The Spine and Its Tingle: ARTstor Digital Library: A Case Study (revisited)

Yet again, I have found some of my grad school assignments kicking around the Internet, due to the oversight of a former professor (or his grad assistant), who forgot to remove them at the end of the semester. Thought I'd post one of them here since it is already out there anyway, caught in the sticky clutches of the vast InterWebs. This one is a case study about ARTstor*. I've removed the original images since I no longer have access to ARTstor. If you have access to ARTstor via your university's nifty academic library, DO please check it out. It is a great resource.

*one caveat: this was written in early 2011. ARTstor has likely undergone changes and improvements in the last two years.

ARTstor Digital Library: A Case Study

Introduction

ARTstor is a non-profit digital library with a collection of over one million images, with accompanying metadata, in the arts, architecture, social sciences, and humanities (Golderman and Connolly 2011, 136). Using the criteria outlined in the article “Search Features of Digital Libraries” as a guide, I will evaluate the design and search features of this digital library. I will be examining the library’s background—its users, purpose, and collection development strategy—interface design, browsing function, search mechanism, and I will be comparing it to other digital libraries. In evaluating the library’s search mechanism, I will be looking for certain features and advanced search options, such as the ability to use boolean searching, field searching, results rankings, controlled vocabulary, language translation, date/range searching, related terms, and refining of initial search.

Background

ARTstor was born out of the need to address the “high costs, a lack of standards, and daunting intellectual property concerns” associated with digitalization (“Mission and History”). It was founded by the Andrew W. Mellon Foundation in 2001 (Marmor 2006, 95). The beginnings of this digital library are recounted:

The Mellon Foundation’s goals for ARTstor were similarly ambitious and global from the outset. The first phase of the project involved partners in China, France, the UK, and the US to build the Mellon International Dunhuang Archive—a digital repository of extraordinary, high-resolution images from 40 cave grottoes in the Gobi desert (one of the largest Buddhist art sites in the world), along with images of silk banners and manuscripts from the caves brought to Western Europe by English and French explorers at the turn of the 20^th century. While the project presented many challenges, it was a resounding success, demonstrating the ability to preserve objects in peril, to reunite for study in one location works of art previously scattered around the world and to join people across geographical and cultural divides. (“Mission and History”).

In January of 2004, ARTstor became an independent, not-for-profit organization (Marmor 2006, 95).

Purpose

ARTstor’s mission is “to enhance scholarship, teaching, and learning in the arts and associated fields,” and lists the following as its primary goals:

To assemble image collections from across many time periods and cultures that will, in the aggregate, have sufficient depth, breadth, and coherence to support a wide range of educational and scholarly activities;
To create an organized, central, and reliable digital resource that supports noncommercial use of images for research, teaching and learning; and
To work with the arts and educational communities to develop collective solutions to the complex challenges that are an inescapable part of working in a changing digital environment. (“Mission and History”).

Collection Development

The images in the collections are contributed by many different repositories—museums, libraries, photo archives, scholars, artists and artists' estates, and photographers. There are over 150 collections in the arts, architecture, humanities, and social sciences including Magnum Photos, the Illustrated Bartsch, The Metropolitan Museum of Art, the Museum of Modern Art, the Berlin State Museums, Erich Lessing Culture and Fine Arts Archives, Scala Archives, Carnegie Arts of the United States, the Natural History Museum in London, the Bodleian Library at Oxford University, the Mellon International Dunhuang Archive, the Asian Art Photographic Distribution (AAPD) from the University of Michigan, contemporary art and architecture from ART on FILE, the Ezra Stoller Archive of Modern Architecture, just to name a few. In building its collection, ARTstor

“aims to include visual materials that are useful for both teaching and research in a variety of subjects from prehistory to the present. We seek the ongoing participation of the community to broaden and deepen our coverage in western and non-western cultural heritage materials, and to expand well beyond the history of art into other areas in the humanities and social sciences. ARTstor welcomes the opportunity to grow the range of images in ARTstor through new partnerships with unique and important collections,” (“Collections”).

Users

ARTstor is used by educators, scholars, curators, and students of universities, colleges, museums, libraries, and K-12 schools worldwide (“Overview: What is ARTstor?”). In order to access the collection, a subscription is required. Users must be affiliated with a participating non-profit institution in order to access ARTstor’s collections as it is not licensed to individual users at this time (“Who Can Access ARTstor?”). According to the site, “…fees help to sustain the ongoing maintenance and development of the collections and software for the Digital Library. Institutions typically enter a site-wide license with ARTstor through IP authentication. Site licenses to ARTstor allow unlimited number of simultaneous users to access the Digital Library, both onsite and remotely via username/password” (“Overview: What is ARTstor?”).

According to a January 2011 article from Library Journal, costs vary by institution: “Institutions pay two fees: the Archive Captital Fee (ACF) and an Annual Access Fee (AAF). Pricing is established based on the Carnegie Classification with the range from Community College level ($1150 ACF fee) to Very Large I ($46,000 ACF fee),” (Golderman and Connolly 2011, 136).

Interface Design and Features

Initial Screen

The initial page functions as a gateway to the collection and contains announcements about upcoming events and a brief sampling of images. Tabs near the top of the page include “Collections & Services”, “Interested in Access?”, “Using ARTstor”, and “News.”

“Collections & Services” provides an overview that gives ARTstor’s background and how it can be utilized. There is information about how the overall collection was built and continues to develop, the names and sizes of individual collections, subject guides, an FAQ page, and other miscellaneous information. There is also a section called “Metasearch into ARTstor” which I will cover in more detail later.

“Interested in Access?” provides information about who can use and access ARTstor, becoming a participating institution, current participating institutions, and trial access. “Using ARTstor” walks the user through the key features of ARTstor, how to search, add images, etc.

The user enters ARTstor’s digital library via a link from the initial screen. Once “inside” the library, there are clickable tabs that allow the user to view the history of the library, find records/images, organize and share retrieved images, and adjust the display. There is also a “tools” tab that allows the user to save and export citations, download an offsite presentation tool, and create a profile.

This page also contains a search box for conducting a basic search with a link to an “advanced search” page beneath it. There is also a browse feature on this page that allows the user to browse the collection by geographical location, classification (painting, performing arts, photography, etc.), and collection. There is also a “featured groups” browse option. This option presents the user with a choice of interdisciplinary themes. Each group listed contains “iconic images mixed with other selections that are meant to trigger new ideas and provoke deeper research” (“Browse”). These groups are not comprehensive but are designed to get the user started on a deeper exploration of certain aspects of the collection.

Browsing Function

Users may browse by geography, collection, classification, and featured groups. ARTstor provides subject guides, which can be useful for browsing purposes or for the user who is new to using ARTstor. These “how-to guides” cover 23 subjects and topics covered include American studies, design, fashion and costume, music history, and photography. The guides come in PDF format (Golderman and Connolly, 2011, 136).

Search Mechanism

There is a basic keyword search and an advanced search. With an advanced search, the user can search specifically for title and/or creator, date/date range, and specific geographical locations, classifications, and collections can be selected to narrow the search—up to 5 filters may be selected. Once results are retrieved, the user is able to search within the results. The user can also sort results by title, creator, date, or relevance. By default, results are sorted by relevance.

ARTstor had, for years, lacked a controlled vocabulary. An article from 2006, written by ARTstor’s Assistant Director for Library Relations, explains why there was no controlled vocabulary:

Because of the diversity of the image collections and the sources from which they come, there is, at present, no single controlled vocabulary that works across all of ARTstor. Since our contributors had cataloged their collections in different ways, targeted to their individual needs, ARTstor has had to bring together rather disparate data. Using thesauri and standards already available in the community, ARTstor has been able to address the heterogeneity of the data between collections by enhancing it so that the values for certain types of information are consistent across collections,” (Rockenbach, 3).

According to ARTstor’s website, they now use the Getty Research Institute's Union List of Artist Names (ULAN) to match artists' names with an authoritative creator record. These matches allow links to be established between the source creator name and the ULAN creator record. There are plans to extend the matching of source data to external vocabularies into other areas, such as repository names, geographical locations, styles, and periods. They also offer metadata enhancements, image clustering (for duplicate images), and they make associated images in groups available to the user through the application of mathematical analysis (“ARTstor Digital Library Metadata Policy”).

ARTstor has developed a search interface called “ARTstor XML Gateway” that supports cross-database searching, meaning ARTstor can be searched from external sources. It is described thusly:

The basis for the ARTstor XML Gateway is the Search and Retrieve URL (SRU) Service. The SRU Service is a web services-based protocol for querying databases and returning search results. Currently under review as a part of the National Information Standards Organization’s Metasearch Initiative, SRU Incorporates elements of the Z29.50 protocal for Information Retrieval. The ARTstor XML Gateway was designed using SRW/U Open Source Software, developed by Online Computer Library Center Search, as a basis for implementation (“Metasearch Into ARTstor”).

Constructing a Query

To test ARTstor’s indexing and searching functions, I constructed several queries on the artist Dorothea Tanning. I wanted to find an image of the painting Jeux d’Enfants. I tried the basic keyword search first and searched multiple collections. I typed in “Jeux d’Enfants” and two results were returned: Children’s Game/Jeux d’Enfants by Dominque Vivant Denan and the piece by Tanning.

I wanted to find out how many images of the works of Tanning ARTstor has in its collection. I did another basic keyword search for “Dorothea Tanning” and retrieved 42 results, including paintings, sculpture, and photos. There were many duplicate images. I then tried the search again, using “Tanning, Dorothea” to see if reversing the order of the names would make a difference in the number of results returned and it did not.

I then initiated some searches using the advanced search options. I did a “creator only” search for Tanning Dorothea, and retrieved 39 results. In the next search, instead of selecting “creator only” I selected “in any field” and retrieved the same 42 results returned in the basic keyword search.

I noticed that several of the images returned in my searches were photographs of Tanning and the artist Max Ernst. I decided to construct a query to retrieve only these photographs. I searched for Dorothea Tanning AND Max Ernst and retrieved 37 results, including the works of Tanning and photos of the two together. In my next search, I used the same terms except I substituted “OR” for “AND” and retrieved 807 results. Some of the retrieved images did not appear to be related to Tanning or Ernst. I then searched for Dorothea Tanning NOT Max Ernst to see which artworks would be retrieved. This search returned 5 results, which were all works by Tanning but most of the works by her available in ARTstor were not retrieved.

I decided to conduct a sort of general search to see how well the Surrealist movement (of which Tanning and Ernst were a part) is covered by ARTstor. I also wanted to see how many results would be returned with a more general search. I did an advanced search using the date range 1930 CE to 1950 CE, limited the geographical regions to the United States, United Kingdom, and Europe, and the classification “paintings.” This returned over 1,000 results. I then removed the United Kingdom as a filter, and adjusted the date range to 1939 CE-1950 CE (what’s known as the “golden era” of the Surrealist movement) and still retrieved over 1,000 results. Removing Europe as a filter also produced over 1,000 results. While you can search within these results, it is necessary for the user to know a bit about what he/she is looking for in order to make good use of this function.

How Images May Be Used

Image results are thumbnail size and can be double-clicked for a larger view. View can be zoomed and panned. Data about the image, as well as metadata, is available. Users can view information about an art piece itself—creator, title, work type, date, material, measurements, the repository it is housed in, ARTstor collection, and ID number—and view the file properties of the image, including image ID, download size, collection, accessibility, format, file name, width, height, and image URL.

Users can print, save, create, and share images, organize them into folders, download them to KeyNote, PowerPoint, and ARTstor’s Offline Image Viewer, save, email, export and print citations for images, and upload and manage personal images and sound files.

Comparison to Similar Digital Libraries

If I were doing research specifically on Dorothea Tanning, I would be more likely to use the image library at www.dorotheatanning.org. This website is free and publically available and it is dedicated to Tanning. While it obviously does not offer all the features that ARTstor offers, it is a more comprehensive collection of the artist’s work. The images are arranged in the order in which the original artworks were created, to show the progression of Tanning’s work. This is a growing archive and includes biographical information about Tanning, information about her collections, exhibitions, and information about writings by and about Tanning. This is a relatively new venture and more features will probably be added in time.

In comparing its search and browse features with those of the New York Public Library Digital Library ( http://digitalgallery.nypl.org/nypldigital/index.cfm ), I see similarities but also some major differences. On the NYPL Digital Library website, the user can browse subjects, including names, library divisions, arts and literature, cities and buildings, culture and society, history and geography, industry and technology, nature and science, and printing and graphics. Advanced searching options include field searching, boolean AND and OR, wildcard characters, limiting by date/range, searching by image ID and catalog number, and you can enter terms that you want avoided in the search. Being able to search by image ID and catalog number is an advantage. I attempted to search file properties fields in ARTstor using the information ARTstor provided for some of the images I viewed (identification numbers, specifically) and was not able to retrieve the files I was searching for.

Finding nothing for Dorothea Tanning in the NYPL Digital Library with a basic search, I tried the author and illustrator Clare Turlay Newberry and retrieved one result. By double clicking the image, I was able to view it on a larger scale.While I was able to increase the size, I was not able to view the image in the detail that ARTstor provides. The NYPL Digital Library does provide similar information about their images but they are more tailored to fit the New York Public Library’s primary role as a “traditional library.” Information provided includes image title, item/page/plate number, source, source description, location, catalog call number, digital ID, record ID, digital item published date, and rights notice. They also provide browsing options within the search results related to the item you are viewing. For example, the Clare Turlay Newberry image I retrieved was from the dustjacket of a book called Babette. I was given the option to browse book jackets and books on cats (because this particular book is about a cat), book jackets from the era the book was written, and other books in the same division. Being able to search by subject is something that is difficult to do in ARTstor and I have to wonder how useful it is in the NYPL Digital Library. For books, it would be useful. But for images, and images of art specifically, a user may have limited success.

Unlike ARTstor, this website is dealing with books, manuscripts, dust jackets, posters, etc. The nature of their collection is a bit different from ARTstor so it’s hard to compare the content but I find it to be an excellent research companion to ARTstor and it has some advantages over ARTstor. It’s a free resource, easy to access, and, at least in my opinion, the layout of the site is much simpler and easier to navigate. However, I believe ARTstor would be much more useful to educators because ARTstor offers many tools and features that are specifically designed to aid in teaching and research.

Conclusion

ARTstor is still adding to its collection. It’s fair to assume it will expand exponentially in the future as efforts to digitize are stepped up. Because of its many features, ARTstor can be an extremely useful tool for scholars, educators, and students. Even the casual user would find it to be a great way to explore the art of specific regions and eras and discover new artists. The fact that the images can be made available (when they otherwise would be difficult to see since not everyone is able to travel to distant museums and art galleries) is a major accomplishment in the move toward digitization.

There may be better sources out there if you are doing research on specific artists, however. In looking for information on Dorothea Tanning, specifically images of her artwork, I was able to find a better resource and using it did not require me to be affiliated with a subscribing institution. The cost of ARTstor, and the fact that individuals not affiliated with subscribing institutions cannot access it, means that it can only be useful to certain groups of users.

As pointed out in “Using ARTstor” by Mirela Roncevic, ARTstor is usually lauded for its accessibility but its user-friendliness is in the eye of the individual user: “The more one knows about art, it appears, the more fruitful ARTstor is,” (Roncevic 36). One user pointed out that browsing through results can be challenging if the search is very specific and that it helps to be a seasoned searcher when using ARTstor. ARTstor does offer free training, however, and this is a useful tool for those users who may not be so familiar with art. In trying to learn about all the available features, services, and terms of usage, I found mountains of information on the website. However, trying to navigate through the different screens for this information and trying to relocate the links later on was difficult because of the labyrinthine way in which it is organized. The Roncevic article also points out that the cost of ARTstor may be a bit pricey for some institutions. Databases aren’t cheap and this is not likely to change anytime soon. But ARTstor has the potential to expand in so many ways, and has already done so in the decade it has existed, that I believe it is a worthy investment for any research institution with a humanities focus.

Bibliography

ARTstor. “ARTstor Digital Library Metadata Policy.” ARTstor.

http://www.artstor.org.libezp.lib.lsu.edu/our-organization/o-html/standards-data.shtml (accessed February 8, 2011).

ARTstor. “Browse.” ARTstor.

http://library.artstor.org.libezp.lib.lsu.edu/library/welcome.html#2|270|Featured20Groups (access February 7, 2011).

ARTstor. “Collections.” ARTstor.

http://www.artstor.org.libezp.lib.lsu.edu/what-is-artstor/w-html/collections.shtml (access February 7, 2011).

ARTstor. “Metasearch Into ARTstor.” ARTstor.

http://www.artstor.org.libezp.lib.lsu.edu/what-is-artstor/w-html/features-and-tools-metasearch.shtml (Accessed February 7, 2011).

ARTstor. “Mission and History.” ARTstor.

http://www.artstor.org.libezp.lib.lsu.edu/our-organization/o-html/history-and-mission.shtml (accessed February 7, 2011).

ARTstor. “Overview: What is ARTstor?” ARTstor.