Images as Evidence
June 1999. This article was first published in Museums and the Web 97: Selected Papers. Pittsburgh: Archives & Museum Informatics, 1997. Pp.347-361 (based on a paper given at the Museums and the Web 97 conference, Los Angeles, March 1997 ). Republished in the VRA Bulletin, Vol.25 (Spring 1998), No.1, pp. 58-66.
Many of the web sites linked in this article have been updated since this article was published.
This article may be used freely for non-profit educational purposes. All other uses must have the written permission of the author and original publisher. Reproduction of any of the illustrations must also have the permission of the owner of the work of art illustrated and of the private individual or institution that posted the image on the web.
Not only art history, but all disciplines that study artifacts in detail rely heavily on images as evidence. Among the humanities and social sciences, this includes not only art history, but also archaeology, anthropology, and increasingly social history. Moreover, some of the sciences, among them astronomy and biology, including medical research, are heavily dependent on the careful study of images. These images are needed not only by advanced scholars. but also by the thousands of graduate students writing seminar papers and theses, by students at all levels taking courses, and by members of the public, many of whom have a serious interest, often specialized expertise, in these subjects. Yet the importance of these images is distressingly undervalued by those who do not rely on them as evidence, and the characteristics that make them useful for research are too seldom recognized.
Digital imagery offers us the opportunity to make available for the first time large numbers of high quality color images which would be prohibitively expensive to reproduce in hardcopy. The opportunity exists for a dramatic expansion in the amount and quality of visual evidence for the manmade and natural worlds. To date, this opportunity has been taken advantage of by only a few pioneering institutions. The remarkable evidence retained in images, which digital imagery should finally bring center stage, continues as undervalued as before.
The Jan.-Feb. 1997 issue of Museum News, the journal of the American Association of Museums, features the article "Perfect Site: Museums on the World Wide Web." The editors asked seven experts (the director of a major art gallery, the executive director of a science center, a museum information management consultant, an exhibit developer, a director of print and electronic media, the assistant director of a museum of archaeology and ethnography, and a director of new media initiatives) to "respond to the question, what qualities make for an outstanding museum web site? We then asked our experts to use these criteria to choose five museum sites they consider to be among the best available on the web." In the published replies, there are occasional mentions of attractive graphics and good design, but nowhere in the list of characteristics or in the characteristics of the five sites each expert selects is the importance of images of objects in museum collections given more than passing mention. To be fair, two of the experts do mention images. Diane Zorich writes that "Museums that have developed outstanding Web sites . . . offer information in a well-organized digestible form that gives the user plenty of options (e.g., "text only" versus full images, browsing versus a search engine or site index) . . . ." and that "Many museum Web sites are beautifully designed but do not merit a second visit because they offer no incentives beyond a view of nice images." Katherine Jones-Garmil is the lone advocate for images, recommending the Minneapolis Institute of Arts as "a great resource for educators, rich in images and developed content." Among these seven discussions of characteristics and the listing of characteristics of over twenty-five different exemplary museum sites, that's it for images, let alone high quality, accurate, or detailed images.
Is this an exceptional example? I don't think so. The commercial world fully recognizes the value of images, though not often of accurate ones. In the museum and academic worlds the importance of images and the characteristics that make them so are consistently undervalued; this in spite of the pervasive influence of images in our modern world. Let me attempt to fill the near vacuum in the Museum News survey by describing sixteen characteristics of images that seem to me important if museum web sites are to provide for the research needs of their diverse publics. These are not technical characteristics that only computer experts might recognize. These are practical, common-sense characteristics needed by any serious user. I shall attempt to illustrate these characteristics with examples from a few exemplary sites that have begun to take the need for images seriously and have found the funding to do so. Clearly this is an expensive and time-consuming undertaking and there is no assumption here that such images can be made available without a major commitment on the part of institutions and funding agencies. There are also major issues of copyright yet to be resolved. Partly for these reasons, it is important to be clear about the importance of high quality images and the characteristics that provide for research. It will help if we divide these into two groups: characteristics of images themselves and characteristics of information about images.
Important Characteristics of Images
1. Quality of individual images
What one means by "quality" varies according to what one values in an image. If one is attempting to sell posters illustrating a work of art, quality may lie in the image's beauty or commercial appeal. But when one is conducting research, we would usually prefer to be looking at the work of art, the archaeological artifact, or the botanical specimen itself. Hence, as evidence, images are valued to the extent that they approximate what one would see if looking at the object itself.
The Spencer Museum of Art Printroom at the University of Kansas has put on their web site numerous high quality images at approximate life-size (fig.1) with enlarged details. Especially notable are seven image maps allowing one to examine all areas of seven prints in detail. These serve as examples of seven basic printmaking processes, the much-enlarged details allowing one to see the important distinctions among print techniques. The site also includes an informative introduction to collector's marks, inscriptions, monograms, and conservation issues, drawn from the Spencer's print and drawing collection, with high quality, much-enlarged details and accompanying text. To date, sites such as this with individual high quality images seem to have resulted from collaboration among a few individuals who recognize the need for high quality images. In the case of the Spencer Museum (http://www.ukans.edu/~sma/prints.html), the three collaborators were Stephen Goddard, Curator of Prints and Drawings, Jeff Bangert, computer specialist, and Robert Hickerson, staff photographer.
Images of this quality are more difficult to achieve for paintings and, especially while issues of copyright and fair use are unresolved, museums have been reluctant to put images of this quality on-line for one-of-a-kind objects such as paintings. For research purposes, however, the more unique the objects, the more important are the images; and the larger and more complex the painting, the more necessary are high quality details. The best I have seen are many of the images on the new site of the National Gallery of Art, Washington, D.C., where Neal Johnson has been working from 8 x 10 inch transparencies in a process approximating facsimile reproduction (http://www.nga.gov/collection/collect.htm).
2. Different overall views of each object
Every image of an object, even the most accurate photograph, presents only one type of information about the object. Therefore, it is often desirable to have several different overall views of an object, taken under different lighting conditions, with different lenses and from different angles. To some extent this is true even of two-dimensional objects, where the impasto and reflectivity of paintings is often eliminated by professional photographs taken for publication, and differences among fresco, gouache, and oil painting nearly eliminated. The need for multiple views is glaringly true of sculpture. I cannot claim to have conducted a thorough search, but I have not yet seen any museum site with more than one overall view of any sculpture. Perseus includes overall views of many Greek vases taken from several different angles (http://www.perseus.tufts.edu/cgi-bin/browser?object=vase) and plans to introduce QuickTime images of sculpture on its Roman site.
3. Details of single objects
In addition to different overall views of each object, whether of a painting, sculpture or building, we need images of all significant details. On its impressively comprehensive new web site (http://www.nga.gov/home.htm), the National Gallery of Art, Washington, has made available unusually large numbers of details for select paintings. Since these are all made from scans of their original 8 x 10 inch transparencies, details for small paintings retain more detail than those for large ones. For The Feast of the Gods, by Giovanni Bellini and Titian, there are no fewer than twenty-six details (fig.2), all of which can be opened to medium size images. The painting, one of the Gallery's prime masterpieces, rewards careful looking. To study or simply to enjoy the incomparable artistry of this painting or its recent cleaning, one longs for details of even smaller areas, which is why in the gallery most viewers move in to have a closer look. What we are missing when such details are not available is convincingly demonstrated where they are: e.g., Giovanni Bellini, Saint Jerome Reading (http://www.nga.gov/cgi-bin/pdimage?361+0). This provision of numerous details for many paintings is one of the most significant contributions of the NGA site. Because the site is still under development, details for large portions of the collection, including British and French paintings, 20th Century, and Sculpture, are still in the process of being added.
In their definitive article, "Toward on-line, worldwide access to Vatican Library materials," (IBM's Journal of Research and Development ; Vol.40, No.2; http://www.almaden.ibm.com/journal/rd/mintz/mintzer.html), the thirteen joint authors describe the aims of the Vatican Library project: "A core goal of the project is to provide access via the Internet to some of the Library's most valuable manuscripts, printed books, and other sources to a scholarly community around the world. A multinational, multidisciplinary team is addressing the technical challenges raised by that goal, including . . . . Capture of images of the materials with faithful color and sufficient detail to support scholarly study. . . . [and] Development of tools to enable scholars to scrutinize images of manuscripts." They also describe their Scholar's Interface Application program (SIA), which "enables user-scholars to locate images of interest, download them to their workstations, display them, and magnify portions of them to view their details, with accurate color."
In a revealing paragraph, the authors write:
This result should not have been surprising to anyone who uses such images for research and it is a bit distressing, though common, that this model project should recognize this need only for material including small inscriptions. Also, it would be desirable for future projects requiring handling of such rare material to scan at least at the Pro Photo CD resolution of 4096 x 6144 pixels per inch for the archival copy.
A stunning demonstration of the importance of details was recently put on the web site of carte de visite photographs from the private collection of Wm. B. Becker (http://www.Photoarts.com/gallery/becker/ - link now dead). Social historians and museum web site directors take note. For an image of "The Funeral of Abraham Lincoln at the Ohio State House Columbus--1865," the site first provides a thumbnail, then a 4 x 6 1/2 inch image (already larger than the standard 2 1/2 x 4 inch size of the carte de viste) with the question "What's that building across the street, facing the seat of Ohio's government? It's not a law firm! Click here for a closer look." Clicking we see an enlarged detail (Figure 3 <www.wwnet.com/wmb> [June 1999 - link changed to http://www.wwnet.com/~wmb/]) noting "A closer look shows Lincoln's name on the catafalque; a sign proclaiming "Billiard Room" on the building facing the State House; a clock--or perhaps a watchmaker's trade sign--reading twelve minutes to three; and next to the clock, ghost figures of boys who failed to stay still during the entire exposure." The message asks us to "Click on magnifying glass for a closer look at the mysterious head-gear of the grooms holding the horses hitched to the catafalque." Clicking on the magnifier, we see an even more enlarged detail with the explanation "Several men at the left of the image are obviously officials on this solemn public occasion. Here, a closer look at two of them. Both have white (or at least very light-colored) placards perched on their top-hats. Since black was the color of mourning, it is likely that these cards were meant to communicate some other message. Are they a badge of office? Directions to the public? Unfortunately, the resolution of this image does not permit us to read what--if anything--was written on them."
Aimed at specialized laboratory-type research, the highest quality images of very small details are usually the result of photography in conservation laboratories. Most notable here has been the work of the National Gallery, London. A few of these images have been made available on the Internet, as demonstrations, through a cooperative project with Kirk Martinez, at that time with Birkbeck College, University of London (http://www.ecs.soton.ac.uk/~km/vips.html).
4. Images of all parts of a multiple object
For research on objects with multiple parts, images of only the most famous parts is not adequate. For medieval manuscripts, we need images of the volume as a three-dimensional object, of the covers front and back, and of every page, with details of the illuminations.
The prototype of the on-line version of the first volume of Diderot and d'Alembert's Encyclopédie, ou Dictionnaire raisonné des sciences, des métiers et des arts (http://tuna.uchicago.edu/homes/mark/ENC_DEMO/) was recently put on the web by ARTFL, the Project for American and French Research on the Treasury of the French Language, a cooperative project of the Institut National de la Langue Française (INaLF) of the Centre National de la Recherche Scientifique (CNRS) and the Divisions of the Humanities and Social Sciences of the University of Chicago. The site informs us that it will eventually include all seventeen volumes of text and eleven volumes of plates that composed the first edition of this famous encyclopedia, published between 1751 and 1772.
5. The ability to move around sculpture and through architecture, moving in to examine whatever details may interest us, at whatever pace one wishes
Various attempts have been made to provide images that allow viewers to control the process of viewing three-dimensional objects, approximating as nearly as possible the natural process by which we study artifacts and buildings. To date, however, the technical complexity of these undertaking has produced only primitive prototypes more interesting for their future promise than for their current usefulness. Few of these have found their way onto the web. Notable are some of the manipulations recently made available on the innovative Perseus Project site which Kirk Alexander and Marilyn Lavin created at Princeton (http://mondrian.Princeton.EDU/pierowww) and the plan to introduce QuickTime views of sculpture on the Roman Perseus.
6. Images of all or nearly all of a collection
It is immensely useful to have all or nearly all of a collection online. I want to stress here not just the number of images but the complete coverage of a collection, which allows one to use the images as a comprehensive search tool, and the rich context that allows one to study secondary works, including works by less famous artists, by an artist's pupils, copies after famous paintings, even imitations and fakes. It is not generally recognized by the public that museums house large numbers of objects that are never displayed but nevertheless serve as an indispensable resource for research, without which our knowledge of the famous objects on display would be limited and our knowledge of culture in general severely distorted.
The most notable site here is of course the "The Thinker" site of the Fine Arts Museums of San Francisco. The "largest searchable art imagebase in the world", the site already contains 65,000 works of art, 50% of their collections, and the museums are committed to making their entire collections available online. To their images of 60,000 works of art on paper, the museums have recently added images of 2500 ceramics and glass and 750 European paintings (Art Imagebase: http://www.thinker.org/imagebase/index.html). This comprehensiveness is invaluable for research. If we search for Albrecht Dürer we find 250 images, including anonymous copies after and copies by lesser known contemporaries. The site includes a brief explanation of how the imagebase is used for curatorial research by Karin Breuer, Associate Curator, Achenbach Foundation of Graphic Arts at the Legion of Honor (http://www.thinker.org/imagebase/tour/1.html). Bob Futernick, the chief architect of the site, tells me that almost none of the 60,000 works on paper had been photographed in any satisfactory form previous to the idea of digitizing the collection. In this way, as in so many others, the potential of the Internet is encouraging museums to engage in systematic documentation that we should have attended to long ago.
7. A large number of images, with details, in a distinct subject area
The most natural foundation for image databases have been already existent museum collections, available in-house. More challenging are concentrations of images in subject areas, drawn from various collections and sites. The most famous and impressive of these is Perseus, an extensive digital library for ancient Greece (http://www.perseus.tufts.edu/). Images are grouped under four categories (sites, architecture, coins, vases, and sculpture) and include plans of archeological sites, aerial views, scale models, ground plans, elevations, other types of diagrams, and thousands of photographs. These vary a good deal in type and quality, but even the best photographic images, scanned from excellent original 35mm slides by Maria Daniels, Visual Collections Curator, have been put on at low resolution and are in most cases not very satisfactory for research. However, even at low resolution, small objects such as coins and details of vases are useful. The most remarkable feature of the Perseus image archive is the inclusion of 15 or more images for 30% of the vases and an astonishing 109 and 110 for two vases at Harvard (for this, see the images of vases listed under: "Cambridge: Harvard," "Museum of Art: RISD," "Tampa," "Toledo," "University of Mississippi," and "University of Pennsylvania" (http://www.perseus.tufts.edu/cgi-bin/browser?object=Vase&field;=Collection). Even at low resolution, these details provide an exceptional basis for the study of the vases represented.
The Perseus Project has recently received a grant from the Teaching with Technology Program of the National Endowment for the Humanities for the development of a Roman Perseus, for which QuickTime images are being explore (http://www.perseus.tufts.edu/neh.ann.html).
In a few cases, commercial companies have put thumbnails of their digital images on line. Especially where these include major concentrations for individual artists, collections, or archeological sites, these constitute a valuable research resource. The most impressive of these sites, Saskia Ltd., has available on its web site thumbnails with identifications for over two thousand images, about two-thirds of which are unique details (<http://www.saskia.com/digital.html> [June 1999 - link changed to http://www.saskia.com/query/Sets.asp?type=d ])). As with some other companies, higher resolution images can be licensed by educational institutions through individual site license agreements.
8. Many different types of visual material
Every image of an object, even the most "accurate" photograph, presents only one type of information. For research, one needs multiple images of different types. Depending on the object and nature of the research, one may need not only many different photographs with details of the object, but also ground plans, cross-sections, other types of diagrams, maps, videos, motion sequences, sometimes even x-radiographs and infrared reflectograms, etc.
In addition to the Perseus Project described above, an exemplary site integrating diverse types of visual images for a focused topic was mounted by the Sackler Gallery, in conjunction with their exhibition, "Preserving Ancient Statues from Jordan" (http://www.si.edu/Asia/jordan/jor_MM.htm - link now dead)
The most impressive example of what can be done when a scholar has accumulated research images of many types and related information for indepth study of a major monument can be seen on Columbia University's Amiens Cathedral web site, based on Stephen Murray's research and photography (http://www.learn.columbia.edu/Amiens.html). This is one of the pioneering projects of Columbia's new Media Center for Art History, directed by Murray and Maurice Luker. On this site, one can study not only color and black-white photographs of many parts of the cathedral, including photographic sequences (progressively closer views, identical photographs at different times of day, from changing angles, etc.) and numerous detail (with tiny diagrams indicating the locations of the detail and angle from which photographed), but also drawings and diagrams of elevations, cross sections, and ground plans (Figure 4 <www.learn.columbia.edu/Drawings/index.html> [June 1999 - link changed to http://www.learn.columbia.edu/Mcahweb/Drawings/Diagrams/dia01.html ), axonometric drawings of the construction sequence, computer graphic simulations of unusual views, and QuickTime panoramas of exterior and interior sequences.
9. Choice of multiple sizes and resolutions
One of the great advantages of computers over hardcopy is the opportunity of making information available in multiple formats, accessible by each person according to desired use and equipment available. Thus, we regularly look at thumbnail images before deciding whether to open larger files. Multiple formats support high quality images because large image files can be made available without imposing themselves on those who do not want high resolution images or do not have the equipment necessary to use them efficiently.
In December 1996, a prototype demonstration database of the California Heritage Digital Image Access Project went on line (http://sunsite.Berkeley.EDU/CalHeritage/), with first-rate images of photographs from the Bancroft Library at the University of California, Berkeley. This impressive site could serve as a model for many of the characteristics of digital images needed for research. Based on the Bancroft's extensive collection, the site produces 858 hits on a search of "Carleton E. Watkins" alone, grouped under 18 categories, many of these hits pointing to several hundred photographs already available as thumbnail, medium, and high resolution images (fig.5), which, in many cases, would fill a 30 inch screen. Recognizing the diversity of users, all museum sites should make available, in addition to thumb nails, at least two additional images at increasingly high resolution and size.
Important Information about Images
10. Multiple search fields for locating images
The Fogg Museum's "Sargent at Harvard" site is "a structured database for over four hundred drawings, paintings, sculptures, and related materials" (http://www.artmuseums.harvard.edu/sargent/archive.html). It provides five keyword search fields (subject, medium/support, project, location, and watermark and artists tools) and opportunity for full text search using a string of whatever words one chooses in order to narrow the search. The site includes also an exemplary list of "Related Collections" of Sargent material, providing not only links to other museum sites but also brief descriptions of their Sargent holdings.
11. The ability to search by visual characteristics
A number of search engines have recently come on the market which allow one to search by color, texture and shape. Although this may have usefulness for art students in design courses and for an occasional design-related curatorial search, it is difficult to see how they are adequate for the types of searches engaged in by art historians and others searching for less design-specific characteristics. Nevertheless, it is useful to explore the potential of even these primitive examples, for which see the Fine Arts Museums of San Francisco's joint Query by Image Content (QBIC) project with IBM (http://126.96.36.199/cgi-bin/QbicStable [June 1999 - link now dead]) also IBM's product description (http://www.software.ibm.com/is/dig-lib/; Product Information; Search and Access; QBIC).
Complicated and multifaceted though it may be, what is needed is the ability to select some area of an image and to ask the search engine to find other similar details in other images. Projects exploring this possibility are underway, but I have not seen anything yet developed which would actually be of use for this type of content search. Most promising is the Morelli Project, which is exploring the possibility of searching not only to match overall compositions but also by details or motifs (http://www.hart.bbk.ac.uk/~cath/morel.htm [June 1999 - link now dead]).
12. The timely provision of images
Newspapers have long taken advantage of the ability to make images available within hours of events or discoveries, and with television we are sometimes shown events as they occur. But these subjects are limited to the popular press, and the number of images and their quality is usually insufficient for anything more than an introduction to the subject portrayed. Like television, the Internet offers the opportunity of making images available quickly with distribution to an international audience. But unlike television, many more images can be made available, of higher quality than those in other media, and in areas that would not find their way into the popular press.
With art thefts, speedy notification of dealers and customs agents worldwide is a key to recovery. Recently, the London office of the international Art Loss Register (ALR), a permanent computerized database of stolen and missing works of art, antiques and valuables, has mounted a web site including its quarterly newsletter which includes images and descriptions of recently lost and stolen art objects (http://www.artloss.com/main.htm). The site includes images of objects that have recently been recovered with the aid of ALR. Understandably, the images are rather poor and would not serve to distinguish an original from a high quality copy, but even low quality images are vastly more useful than mere text descriptions, and this format would provide for high quality images when available.
13. Multiple types of information about images, closely integrated
The "Sargent at Harvard" site is also a model of extensive content systematically organized. Information for each drawing or watercolor may include as many as fifteen categories (accession number, title, medium, date, dimensions, donors, inscriptions, signatures, notes about the object, provenance, other accession numbers, bibliography, exhibition history, sources of information, and keywords). This impressive site, which approximates a scholarly catalogue raisonné with the added benefits of computer search, etc., is the result of a year of intensive work by Melinda Linderer, Lynn and Philip A. Straus Intern in the Drawing Department for the academic year 1995-96, in consultation with Sargent experts, and by Lee Mandell, Computer Support Manager, Harvard University Art Museums. It is indicative of the professional hours required to bring museum files and photography up-to-date for a research site that there should still be "Sorry no image available" notices for so many objects, though there seems no reason why these should be the most intrusive aspects of the design.
If we check the "literature" section for different paintings in even the best museum sites, we discover that the listings are in some cases extensive and nearly complete, in other cases rudimentary. Anyone who has carried out detailed research on paintings in the collection of major museums will understand the problem. To locate all of the material on a single painting, it is necessary to check, at a minimum, the registrar's files, curatorial files, and conservation files, any of which may temporarily be in a different museum office, and the most current information may reside only in the heads of the professional staff. For photographs and slides one must look in different locations and for laboratory photographs elsewhere again.
It is no wonder that museum sites rarely approach the status of published catalogues raisonné, but they have the immense advantage of making information available widely without the ten year delay common for the research and publication of major catalogues. And they have the potential of eventually persuading museums to get their own files in order and to coordinate valuable information among their many departments.
14. A complete image trail
In order to understand the evidence contained in any image, what is reliable, what not, we need to have as complete and detailed an image trail as possible. We need not only to have a description of the object being imaged, but also to know how the original photograph of it was made: what type of lens and film were used and how exposed? Or was the image made by direct digital capture and if so how? We need to know how the image was scanned and how manipulated at each stage. Few things are as important for the reliable use of digital images as the establishment of a widely agreed upon template for recording such information as a regular expectation for research photographs. Some watermarks under development would facilitate recording at least part of this image trail. Of course, this information is more crucial for certain types of objects and for some types of research than for others. To the extent that the visual appearance of the objects is important, we need to know as clearly as possible the relationship between what we are seeing when we look at the digital image and what we would see were we looking at the object itself.
For research purposes it would be desirable for color transparencies of museum objects to include color bar, gray scale, and ruler, for these to be retained in the digital images, but easily hidden or displayed at the click of a mouse.
A few commercial companies lead the way in providing such information. Accompanying its Kodak Photo CDs, Luna Imaging (http://www.luna-imaging.com) provides a disk of management data, including, for each slide or photographs scanned: the size and type of photograph or film scanned, format of the master digital image (e.g., 16 base), format of any derivative digital images generated from the master image, and, for both master and derivative images, any image capture variables during scanning, resolution, bit depth, and date of scanning.
IBM's exemplary Research site (http://www.research.ibm.com/image_apps/) provides wonderfully detailed information with explanatory images for their joint project with the Vatican Library, describing the process by which they capture and display digital images, including the six-steps used to prepare images for the Internet (fig.6), and a seven step description of the "flow of images through [their] system."
15. Information on how to order or site license high quality images
Not surprisingly, instructions for ordering high quality photographs or for licensing larger, higher resolution digital images than those available on the web are readily available on commercial sites. Among museum sites, this information is rarely presented as fully as by the Detroit Institute of Arts, which has a wonderfully clear and instructive description of publication rights, lists of different conditions for ordering print media and electronic media, procedures for ordering different types of images, methods of payment, and a printable orderform to be mailed or faxed to the Institute (<www.dia.org/main/pageIE.html> [June 1999 - link changed to http://www.diamondial.org/]). The amount of detail may frustrate some viewers, but to read the instructions carefully is to be educated in the complexities of contemporary images and their proper use.
16. The ability to add ones own images to a site, for sharing and peer review
For interchange at all levels, it is important to share images and to exchange information about and interpretations of images. This is especially important for advanced research where the images, information and interpretation may not exist in any hardcopy source. The Dermatology Online Atlas at the School of Medicine of Friedrich-Alexander Universität Erlangen-Nürnberg, provides not only large numbers of images on a well-defined subject, reasonably good images of details, but also allows other researchers to add their own images for peer review and discussion (<www.uni-erlangen.de/docs/derma/bilddb/db.htm> [June 1999 - link changed to http://www.uni-erlangen.de/RRZE/ausbild/schrift/mitteil/mb70/]). In Dec. 1994, Andreas Bittorf added to the DOIA the first version of a detailed account of "How to put your own images on the web", including exemplary discussions of image compression and storage, problems encountered with digital images, etc. (<www.uni-erlangen.de/docs/derma/bilddb/how_to/how_to.htm> [June 1999 - link changed to http://www.uni-erlangen.de/RRZE/ausbild/schrift/mitteil/mb70/]).
This paper does not pretend to be a comprehensive survey of museum sites that are leading the way in the development of useful images; many other sites would deserve recognition in such a survey. Instead, this paper attempts to identify characteristics of images of museum objects that are important for research and to illustrate these with exemplary pages from a few sites.
The paper deals only with images on the world wide web. We should note, however, that there are additional images, sometimes of superior quality, available to subscribers of Internet services or to a limited number of institutions jointly engaged in research projects. As the Internet continues to explode, it seems almost inevitable that additional research sites, not freely available to the public, will continue to be developed.
Digital images currently available on the Internet must be considered primitive prototypes for images that will be developed in the future. All the more reason to make clear what directions these developments should take if they are to serve the needs of research (see Rhyne "Computer Images for Research, Teaching, and Publication in Art History and Related Disciplines," Visual Resources, Vol. XII, No. 1, 1996, pp. 19-51; a portion of which was published as a special report under the same title by the Commission on Preservation and Access, 1996). As digital image technology develops and as the potential of these images for important new types of research and discovery is recognized, many objects will need to be rephotographed with new equipment and specialized lighting, often under laboratory conditions. Museums, which house so many key objects for the study of the manmade and natural worlds, bear a special responsibility and opportunity for making these images available. The potential benefits to society are immense,
_ _ _ _ _
Home pages for museums, university departments and commercial companies whose web sites or individual pages are cited in this article
Art Loss Register (New York and London)
Arthur M. Sackler Gallery, Freer Gallery of Art, Smithsonian
Institution (Washington, DC)
Becker, Wm. B.: "Photography's Beginnings"
Birkbeck College, University of London: History of Art
Columbia University: Department of Art History and Archaeology
(New York, NY)
Detroit Institute of Arts (Detroit, MI)
Fine Arts Museums of San Francisco: M.H. de Young Memorial
Museum and California Palace of the Legion of Honor (San Francisco,
Friedrich-Alexander Universität Erlangen-Nürnberg:
Department of Dermatology, Medical School (Erlangen)
Harvard University Art Museums (Cambridge, MA)
Institut National de la Langue Française (INaLF)
of the Centre National de la Recherche Scientifique (CNRS), Paris
International Business Machines
LUNA Imaging (Venice, CA)
National Gallery of Art (Washington, DC)
Princeton University: Department of Art and Archaeology
Saskia, Ltd. (Portland, OR)
Spencer Museum of Art, University of Kansas (Lawrence,
Tufts University: Classics Department (Medford, MA)
University of California, Berkeley: Departments (Berkeley,
University of Chicago: Divisions of the Humanities (Chicago,