies) has developed a "revolutionary" chip that "makes it possible to send and receive multiple programs simultaneously on various combinations of PCs, phones, fax machines, and TVs - all over a single, ordinary telephone line. Within two years, a $20 decoder chip may be streaming multimedia data across your old copper wire at a mind-numbing 16 megabits per second..... Phone companies, of course, are eying this as a way to immediately deliver multimedia services without costly rewiring."
We also have a case here of an Australian company, Jtec providing internationally competitive high speed data transfer hardware and an Australian firm, Search Tech that has been consistently innovative with imaging projects. I have no links with these firms but suggest that their track records are such as to merit their inclusion in RFT address lists.
State Library of New South Wales is the only one of the institutions surveyed that had initiated a project involving external networking of images (to a test audience of five public libraries). The New South Wales project is described in more detail in Part B of this report.
Nearly all other institutions surveyed have set up working groups for monitoring and evaluating the technology, often within broader frameworks at a whole of government level to achieve more effective networking of government information generally. Governmental interest at both Commonwealth and State levels in information services is, in my experience at least, at an unprecedented level. Imaging is not confined to the needs of pictorial material but is of interest also for analogue imaging of manuscript and other documentary material.
During the course of the survey, I also visited or spoke with several other organisations about image networking plans. The National Botanic Gardens and the Art History Department at Australian National University are perhaps the two that have made the most progress and both have image databases available on Internet's World Wide Web using the remarkable Mosaic text presentation format.
There is, in many organisations, a concern about "reference" quality versus "reproduction" quality of images:
In museums, possibly more often than in libraries, electronic publishing and exhibitions concerns dictate some of the considerations about image quality. (This comparison between the two sectors is a generalisation but holds more-or-less true) Reproduction images are ones that can, say, be incorporated into an annual report or sent off to a bureau to create a 10 square metre photographic backdrop for an exhibition.
There is a concern in many organisations about giving users access to reproduction quality images because of potential problems with unauthorised copying. Consequently, many organisations are talking about the provision of "reference quality" images - images detailed only enough for a user to identify that they might want a copy of the original. They would then be forced (unless, be it noted, they had access to image enhancement software and were prepared to go to the trouble) to approach the institution for a reproducible copy.
One of the attractions of Kodak's Photo-CD appears to be that it simultaneously gives you five different levels of digitisation, ranging from a thumbnail identification version to the full digitised image. (The issue is that, rather than having film developed conventionally, Kodak "develops" it as digitised images, up to 100 images per one-off CD-ROM). I was told of another firm with a similar product/plan and it is easy to imagine a variety of digital photograph developing services springing up with different products (before they get overtaken by digital cameras?).
Photo-CD is being viewed with interest by some organisations which take a lot of photographs themselves as a way of by-passing the developing/scanning stages and simultaneously providing both reference and reproduction quality images.
David Bearman, the American museums consultant who visited Australia in May referred to the concerns about unauthorised copying. If I understood him correctly, I think his recommendation was that cultural institutions should make the best quality images available but coupled with clearcut guidelines and warnings about reproduction and publishing rights. This would reduce the call on institutional copying facilities, would be more timely and convenient to users and would create unequivocal evidence, in the event of a breach of copyright, that a breach had occurred.
Against this, one imagines the debate will be influenced as will much else in this report by considerations of bandwidth, speed of response and data storage costs.
Only two interviewees volunteered that they saw image digitisation as a potential conservation tool for at least parts of their photographic collections, ie, volunteered that they accepted that digitisation might prove a cost-effective alternative to conserving the original photographs as artefacts. It seems possible that, as digitisation techniques mature, there will be a similar national debate during the next decade to the debate of the 1970s and 1980s about maintaining original newspapers in addition to microfilming them. By the next decade, however, one imagines that many photographs will be digital artefacts anyway.
These technological considerations (and others) will have a significant bearing on how Australian cultural institutions make their pictorial collections accessible to the next generation of users. One can predict too that CD-ROM publishing of encyclopaedias and curriculum materials are here to stay and that there will be much use made of images currently held in these institutions. One can hope that there will be innovative publishing programmes for publishing selected images thematically using CD-ROM. One imagines too, particularly with improving compression techniques, that CD-ROM will prove an attractive option for organisations with relatively small collections of pictorial materials wanting to provide easy access to images of the collections.
My feeling is that four factors militate against setting up a central database of pictorial data at the item/frame level:
(a) the sheer size of the database (many hundreds of gigabytes);
(b) lack of the important ABN quid-pro-quo element: the absence of shared
cataloguing opportunities as there is with monograph cataloguing;
(c)
ownership and control issues (in an environment where some have expectations of
turning in tidy profits on their images. Whether this is realistic or whether
the best that will be achieved is a small percentage of cost-offsetting remains
to be seen.)
(d) a rising tide of acceptance of the practicality of
virtual networking as evidenced by the success of and wide interest in AARNet
and Internet.
Others interviewed during this survey would add a fifth reason: that much of the material held in state libraries is of primarily local interest. I think I would dispute this while claiming simultaneously that much of the pictorial subject matter of state libraries is similar, only the geographic setting is different.
This technical consultancy has plainly not been commissioned to determine the future design of Australia's distributed national collection but it has to be based on some premises about its design. My primary premises are that:
Importantly, with ABN itself, emphasis in the National Library's National Document and Information Service (NDIS) redevelopment proposal is given to automatic synonym linking and to hierarchical search capabilities. Another important emphasis is on whole-language translation, via the meta-thesaurus concept where different linguistic classification systems can be automatically translated between or "mapped".
My view is that these devices are fundamentally interesting and intelligent whereas past debates along the lines of which of the terms "Australian Aborigines" or "Aborigines, Australian" was preferable did little to improve services to users. Under an intelligent meta-thesaurus, you might for example simply click on the version you want to appear in your library products and on your library screens and not worry about what anyone else is doing (so long as what they do conforms to a recognised alternative.) (A meta-thesaurus is something which provides automatic linkages between different thesauri, eg that makes possible automatic translation between say Medical Subject Headings and Library of Congress Subject Headings).
Stemming from these considerations about standards, the report below describes two important practical standardisation issues: the depth of subject cataloguing and the use of hierarchies. The latter has represented a major problem for all pictorial cataloguers (as it has in just about every other documentation field.)
There is, let it be said, minority support for the concept of a single database. The National Museum of Australia, for example, was keen to load text records of its individual photographs onto ABN (without images.)
One should not pretend that the virtual network model will not have its drawbacks. A single database would, self-evidently, be a wonderful boon to users and a drawback only perhaps on occasions where they were overwhelmed by the large amounts of information retrieved. A single database would also provide exceptionally strong support for cataloguers and volunteers trying to identify and caption photographs and for national level identification of collection strengths and weaknesses.
An option that has been canvassed during this survey is one of maintaining image data locally but providing descriptive cataloguing about it at a collection group or sub collection level. Under this scenario, the State Library of the Northern Territory might, for example, provide ABN with, say, 500 records that described the collections that make up its pictorial collection. Several institutions said they would be willing to do this, some circumscribing this willingness with an "If we had the resources".
I would give a low priority to this and make the image capture and data documentation backlog which is considerable everywhere the undiluted top priority for pictorial sections of libraries. My strong supposition is that many (though by no means all) of the collection-level records held in many organisations can only be described through fairly bland and general text descriptions "Smith collection of Braidwood scenes", "Jones ship collection", "Brown collection of portraits of prominent Sydney personalities".
If the AARNet virtual network model prevails, as I suggest it will, this problem can and should be revisited in the context of navigation devices generally required for such a network. Reference librarians have traditionally been designated as society's information navigators and it will be interesting to see what role they play in defining these devices. My view is that present constructs of AACR collection-level descriptions of photograph collections are unilluminating but that a reconsideration of the issue in about 1998 would be worthwhile.
Imaging embraces more than just the requirements of pictorial collections. Increasingly, institutions are looking at scanning large textual document and archival collections. As described above, the technology is moving very quickly and many institutions surveyed had established in-house committees to evaluate and monitor the technology for a variety of information service, exhibitions, conservation and commercial purposes.
In my view, it would be worth considering the establishment of an Australian information sharing forum designed
(a) to compare experiences; and
(b) to monitor the technology
(bearing in mind that the Internet hosts similar international forums).
In addition or alternatively, we have here a technological issue group that would warrant the holding of two conferences during the next 3-4 years, possibly under the sponsorship of ALIA. There is no doubt in my view that a well-designed conference programme on imaging technology and experiences would more than pay its way.
It is important in my view that, whatever information sharing mechanisms are used, they take into account the needs of libraries, museums and archives and that inter-sectoral communications on the topic be maintained and enhanced.
There are, generally speaking, significant differences in documentation practices and in the reasons for holding photographs between libraries and large museums. Smaller museums will be described separately.
A. Documentation Practices in Museums
In most large museums, the focus has been on "registration" of objects in collections, which, particularly in the natural sciences, might contain millions of objects or specimens. Provenance and exhibition, loan and conservation history are important concepts in museum documentation. "Registration" has often amounted to little more than the briefest description of the item in an accessions register. "Cataloguing" in museums usually involves incorporating data from the accessions register and the original acquisitions file and progressively building up a contextual history of the object. There is no standard such as AACR whereby a museum can say that it has "finished cataloguing" an object.
There is no common thesaurus such as LCSH though bodies such as the National Museum of Australia and the Powerhouse Museum have tried to develop thesauri and promote their standardisation within other museums. In part, this reflects the enormous range of subject material held by museums and widely differing views among different museums about the contextual significance of similar items. Importantly too, the vast majority of Australia's estimated 1900 museums and galleries have collections numbering in the hundreds or few thousands and complex subject cataloguing practices are usually beyond their requirements.
The lack of standardisation also reflects the fact that museums are often dealing with unique objects and do not benefit in the way that libraries do from shared cataloguing. There has not been the same economic imperative to standardise that there has been in libraries.
In many larger museums, small systems have been set up often on the initiative of individual curators or curatorial departments to handle their own materials. These are often idiosyncratic systems and often use purpose-built coding arrangements in lieu of text or a mixture of code and text. Where an institution-wide database product has been introduced into a large museum, it usually supports a range of different database structures tailored to the needs of individual departments or collections. The Museum of Victoria has over 150 different databases.
Finally and importantly, there is no tradition in museums of public access to the collections. The library catalogue, be it a card catalogue or an automated catalogue, has always been seen simultaneously as a finding tool for readers and staff alike. There is no perceived fundamental dichotomy in libraries between these two purposes despite occasional requirements for different front-ends, ie OPACs. Library catalogues are also geared for retrieval of items that can usually be retrieved from the stacks without difficulty either by staff or by readers themselves.
Reasons Museums Hold Photographs:
It is not quite as simple in a museum to physically assemble all the items retrieved in a search on, say, "Agricultural machinery"! A significant incentive for museums in incorporating images into their databases is to provide "surrogate" access to the collections without having to move or handle objects. There are analogies here with valuable art works, glass negatives etc but not, generally speaking, with conventional library materials.
With the notable exception of the Museum of Victoria, state museums do not hold large general collections of social history photographs. Arguably, the emphasis in museums is usually that photographs are collected to provide interpretative information about other collections. Even with the Museum of Victoria, the photographic curator's interest is in the information content of the photographs (which are captured on videodisc) and not in the physical photographs themselves (which are not necessarily collected). In other cases, for example the Northern Territory, the NT Museum has traditionally given social history photographs to the NT Library Service.
Photography is a vital research tool in the field work of many museum curators. Many are hired on the basis of their particular scientific experience and come equipped with already formed personal collections of slides documenting their earlier research. As a corollary, many take their collections of slides with them when they leave. I was struck by the number of curators during this survey who said they would have difficulty distinguishing between their own photographs and those that belonged to their employer. In some cases, employing museums did not provide cameras or photographic film for paid or vacation field work and some curators suggested that this complicated copyright issues somewhat.
Ignoring this particular copyright complication, however, it is worth noting that, to some extent, progress made in cataloguing photographs in museums merely reflects decisions by individual curators that a database would help them control their growing collections of photographs as an aid to their own research.
Museums frequently acquire photographs of items from vendors or donors before or at the time they acquire an object itself. These are usually filed on files for the object or collection name. Many museums recognise the importance of having a photographic record of items in the collection ("registration photography") and most recognise the desirability of photographing all objects as they are acquired (if only for auditing and insurance/security purposes). At the same time, few large museums have the resources to do this systematically let alone to do it retrospectively.
A particular attraction for some museums in having images incorporated into their databases is the potential for enhanced cataloguing.: Instead of requiring an object description of the type: "Ornately decorated Australian hand-painted slipware beer mug overpainted with idyllic pastoral scenes showing cattle and stockmen and bordered by a stylised wreath of wattle blossom", an incorporated image could provide a preferable substitute for much of the subjectivity inherent in this description.
There are also significant collections of photographs taken during conservation work in museums showing the original condition of items and the progress of physical restoration. Also, there are public affairs type collections showing corporate events such as exhibition openings. The latter category also exists in libraries and often will be of little interest outside the institution. The War Memorial's large collection of public events photographs falls into a different category because it largely documents national ceremonies held at the Memorial
The survey covered only two archives, Australian Archives and the Archives Office of Tasmania, the latter because it, rather than the State Library of Tasmania, is responsible for collecting photographs in the same way that state libraries do in the other states.
It is of passing interest that many of Australia's videodisc projects comprise mainly photographs that might theoretically have ended up in archives rather than libraries, ie, photographs taken by government photographers on official business. While this obviously reflects partly on copyright, it also indicates something about the wealth of material held in Archives.
The entry in Part B of the report under Australian Archives will hopefully give some indication of the wealth of pictorial materials held in state and national archives. The joint projects between the Museum of Victoria, the Victorian Public Record Office and the State Electricity Commission and Melbourne Water are described below under Section 6.2 as an "exceptionally interesting model" for providing access to archival pictorial material.
As with manuscript collections in libraries, the basic collection unit in archives is the "series", a set of records, regardless of physical format, which shares the same provenance. Physical format is a secondary consideration. Although it is a widespread practice in the libraries surveyed for interesting pictorial items in manuscript collections to be copied by Pictorial sections, the survey was unable to elicit how many pictorial items were held in manuscript sections of the libraries surveyed or in Australian Archives.
Although outside the terms of reference of this study, it is worthwhile considering the pictorial collections held in local public libraries and in local museums and historical societies. Photographic collections in such institutions seem typically to number in the high hundreds to low thousands.
There are many instances where state libraries or the National Library and the Northern Territory Museum of Arts and Sciences have acquired photographs from local historical societies and/or where the central agency plays a supporting conservation role by maintaining negatives in preservation environments which are not available to the local societies.
The Australian museums association (known simply as "Museums Australia"), state museums and state arts ministries have been variously active in promoting de facto documentation standards for local museum collections. Victoria, South Australia and Western Australia have all promoted subject cataloguing based on Patricia Summerfield's Historical Collections Classification Scheme for Small Museums with some local modifications. Summerfield's scheme is also used by some museums in other states. There are initiatives in all states to promote standardisation.
Many small museums and local history societies, even where they have not automated their collections, are following these standards. The Victorian standardisation program is possibly the most comprehensive in terms of the detail of its specific documentation and its market acceptance though the Western Australian Museum's WamFoto system provides a similar degree of specificity aimed at photographic collections. An extract from Arts Victoria Museum Unit. The Museums cataloguing manual, 2nd ed, 1988 showing parts of a specimen worksheet for cataloguing an oil painting is shown overleaf:
Note the hierarchical subject and place of manufacture descriptions, the latter analogous to MARC geographical area codes. Such concepts of primary, secondary and tertiary classifications are frequently encountered in museums (as they are, incidentally in many registry systems. Hierarchies are an issue everywhere!). Note too that the sample above illustrates a de facto documentation standard that is in widespread use in Australia's smaller museums, bearing in mind that there are other less widely-used competing standards.
Arts Victoria Standard for Small Museum Cataloguing:
MUSEUM CODE REGISTRATION NO. COLLECTION TYPE Local History OBJECT NAME Painting, Oil NEG. NO STORAGE LOCATION 04.02 SUPP. FILE TITLE "Lost Boy" SUBJECT/S Illustration - Painting, Oil BRIEF DESCRIPTION Painting depicts a child, lost, in the main street of Castlemaine. Scene of the early Castlemaine buildings (streetscape) can be seen in the background. The frame is a simple wooden turned structure. INSCRIPTIONS Marked on lower righthand front: "J. Tyler" MATERIALS Canvas - Paint - Wood - Metal PRODUCTION METHOD Painted WHEN MADE 1890 WHERE MADE Australia - Victoria - Castlemaine MAKER/S Tyler, James HISTORY OF OBJECT This painting is considered an important early painting of the town. It was kept in the artist's family home for many years before acquisition by the museum KEY ASSOCIATIONS Tyler Family SIZE IN CM. 137.5 x 198.5 CONDITION Fair DATE 3 Jun 1988 CONDITION DETAILS Some of the paint is flaking off near the lower lefthand corner (Plus various other data fields relating to PROVENANCE VALUATIONS etc.)
Another example of local history collections photograph cataloguing was recently contained in Martin, Megan and Heath, Andrew. Mainstreaming local studies: integrating and automating intellectual access and collection management of local studies materials. Aplis 7(1): 29-44, March 1994.
The following is the example given there of a full MARC record for a photograph:
LDR Leader km a BRN 001 205166 DFF 007 khrbo DFF 007 kgobj FFD 008 930628s1919 xna CNO 090 0 $a photograph - please ask staff MEU 130 0 $a Kidman & Mayoh shipyard / 1 TIL 245 10 $a Wooden ship under construction at Kidman & Mayoh's shipyard at Kissing Point on the Parramatta River around 1919 $h [photograph] COL 300 $a 1 photoprint $b : b&w $c ; 16 x 21cm COL 300 $a 1 photonegative $b : b&w $c ; 8 x 15cm NOX 533 $a 1 positive photograph $b Sydney, N.S.W. $c Rye Library & Information Services $d July 1992 $e 16 x 21 cm NOJ 541 $a Bloggs, Joe, $b 24 Smith Street, Jonesville, NSW $c original negative donated $d receive June 1992 NPB 581 8 $a Photograph was reproduced as item 5.12 in "Focus on Ryde: a local studies resource' published by the Ryde Bicentennary Schools & Youth Task Force in 1992 SUP 600 10 $a Kidman, Sidney $c Sir $d 1857-1935 SUP 600 10 $a Mayoh, Arthur $d c.1860- SUP 600 10 $a Mayoh, Joseph $d c.1865- SUC 610 20 $a Kidman & Mayoh shipbuilders (Sydney, N.S.W.) SUT 650 0 $a Shipbuilding $z -- New South Wales $z -- Parramatta River $x -- Photographs SUG 651 0 $a Kissing Point (Parramatta River, N.S.W.) $x -- Photographs SUG 651 0 $a Putney (N.S.W.) $x -- Photographs AEP 700 1 $a Gallop, Henry Reginald $d 1890-1958 $e photographer
It has frequently been noted that the cataloguing of photographs differs fundamentally from the cataloguing of traditional library materials in that photographs do not name themselves as books or serials do on their title pages. Rather, they have to be captioned or described or classified or elements in them highlighted to be retrievable from a subject viewpoint. This frequently involves an element of subjectivity on the part of the cataloguer or subjectivity/propaganda on the part of the photographer undertaking the captioning.
A meteorologist confronted with a photograph of a squadron of Meteors returning to base through a storm would very likely caption the photograph: "Spectacular example of anvil-shaped cumulo-nimbus formation presaging a major anti-cyclonic event" or whatever (there is no claim to scientific accuracy here) and hardly notice the presence of the aircraft. A military buff might, on the other hand, describe the aircraft in detail and the mission they were engaged in without saying anything about the weather they were flying through.
This is not a minor consideration. At the Queensland Museum, I was shown an early photograph as revealing as anything one could hope not to see about the degradation of Aborigines showing a broken old man sitting in a squalid humpy with some artefact. The photograph had simply been captioned with the name of the artefact by the anthropologist who took the photograph.
This in itself is historically revealing about contemporary attitudes as are many original captions in social history collections which reveal attitudes that today would be regarded as prejudiced or at best as quaint or unobservant. Undoubtedly, future generations will make similar judgements about 1990s language patterns and cataloguing practices.
A generalist confronted with the undocumented photograph of the Meteors flying through the storm and without any guidance about the origin of the photograph is faced with the dilemma of how much research to devote to describing the photograph. Does one simply describe it as "Formation of military aircraft against a background of heavy cloud."? Or does one undertake the necessary research to identify the aircraft which in turn will give some approximation of the date of the photograph? In practice, photographic curators are confronted with these problems all the time, of using the evidence of parked cars or women's fashion or landmarks in the background to date and locate photographs. This is where volunteer enthusiasts - car buffs, train buffs, ship buffs and the like - can and do play such an important part in helping catalogue photographs.
A related issue, which arises particularly with streetscapes and similar photographs, is the level of captioning detail required. Does one indicate that the photograph shows horsedrawn vehicles and horse troughs and a queue outside a bank and a tram in the distance etc.? Here again, the best results are probably achieved by people with the experience to know that the horsedrawn vehicles are pretty much standard for the time but that the horse troughs are unlike most others of their time and are worth cataloguing. There is, however, in my view at least, a frequent tendency towards over-inclusiveness and there are many instances in videodisc projects where if you search, say, for "Cats", you will retrieve a photo of a class of schoolchildren where the girl in the back row is holding a barely-discernible cat. This causes immense potential problems in searching, say for the Sydney Harbour Bridge, where thousands upon thousands of photographs may show the Bridge in the background but not as the primary subject of the photograph.
Note, however, that these sort of problems only loom large with large collections. If one is cataloguing a small local history collection of a few thousand photographs, one might tend instinctively to be more inclusive.
Drawing these discussion threads into some practical observations:
6.1.1. Cataloguing styles are often conditioned by the capabilities of the software used for searching. There is no tradition of differentiating between occurrence of a word in a title and as a subject heading to give different weighting to retrieval concepts. There is no tradition of other forms of term weighting, eg "major" vs. "minor" descriptors to differentiate between subjects captured incidentally in a photograph versus subjects that are the focus of the photograph. My view is that, if there were an attempt to consolidate various image databases into a single large database, (which, as suggested elsewhere, is unlikely in my view) this will create significant retrieval difficulties with frequently occurring subjects such as Trains, Horsedrawn vehicles, Buildings, Agricultural machinery etc.
My view is that the distinction between text and subject heading fields should be promoted as a national "guideline", ie, as something a little short of a "standard": Major/minor distinctions are difficult to implement and slow cataloguers down.
My recommendation, and it applies to nearly all the institutions surveyed, is to use text fields to describe "any mention" of a subject. Use subject heading fields to describe the main subject only. This will give a straightforward way of restricting searches to important images about a particular subject. In system design terms, it also represents a simple proposition for all database designers and, at the user end, is relatively simple to cater for.
The issues raised here should be considered by organisations with collections that have not yet been automated. I do not suggest that organisations that have already automated should rework their data.
6.1.2. While large organisations have access to a range of research tools useful for cataloguing photographs, albeit via a diverse range of encyclopaedias, pictorial reference works, etc., the absence of such tools represents a real problem for smaller organisations. In my view, an illustrated thesaurus/dictionary of major social history objects (eg, types of horse-drawn carriage; different types of dairying technology) would be extremely useful to all involved in photographic cataloguing and have enormous potential for time-saving and improved documentation. In larger organisations, one could imagine too that an online version of the thesaurus would enrich photographic databases for both cataloguers and users. This would be the case particularly if traditional thesaurus concepts were extended to include the notion of the thesaurus as an encyclopaedia (ie, gave the history of the object and information about its social significance.) In small museums, the thesaurus would have applicability for all the cataloguing of museum objects, not just photographs.
This concept is not unconnected with other options that are becoming available to libraries generally to employ their authority files better as information navigation tools.
6.1.3. This paragraph is meant to illustrate further the observations made under 6.1.1 and 6.1.2 as well as to query the utility of some of the more complex database structures encountered during the survey:
The best style for the subject description of photographs in my view is via a formalised caption such as is used by the Australian War Memorial. (I am not making any points here about the physical description of photographs as artefacts or about collection management elements in database descriptions.) Quoting from the Memorial's entry in Part B of this report:
"The preferred database description is one where most significant data is entered in a single data field via a formalised caption of the form: Date - Place - Subject - Associated information - Detail, eg:
"1946-05-08. Heidelberg Victoria. Men of the 2/3rd Australian Motor Reserve Transport Company playing a game of whist in the grounds of the 115th AGH. They had recently been repatriated from Changi. Identified personnel, from left to right, are (1) Pte R. Goldsworthy, (2) Pte K. Benson, (3) Sgt P. Phillips and (4) WO1 R. Sharpe.
"Thesaurus headings are added automatically into separate fields:
Places: Heidelberg Changi Units: 2/3 Reserve Motor Transport Company 115 Australian General Hospital Topics: Card games Conflict: W2"
We can contrast this with the style used by National Museum of Australia (This is by no means the most complex data structure encountered but it serves as an example):
Brief Description: Framed b/w photo of the 1937 Australian Women's Cricket Team, England, June 1937. Timber frame stained teak colour. Association Details: Person: Ms Mollie Flaherty; Ms Amy Hudson; Marie Jegust; Mrs Alice Smallman (and 13 other names) Event: Date: Jun 1937 Jul 1937 Place: England/Great Britain Subject Terms: Sports; Cricket; Women's cricket; Photographs
The War Memorial would describe this something as follows:
"June 1937. Somewhere in England. Team photograph of the Australian Women's Cricket Team taken during their 1937 tour of England. Front row (L) to (R): (1) Mollie Flaherty; (2) Amy Hudson (Vice Captain); (3) Marie Jegust (Captain) etc. etc."?
This style is a lot more informative and a lot simpler than the one used.
Taking the subject-related data from Ryde Public Library's MARC record:
MEU 130 0 $a Kidman & Mayoh shipyard / 1 TIL 245 10 $a Wooden ship under construction at Kidman & Mayoh's shipyard at Kissing Point on the Parramatta River around 1919 $h [photograph] SUP 600 10 $a Kidman, Sidney $c Sir $d 1857-1935 SUP 600 10 $a Mayoh, Arthur $d c.1860- SUP 600 10 $a Mayoh, Joseph $d c.1865- SUC 610 20 $a Kidman & Mayoh shipbuilders (Sydney, N.S.W.) SUT 650 0 $a Shipbuilding $z -- New South Wales $z -- Parramatta River $x -- Photographs SUG 651 0 $a Kissing Point (Parramatta River, N.S.W.) $x -- Photographs SUG 651 0 $a Putney (N.S.W.) $x -- Photographs AEP 700 1 $a Gallop, Henry Reginald $d 1890-1958 $e photographer
The marginal utility of just about everything outside the Title field here is questionable in my view or, more accurately, is expensively obtained.
I think that what is really wrong with much of this stems from 1960s-1970s concepts of database construction. The War Memorial would deal with the photograph above in something like the following manner:
Title: 1919. Kissing Point, Parramatta River. Wooden ketch (possibly the Aardvaark?) under construction at Kidman and Mayoh's shipyards. Kidman and Mayoh was established by Sidney (later Sir Sidney) Kidman, Joseph Mayoh and Arthur Mayoh in the 1890s and was employed by the Commonwealth Department of Defence to build a number of coastal vessels during World War I" Place: Sydney Topic: Shipbuilding Ketches Conflict: PTW1 (ie, Post World War I)
(There is no claim on my part to historical accuracy here).
Once again, I think this style is significantly more informative than the one used by Ryde Public Library. It is also, importantly, one that can be taught relatively easily to volunteers or anyone else without the arcanities of MARC and it permits dozens of photos rather than a handful to be catalogued per person-day.
The style, however, also illustrates the weaknesses of flat file database constructs because, if you have several photos of Kidman & Mayoh's shipyards, you then have some problems. Do you repeat the historical information about the shipyard for each photo? If you do, you get a lot of irrelevant photos in a text search, say, on "Department of Defence". You can solve the problem by using authority file or hypertext constructs or by adding a new, unindexed or differently indexed field, eg:.
Title: 1919. Kissing Point, Parramatta River. Wooden ketch (possibly the Aardvaark?) under construction at Kidman and Mayoh's shipyards. History: Kidman and Mayoh was established by Sidney (later Sir Sidney) Kidman, Joseph Mayoh and Arthur Mayoh in the 1890s and was employed by the Commonwealth Department of Defence to build a number of coastal vessels during World War I"
You can do better still, however, with hypertext and hyperthesaurus constructs where hyperlinks are used to open up progressive views of associated data, eg:
Title: 1919. Kissing Point, Parramatta River. Wooden ketch (possibly the Aardvaark?) under construction at Kidman and Mayoh's shipyards
and where clicking on one of the highlighted text items gives you further information, eg:
Kidman and Mayoh
A shipbuilding company with shipyards at
Kissing Point on the Parramatta River and at Bermagui NSW. Kidman and Mayoh was
established in the 1890s by.....
Aardvaark (RAN ketch)
The Aardvaark was a vessel of xxx tons
commissioned in 1919 by the Commonwealth Department of Defence from Kidman and
Mayoh shipyards and used by the Royal Australian Navy for hydrographic work in
the Pacific until 1935 when it was shipwrecked off .....
Ketches
A Ketch is a fore-and-aft rigged vessel with a large
mainmast and a smaller mast aft, but forward of the rudder post (Macquarie
Dictionary). See illustration
Click on illustration to see an illustration of a typical ketch.
It is this sort of encyclopaedic and illustrative material that I referred to under 6.1.2 above that I believe would, if shared between institutions, provide extraordinary interest for the documentation of pictorial material.
The example quoted above also illustrates the potential benefits of relational or hyperlinked databases over traditional flat file structures.
6.1.4. It is a pity in my view that information retrieval systems do not routinely provide a data field or menu option to allow users to comment about the accuracy or adequacy of cataloguing records. This facility would be a major boon for photographic cataloguing where many users will have particular knowledge and can contribute much detail that is beyond the scope of generalist cataloguers. I do not suggest that the facility would attract only constructive or accurate comments but I do suggest that, over time, it would lead to significant enrichment of databases.
My impression is that volunteers have been used in a restricted range of subject areas, ie usually trains, aircraft, cars and ships. Under the arrangement suggested above, it would not be unreasonable to predict that various retired health care, architectural, farming and other experts could be drawn into the web of turning from occasional commenters into regular volunteers helping with systematic data enhancement. At the same time, I recognise that managing volunteers is not cost-free.
6.1.5. As someone who has worked with computerised information retrieval systems since 1970, it is striking (to me at least) how much the application of automation was deferred in pictorial collections generally. While state and national libraries have had access to computers with mainframe grunt for many years, this has not generally been the case in corresponding museums. Even within libraries, the application of automation to pictorial materials has been generally deferred. I am still at a loss to understand why typed captions lists are being maintained anywhere!! This is not a reflection on Pictorial sections.
6.1.6. The State Library of South Australia provides an interesting example of a case, the Arbon-Le Maistre collection of shipping photographs, where a donor is voluntarily cataloguing the collection onto a library-specified database at home. This may be something for other institutions to bear in mind, where they can anticipate a collection that is likely to come their way, and where they can persuade the donor not only to document it properly, but to do so in a way that avoids future rekeying and reworking.
The State Library of Victoria's system has the capacity to link a single text record with several videodisc images, eg six images of the same building taken simultaneously but from different angles. Elsewhere (and this is a generalisation) there has been a tendency in some of the videodisc projects to create individual text records for each item at the frame level. I am not certain whether this reflects software or logistical or tactical constraints though some elements of each may be present. (Technically, it is easiest to link a caption to a single image; logistically, it is easiest to batch items for videodiscing in accession number order or whatever; tactically, since the money for the project has been found at last, why not videodisc all the images?) This is not necessarily a bad approach though it does occasionally require repetitive data entry and can cause frustrations to searchers who encounter a batch of similar images that do not interest them.
SLNSW, NFSA and AIATSIS employ relational database structures linking collection-level records with individual images. There may be other examples that I saw without recognising the underlying data structures, but my impression is that most systems used for cataloguing photographs are record-based. It is a characteristic of these three relational systems that false drops are encountered more often than in searching flat file databases owing to the textual complexity of collection-level records and this represents an interesting system design issue.
The Museum of Victoria is the only organisation surveyed which had taken a highly-selective or representative approach to cataloguing a major collection. The Museum reached an agreement with the Victorian Public Record Office and the State Electricity Commission to videodisc the 10,000 most interesting of the 250,000 SECV photographs held by PRO. It has made similar arrangements with Melbourne Water and with Esso Australia. In my view, this is an exceptionally interesting model because it allows access to representative photographs of switchyards, substations, transmission lines and the like without over-burdening the system with dozens of repetitive images.
My supposition is that this would be an excellent way to open up access to archival and corporate images without the potentially crippling burden of imaging everything.
This issue, of whether to digitise or videodisc every photograph in a collection, is one about which there is very little agreement and little prospect of national standardisation. The issue is partly related to the size of collections. For example, the State Library of Northern Territory happens to have a collection that handily fits on a single videodisc. On the other hand, the Australian War Memorial is proceeding with a comprehensive imaging project for its 1 million images. The issue is also related to the significance or pictorial interest of the sub collection, eg an album containing family scenes of marginal general interest might merit only a single or few representative images as a guide to the general content of the album or the selective cataloguing of the most interesting images.
The prospect of not having a national standard probably does not matter under the virtual network model that it is suggested will prevail.
Posters constitute an interesting smaller group with two outstanding thematic collections of an estimated 65,000 posters (primarily cinema, music and performing arts posters) held by the National Film and Sound Archive and of 9,000 held by the Australian War Memorial. State Library of Victoria and State Library of Tasmania hold relatively large collections of 21,000 and 6,000 respectively. Australian Archives holds series of army recruiting, tourist and other government posters which almost certainly constitute an interesting resource. Elsewhere, poster collections number in the high hundreds to low thousands and typically comprise a mixture of theatrical, tourist, political and government posters. The total number of posters held in the organisations surveyed (excluding archives collections where their numbers are unknown) is about 120,000.
Cataloguing progress is most advanced at the War Memorial where all posters have been catalogued and many imaged. At the National Film and Sound Archive, posters are retrievable via manual or database entries at the production name level.
At State Library of NSW, Mortlock Library and State Library of Tasmania starts have been made to catalogue individual posters and formats and styles used are described under the relevant entries for these organisations in Part B of the report. Elsewhere, posters have usually been treated as ephemera, ie, sorted into broad subject categories and simply stored in map cabinets.
The survey did not aim to cover art galleries though it did encounter two organisations that are simultaneously museums and galleries (Tasmania and the Northern Territory museums). Generally speaking, few of the institutions surveyed collect art works and those art collections held in libraries tend to be part of important earlier collections donated to the libraries concerned, eg Dixson and Mitchell collections at State Library of New South Wales, Allport collection at State Library of Tasmania and Nan Kivell collection at National Library of Australia. Specimen cataloguing records for these collections are described under Part B of the report. The Australian War Memorial has a collection of 21,000 art works mostly by official war artists. This is also described in Part B.
There is a palpable common lineage in the systems used by the state libraries, the State Archives of Tasmania and the National Library for cataloguing photographs which reflects some similarities in the social historical content of their collections and past liaison between pictorial librarians. Despite some linguistic and organisational differences, the primacy given to place and the subdivisions available for places are common features of most systems. Other similarities relate to the use of hierarchical constructs to group suburbs together, eg Brisbane - Suburbs - Indooroopilly; Canberra - Suburbs - Hughes.
Manual systems based on self-indexing principles tend to have the most generic forms of topical headings. In other words, systems where all photographs on a particular subject are placed on a single file tend to have broader subject headings. One suspects that this reflects an understanding of the importance of generic searching for most users.
Hierarchical or generic searching capabilities were absent in all software surveyed except the Australian War Memorial, and the yet to be implemented thesauri/systems at the National Museum of Australia and the Powerhouse Museum. The War Memorial is the only institution surveyed that has implemented an hierarchical approach to cataloguing its pictorial materials to date (or, to be precise, that has used formal hierarchies in indexing and which has software that can utilise these hierarchies automatically).
Devices used to offset the difficulties the absence of hierarchical searching creates in other institutions include:
:: standard AACR-style parenthetical additions for place names, eg Adelaide (S. Aust)
:: chaining descriptions for suburbs, eg:
Brisbane - Suburbs -
Indooroopilly
Canberra - Suburbs - Hughes
:: chaining descriptions for topics, eg:
Restaurants -- Retail
stores -- Buildings
:: the use of broad categories to facilitate generic searching, eg:
AIATSIS assigns one of about twenty broad categories to each of its
photographs, eg ART, HOUSING, HEALTH. Similar restricted list of "keyfields"
used in WamFoto at Western Australian Museum and elsewhere in museums.
SLNSW assigns codes to indicate regions within the state, eg Far North Coast, Central West etc.
The curator of the ethnology photography collection at Museum of Victoria is reviewing subject cataloguing practices in a deliberate attempt to cut back to about 500 headings.
Codes in libraries. In addition to the SLNSW's geographic codes described above, the Mortlock Library employs a coding system to restrict searches to "city acres", ie particular survey grids in the Colonel Light planned part of the city.
Codes are frequently encountered in museums. The historic sites register at the Northern Territory Museum is heavily-coded. (This is not strictly a pictorial database but it provides access to files that are supported by photographic documentation.)
The spreadsheet used by the Western Australian Maritime Museum to catalogue its significant underwater archaeology collection uses codes which function simultaneously as a filing device and as an hierarchical search tool, eg under each shipwreck, .6 means metal objects, .61 means bronze objects etc. Queensland Museum uses survey map grid references to describe places at a regional level.
Other vocabulary control issues in museums:
National Film and Sound Archive, which has to date concentrated largely on cataloguing its newsreels and has yet to make inroads into its pictorial materials, started by using the APAIS thesaurus as a model for topical headings. It is currently reviewing its subject cataloguing practices which have already become far removed from their APAIS origins.
Australian War Memorial and National Museum of Australia have developed their own thesauri to handle all their collections, not just photographs. These thesauri bear very little relationship to LCSH or to the general styles used in libraries to catalogue photographs. The Powerhouse Museum has developed a thesaurus for museum objects (eg household objects, musical instruments etc.) and hopes to further the development of a subject thesaurus for use in cataloguing the social context of its collections including its photographic collections.
Finally there are cases in museums where no vocabulary control is used other than broad syntactical or stylistic conventions for entering the captions of photographs and where no attempt is being made to subject catalogue items.
There are three vocabulary control problems which information specialists need to address in organising information for users:
(a) Homonyms:
(The same word or phrase with a different
meaning according to context)
In essence, authors with the same names are homonyms and librarians are familiar with qualifying devices such as dates of birth and death to distinguish between them.
Examples of homonyms are words like "act" which could be the Australian Capital Territory or an enactment or "aids" which could be a disease or an audiovisual presentation or "pup" or "camel" which could be an animal or an aircraft.
In specialised and smaller collections, homonyms generally represent only a minor nuisance. In large and generalist collections, they need to be catered for unless one is prepared to accept a degree of noise in searching.
(b) Synonyms:
Synonyms generally represent a bigger problem
because many things can be represented via a variety of alternative forms,
initialisms and acronyms, eg:
US, USA, United States, America(n)
Libraries generally have established means for handling synonyms, either through the formalities of AACR or through authority files which provide see or use references. The concept of "vocabulary control" is too often thought of as being equivalent to the problem of linking synonyms and labelling the result as a "preferred term". Libraries, in a minority of cases but increasingly, have means of automatically translating non-preferred terms into preferred terms, ie of linking synonyms.
(c) Hierarchical relationships:
The problems of homonyms
and synonyms pale into insignificance when compared to the problems of
hierarchical relationships. This is an issue group that embraces questions of
broader and narrower terms, questions of specificity and granularity, questions
of see under relationships, questions of "effective synonymy" and the
like.
As described above, the representation of hierarchies represents a major problem in describing pictorial collections as it does in most documentation endeavours.
Take a simple hierarchy:
Motor vehicles
Cars
Austin cars
Ford cars
Oldsmobile
cars
Trucks
Ford trucks
In conventional thesauri, this is represented as:
Austin cars Ford cars
BT Cars BT: Cars
Cars
BT Motor vehicles Motor vehicles
NT Austin cars NT: Cars
Ford cars
NT: Trucks
Oldsmobile
etc etc...
Take another hierarchy:
Victoria
Victorian Western District
Casterton
Coleraine
Hamilton
etc etc...
Say a user wants images of vehicles in the Western District of Victoria and the system does not link these terms systematically. You then have to conduct a Boolean search of the type:
(Motor vehicles or cars or Austin cars or Ford cars or Oldsmobile cars or Trucks or Ford trucks etc etc) and (Victorian Western District or Casterton or Coleraine or Hamilton etc)
or, you ignore the thesaurus and search something as follows:
(car or cars or truck# or vehicle#) and (western district or casterton or coleraine or hamilton)
With a systematically hierarchical system, all you need to do is search:
Motor vehicles and Victorian Western District
and you will retrieve all required items whether they are images of an Austin A40 in Coleraine or a Ford Prefect in Hamilton or an Oldsmobile in Casterton.
The Australian War Memorial is the only institution surveyed to have implemented a systematically hierarchical system though National Museum of Australia and the Powerhouse Museum are in the process of implementing similar systems.
Once you have this capacity, then nearly all the concerns people have about granularity fall away, you can be as specific or general as you like in your subject cataloguing provided the system makes links between general and specific subjects.
The two extremes in thesaurus construction are represented
(a) by LCSH which tends to create a new heading every time a specific
new subject is described (the "expansive" style); and
(b) by the APAIS
thesaurus which deliberately restricts itself to about 1800 terms and which
groups numbers of specific subjects under these terms (the "restrictive"
style). There is, in my experience, a great divide in the library profession
between those with a cataloguing background who tend to prefer LCSH styles (and
who generally have little experience in searching) and those with an
information retrieval background who tend to prefer APAIS styles. This again is
a fairly sweeping generalisation.
In systems without systematic hierarchical searching capabilities, my view is that users have been done little service by "expansive" concepts of subject cataloguing. They simply create a requirement for users to conduct several search reiterations or complex Boolean searches of the type described above. The problems are compounded by data field restrictions in searching - the usual construct in library search systems - where one elects to do a search either on a subject or on a text field. It is often cumbersome or impossible to both simultaneously.
Taking the Textpert OPAC system at the War Memorial (in which I have a pecuniary interest), this is the only system I know of in Australia that simultaneously employs the benefits of both text and thesaurus searching and which does so automatically, ie it is programmed to recognise when and how to combine thesaurus and text searching.
Alan Walker's thesaurus for Australian historical photographs constructed during the NSW Government Printing Office's videodisc project now managed by the SLNSW is, in my view, the closest thing Australia has to a systematic thesaurus of terms appropriate for indexing Australian social history photographs. With the SLNSW's agreement, I believe, if a standard Australian thesaurus is to be promoted and developed, large elements of the work done on this project will prove invaluable though there is still some editorial and intellectual work to be done to made it more widely useful and inclusive.
An excellent description of the design criteria and philosophy underlying the thesaurus is contained in Walker, Alan. Building and Australian thesaurus : indexing Australian historical photographs. Paper delivered at the 10th National Cataloguing Conference, Australian Library and Information Association, Fremantle, 4-6 November 1993.
One of the virtues of Alan's thesaurus is its hierarchical arrangement. My view has been for a long time that thesauri need to be based on formal hierarchical principles. Otherwise one finds, as one does sometimes with LCSH where if you chase through the BT/NT reference structures, terms that are defined simultaneously as great grandchildren and as parents of other terms. These considerations will loom very large as system obstacles in designing the new generation library systems.
While they are hierarchically arranged, Walker's hierarchies are not built into the SLNSW's search system so that one still needs to formulate complex Boolean searches of the type shown above.
This has been identified as a subject for further study in the Towards Federation work programme. I offer the following observations as being selectively pertinent to this question and as being central to the bibliographic description of pictorial material in the social history area.
1. Place Names
Place is an important delimiter in many searches for photographs and often the primary search key. Descriptors are usually at about the level of towns or suburbs, ie they are usually at about the level of a postcode.
My view is that an hierarchical arrangement of places can be easily fashioned which includes state, ABS Statistical Division, conventional region, Local Government Area and places down to the postcode level, eg:
New South Wales Greater Sydney Sydney NSW 2000 The Rocks NSW 2000 Sydney Inner West Burwood Municipality Burwood NSW 2134 Croydon NSW 2132 Croydon Park NSW 2133 Leichhardt Municipality Annandale NSW 2038 Balmain NSW 2041 Birchgrove NSW 2041 Forest Lodge NSW 2037 Glebe NSW 2037 Leichhardt NSW 2040 Lilyfield NSW 2040 Rozelle NSW 2039
Note that in a systematically hierarchical system, searches at any level of the hierarchy will optionally retrieve all subordinate terms, eg "Fire stations Sydney Inner West" will search for Burwood or Balmain or Glebe etc etc. This represents a significant improvement on AACR capabilities.
My preference is to employ postcodes in the manner shown because users are familiar with their meaning and because they are easily applied. They also permit a new level of searching by postcode, eg NSW 2040 or NSW 204# give new useful access points.
The Sydney NSW 2000 device enables one to restrict searches to the city itself, a capability that is unavailable with AACR constructs.
Similar hierarchical arrangements can be simply designed for river/drainage systems (which are tailor-made for branched hierarchies) and for such things as mountain ranges, (eg Victorian snowfields), state and national parks etc.
With geographic information system data behind such a thesaurus, another layer of utility could be constructed. I am not certain what priority this would have, but it would be relatively straightforward to automatically assign geographic coordinates every time an item was indexed "Glebe" or "Leichhardt" or whatever.
In such a case, the geographic coordinates simply represent an example of an attribute of the thesaurus term. Systems which automatically assign enriching attributes are potentially extremely useful in promoting both consistency and retrievability:)
Take as an example Paterson's curse which has a variety of "attributes": its scientific name is Echium plantagineum and it is a member of the Echium genus of the Echiaceae family or whatever. It is known as Salvation Jane in South Australia. It is an introduced plant and it has been declared a noxious weed in some states. Most of these attributes are potentially useful in searching and, using a reasonably intelligent thesaurus, whichever of the attributes one thinks of as potentially useful to users can be assigned automatically by the system. This sort of approach to thesaurus design has extraordinary potential.
Take another example: Bronte, Coogee, Bondi Beach, Manly etc. are all beachside suburbs. Say the system routinely attached "beachside" to anything indexed with one of these terms. You then have another hierarchy, say:
Coastal Beachside
and you search for "Sydney coastal vegetation". The system automatically finds items on vegetation in Bronte, Coogee etc.
An hierarchical thesaurus of place names specified down to about the level of postcodes as described above has the makings of a national standard which could be applied to a wide variety of library material. In search systems with an hierarchical search capability, it would remove the requirement for parenthetical qualifications, eg "Melbourne (Vic.)
Pictorial collections frequently require vocabulary control at a lower level than this:
(a) at the street-level, eg Pitt St, George St, Martin Place
(b) at
the landmark level, eg Cockatoo Island, Observatory Hill, Domain.
Some of these in turn are amenable to hierarchies, eg:
Sydney Cove Campbell's Cove Circular Quay
(c) at the building level, considering buildings for the sake of the discussion as places rather than topics, eg Sydney Harbour Bridge, Sydney GPO, Hyde Park Barracks.
Extending this line of argument to topical headings:
the issues for a nationally standardised thesaurus for pictorial materials relate essentially to reaching agreement on a common level of specificity to be applied somewhere within each indexed record, supplemented by local thesaurus or textual enhancement.
The photograph collections of the National Library and state libraries/archives of Queensland, Tasmania and Western Australia have not been automated. There is an opportunity to promote a common thesaurus for these collections. There is also scope for other smaller collections to use the same thesaurus. My view is that it would be highly worthwhile promoting this common thesaurus.
The main elements in this standardisation process are:
Agreement re the general concept of postcode level specificity for places plus agreement about narrower exceptions. Exceptions might be notable or iconic landmarks of the type described above as well as such things as Jenolan Caves, Twelve Apostles etc.
City street names are probably of state rather than national interest and can either be handled via local extensions to a national thesaurus or via text. A useful de facto standard would be to agree on whether to abbreviate or spell out street names, eg Collins St vs Collins Street, Wickham Tce vs Wickham Terrace etc.
Alan Walker's thesaurus contains a number of names of individual yachts, eighteen-footers, racehorses and the like. SLNSW is currently in the process of editing these out of the thesaurus while maintaining them via authority lists. This is a welcome move. However, there is something to be said in my view for maintaining thesaurus terms for notable instances of generic subjects, eg:
Bridges Derwent Bridge Gladesville Bridge Princes Bridge Sydney Harbour Bridge West Gate Bridge
rather than indexing all of these "Bridges" and creating retrieval difficulties.
Similarly:
Airports Bankstown Airport Essendon Airport Kingsford Smith Airport Moorabbin Airport Tullamarine Airport
but indexing, say, Bowral airstrip as:
Airports Bowral NSW 2576
The selective considerations are not meant to address all the issues involved in an Australian thesaurus but they may be of interest.
|
Continue to Part B of the Report (*edited version) or Return to Contents Page |
NATIONAL LIBRARY OF AUSTRALIA, Canberra, ACT 2600,
AUSTRALIA, ABN: 28 346 858 075
Telephone + 61 2 6262 1111; Facsimile +61 2
6257 1703; TTY: 1800 026 372