The Interactive Archivist

Tiah Edmunson-Morton
June 19, 2009

Introduction

The Oregon State University Archives has a robust photograph collection, which is reflected in its hearty companion digital collection. These collections document the history of the university, as well as the agriculture and natural resources of the Pacific Northwest. While the physical and online collections are popular, we were interested in finding a new way to showcase our images and interact with users; in particular, we wanted a place to house online image-based tutorials, share pictures from current programming events, and highlight the photographs in the Gerald W. Williams Collection. Flickr, the popular website for managing and sharing pictures online, offered an alternative to the OSU Libraries’ current online image management system, CONTENTdm. The OSU Archives set up a Flickr account (osu.archives) in the summer of 2008; additionally, the Archives joined Flickr Commons (osu.commons) in February 2009.¹ Framed in the context of analyzing user interaction with online image content, this case study will discuss and compare the OSU Archives’ use of CONTENTdm and experiences with Flickr.

Context of the Case Study

In many ways, the OSU Archives is a “traditional” university archives, housing the official and unofficial records that document campus history. These records contain materials generated by faculty, academic departments, administrative offices, students, and campus organizations. We also have a large and rapidly growing photographic collection, which includes over 450,000 images taken by university staff, commercial photographers, and individuals showing campus buildings, people, and the state of Oregon (with a focus on the mid-Willamette Valley and rural areas of the state).

Like many cultural heritage organizations, we wanted to increase the discovery and ease of use for our physical and online collections. However, in an environment where both financial and staffing resources are limited, users are turning to the Google or Yahoo! search engines more than the search box on the Archives’ Web page. As the archival profession is being compelled to reconsider its presumed omniscient authority, and control over archival materials, public access and interaction are a challenge. In many ways, the promise of a non-hierarchical, community-driven folksonomy, which inherently allows for a more personal identification between users and collections, seemed like an avenue for exploration. At the same time, the intrinsically dynamic and reciprocal relationship between an archive and its users, both known and unknown, might simultaneously enhance the dialog with our user communities and challenge us to examine our own professional jurisdiction over how materials are accessed, cataloged, and used.

Before establishing a Flickr account, there were two access points to our photographs: the Archives’ physical collections and the library-wide digital collections site (stored in a CONTENTdm image management system; see “Implementation” section).² Both the physical and online collections were used frequently, with regular research projects utilizing the tangible photographs and download requests for image files. However, we were interested in expanding our user base and reaching a different audience. To quote the Library of Congress report, “For the Common Good: The Library of Congress Flickr Pilot Project“:

One solution worth exploring is to participate directly in existing Web 2.0 communities that offer social networking functionality…Taking collections to where people are already engaged in community conversations might also encourage visits to a library’s Web site where the full wealth of resources are available.³

Flickr was an economical and easy way to reach this broader and more diverse audience; additionally, it allowed for both user tagging and commenting, which served to engage the user community and could potentially provide us with more information about our collections.

The University Archives set up a Flickr pro account in July 2008 to showcase photographs from our programming events, create online image-based tutorials, and upload various historic photographs—without making another Web page. The Libraries had an active account, and we assumed that setting up our own for the Archives, but targeting the same general user base, would result in “success.” We uploaded two types of images: contemporary and historic. Within the “historic” category, we included shots of buildings connected with a historic campus walking tour and linked to a Flickr Map, images of sporting events and assorted ephemeral paraphernalia, and items from the Gerald W. Williams Collection.

The Williams Collection

The OSU Libraries acquired the Gerald W. Williams Collection in 2007. Williams worked for the U.S. Forest Service from 1979 until his retirement in 2005, serving as their national historian in Washington, D.C. from 1998. He was, and is, an avid collector, and his collection reflected this. In addition to Williams’ research library, the Libraries and Archives also acquired his personal papers, which included historic photographs, oral histories, maps, moving images and sound recordings, as well as posters, ephemera, and artifacts pertaining to forestry, environmental history, Native Americans, and geography of the Pacific Northwest. It was a rich collection meant to be explored and used.

The historic photographs collected by Williams included thousands of postcards from the 1900s through 1940s, documenting watersheds, forests, and communities throughout the Pacific Northwest, with a focus on Oregon. Other historic photographs included a large collection of Oregon and California postcards made by photographer Frank Patterson, images of Depression-era Civilian Conservation Corps (CCC) camps and activities, pictures documenting the U.S. Army Spruce Production Division in Oregon and Washington during World War I, logging photos of northwest Oregon taken by John Fletcher Ford, slides of Celilo Falls taken by Williams’ father in September 1956 months before the falls were inundated by The Dalles Dam, large format prints of early twentieth-century forestry scenes in Washington and Oregon by renowned photographers Darius and Clark Kinsey, and images depicting Native Americans in the Pacific Northwest region. Other photographs, taken by Williams, include pictures of national parks and forests in Oregon and Washington, the McKenzie River region of Oregon, and personal vacation photos.

In addition to press releases, promoting the collection to faculty, and professional conference presentations, the Archives wanted a unique venue to increase our audience for the Williams’ Collection photographs. Flickr seemed like a natural fit.

Implementation and Interaction: Flickr and CONTENTdm

What is Flickr?

Flickr is an online photo management and sharing application. It allows for user tagging and commenting, mapping, editing, and sharing, and has an active user community. Characteristically quirky, funny, and idiosyncratic, Flickr believes that their site is “almost certainly the best online photo management and sharing application in the world.”⁴

What are Tags and Clouds?

Tagging and tag clouds are not unique to Flickr. Many social software websites allow users to assign tags, which are essentially a form of personalized metadata. Tagging allows for a personal level of content identification, categorization, and organization. Simply put, tags help people find items and people that have something in common. Beth Kanter, writing for the website TechSoup, says that “tags are keywords that help describe a photo—both the type of image and its contents.”⁵ By tagging your photos in Flickr you provide an easy way to search and find particular photos in your photo stream, as well as connect your photos to other people who have similar interests. Additionally, depending on your privacy settings, other people can tag your photos, and this is what makes it a communal activity!

Flickr was also one of the first sites to use tag clouds, which give users a graphic representation of the most popular tags or keywords assigned to certain images. According to the Wikipedia definition, a tag cloud or word cloud “is a visual depiction of user-generated tags, or simply the word content of a site, used typically to describe the content of web sites.”

Tags are usually single words and are typically listed alphabetically, and the importance of a tag is shown with font size or color. The tags are usually hyperlinks that lead to a collection of items that are associated with a tag.⁶

Because tag clouds are useful visual representations of subject or topic strengths, they can provide an alternative way to mentally access a collection of items.

What is a Folksonomy?

To quote Alfred Hitchcock:

I’m full of fears and I do my best to avoid difficulties and any kind of complications. I like everything around me to be clear as crystal and completely calm. I don’t want clouds overhead. I get a feeling of inner peace from a well-organised desk. When I take a bath, I put everything neatly back in place. You wouldn’t even know I’d been in the bathroom.⁷

Unfortunately for him, any discussions of systems that support user-generated content tend to address the uncertainties implicit in an open-access approach to metadata and content organization. Tradition, hierarchy, chaos, classification, and authority are all words that swirl around talk of taxonomy and folksonomy. Because of its use of tags, Flickr has been cited as a prime example of a folksonomy. Folksonomy, a word formed by combining “folk” and “taxonomy,” is simply a user-generated taxonomy; it is also known as collaborative tagging, social classification, social indexing, and social tagging. Thomas Vander Wal, who is widely credited with coining the term, says that a folksonomy is “the result of personal free tagging of information and objects (anything with a URL) for one’s own retrieval… [It is] done in a social environment (usually shared and open to others).”⁸ He finds that the “value in this external tagging is derived from people using their own vocabulary and adding explicit meaning, which may come from inferred understanding of the information/object;” however, in his view, “[p]eople are not so much categorizing, as providing a means to connect items (placing hooks) to provide their meaning in their own understanding.”⁹

While it is primarily a personal means for creating and managing tags and then annotating and categorizing the content, a folksonomy is unique in that it is a “bottom-up” (or personal) classification system, formed in a social setting by the creators of the content. This is in contrast to the more orderly, traditional “top-down” (or organizational) approach to cataloging or subject indexing, where metadata is generated by “experts” and based in a rigid, controlled vocabulary.

In many ways, this “democratic approach” avoids many of the social, political, and economic implications inherent in the top-down, centrally imposed systems.¹⁰ As David Sturtz says

[i]t allows the users of the system to establish their own sense of balance within the system, to use their own vernacular for indexing and retrieval, and prevents exclusion by creating new categories as needed. However, with no guiding hand at the helm, this communal approach has the power to shut out unpopular or misunderstood terms. The will of the community may flood them with useless content, use them in an unintended way, or marginalize them so that they essentially disappear.¹¹

Yet, in the midst of what many may perceive as complete chaos and meaningless content, we might find new ways of looking at our collections, new ways to interact with our users, new information about the photos we traditionally house in boxes and look to AACR2 to catalog. As the Alfred Hitchcock quote at the beginning of this section illustrates, the folksonomy is not for the faint of heart—or those who eschew complication.

What Can You Do With Flickr?

Uploading photographs or videos is simple, with both advanced and basic uploading options, as well as alternatives for uploading via email or mobile phone. Flickr also offers different ways to organize your content and connect with your users; we found that mindful use of tags, titles, and descriptions from the beginning made it easier to find photographs once we had a large number of images in the system, as did organizing them into “sets,” or placing our sets into larger “collections.”¹²

There are two main tools in Flickr for connecting with other users: contacts and groups. You can make anyone in Flickr a contact, and then you can organize contacts into friends, family, or friends and family. However, because you set up a Flickr account as an “individual” (even if your account is maintained by a group), you use “groups” as an organizational hub. Flickr’s Director of Community, Heather Champ, explains on TechSoup how members can use groups to connect with others through their photos and offers suggestions for ways that groups can offer a space for collaborators to upload and share their photos, either publicly or privately.

Like other photo-sharing sites, you can also create a one-to-one experience where you upload your photos and share them with a known set of people… But the essence of Flickr is more elastic. Flickr creates an environment where people come together to share their interests around photos. Using the Group features, people can share their photos and their thousand words.¹³

How Does This Compare with CONTENTdm?

The University Archives provides online access to its images in a CONTENTdm based digital collection. CONTENTdm is a collection management software system that allows libraries and archives to store and provide access to digital versions of primary source materials. It is billed as easy to install and use; flexible and fully customizable; scalable; open, extensible, and interoperable; and accessible through WorldCat searches.¹⁴ However, it is also capable of managing documents, audio and video files, PDFs, finding aids, journals, maps, and newspapers; it can also connect content, which allows administrators to upload and organize items into chapters, books, or collections.

When the OSU Libraries began its digitization program in 2001, CONTENTdm was chosen for several reasons including a reasonable cost, an active user-base, a visible migration path and ability to extract data in both custom XML and OAI, and the lack of available options; we continue to use it for many of the same reasons, adding in flexibility, ease of use, and the work the OSU Libraries staff has done to customize our program.

There were no substantial complaints with the CONTENTdm system; it provided the Libraries with a reliable system for managing digital images, users with both low- and high-resolution images and related metadata, and many possible “extra” features. While both CONTENTdm and Flickr are digital image management systems, at this time, only Flickr automatically offers an interactive or dynamic experience for users.

Currently, the default version of CONTENTdm does not allow users to tag or comment on the images; however, Terry Reese, the Gray Chair for Innovative Library Services, has modified our version and utilized its flexibility to “Flickrize” CONTENTdm, thereby allowing user-submitted tagging and comments. For this case study, this is important: while allowing user comments was a possibility, it was not an active feature on all collections. At the time this study began, the commenting and tagging features were manually activated on one of our digital collections, the Braceros in Oregon Photograph Collection.¹⁵ In general, we found that neither feature was used regularly by patrons. In fact, there was a single comment on an image of barrels of cucumbers harvested by Mexicans in Columbia County, left in July 2008, which read “I love pickles.” For the purpose of this study, the commenting feature on the Gerald W. Williams digital collection was enabled, and to date there have not been any comments.¹⁶

How the OSU Archives used Flickr

Based on our initial research, we determined that the staff time for account management and cost involved in joining Flickr was ideal. The upload process was simple, the price was low, and the project was unconstrained by any rigid standards or expectations; therefore, we didn’t feel any pressure to ensure impressive use statistics to make it worthwhile.

Once we set up the account and explored the features, we had to think about the workflow and logistics for maintaining the site. How would we organize our content? How could we provide access to and encourage interaction with our images? How could we distinguish between current and historic images? The OSU Libraries had an active Flickr account that was used by the staff and students in Reference Services; they had created a document that detailed their process for uploading images, assigning tags, and moderating comments. Although such a document was attractive for consistency’s sake (our account would also be managed by staff and students), we were also reluctant to be prescriptive or controlling. Chaos was not our goal, and wouldn’t serve users or site managers, but we also wanted to encourage exchange and contribution.

We created our own “rules” document, which defined who could change content, create descriptive information, and add metadata; which tags would be persistent and consistent; and how much metadata we wanted to include. Predictably, when I revisited this document months later in preparation for this case study, I found that we had made up a few policies that were unnecessary and a few that were actually inconsistent with our stated purpose of interaction! For instance, we had complex rules for which photographs we would allow users to tag, which upon reflection and with experience of use seemed irrelevant.

After addressing administrative concerns, we focused on the “social” aspect of this particular social software; namely, how could we reach Flickr and non-Flickr users to share our images? In other words, how could we actually get people to our site and how could we get people to say something about our photographs? We knew that groups offered a place for collaboration and, recognizing that history is made every day at OSU, we wanted to capture it and invite others to do the same. Unfortunately, since we couldn’t offer users the option of contributing directly to our page, we explored the “group” function. We didn’t find any group dedicated to this topic so we created our own, called “What did you see, Where did you go: History in the Making.” We asked students, staff, alums, and friends to view and contribute images that showed the “everyday” historical happenings at OSU and in its surrounding areas. We also joined several groups related to OSU as a means of connecting with those who might not find our account directly.

As we moved into fall quarter and into Oregon Archives Month in October, we increased our use of the account to promote our “visual” microform machine tutorials to reach new students.¹⁷ We also started to upload historic photos, including many directly related to programming events or from collections we wanted to publicize; in particular, we were interested in promoting the Gerald W. Williams digital collection. We used our own “real life” social networks to promote the Flickr site and CONTENTdm commenting feature, sharing information with others in the library, as well as our users. We knew from Flickr’s use statistics that people looked at our photos, though the number of views wasn’t high for either in-house or online access points—and no one left comments or added tags. Interestingly, this was consistent with what happened in the CONTENTdm collection: people looked, but no one left comments.

Concurrently, in September 2008, we applied to join the Flickr Commons, a project created to showcase hidden historic treasures and gather the public’s knowledge about them.¹⁸ We were impressed and intrigued by the user engagement with historical images and the substantial impact the Commons photographs were having on the larger Flickr community. On February 14, 2009, and Oregon’s 150th birthday, the Archives became the twenty-first institution to join Flickr Commons, and the first university member. At that time, we set up a second Flickr pro account (osu.commons) and began to gather use data. Joining Flickr Commons also provided another means for promoting the osu.archives Flickr account and our CONTENTdm digital collections.

Assessment of the OSU’s Use of Flickr

We began the project without many preconceived expectations; however, we suspected that CONTENTdm and Flickr users came to the sites (and the images themselves) with very different expectations. Unlike those who view the historic images in our CONTENTdm collections, visitors to Flickr expect that they will be able to interact with contemporary images. While, in theory, this desire for interaction holds the potential for a more dynamic relationship between users, as well as user and archivist, we know that not all images are created the same; in other words, not every historic image is interesting to the general, casual Flickr-user community.

As previously stated, we were eager to reach a broader and more diverse group of users with our Flickr account, but recognized that our interest needed to be coupled with both statistical evaluation and anecdotal assessment to justify the investment of staff time and financial resources. In many ways, it was important for us to delineate between the eras “before Commons” and “after Commons.” Joining Flickr Commons increased our visibility and traffic considerably, and we wanted to look at how entering a larger, more extensive community would result in additional visitors to both our original Flickr account and our CONTENTdm collections.

Pre-Flickr: Image Access Points and Assessing Use

As previously mentioned, before setting up the Flickr account, there were two access points to photographs: the Archives’ physical collections and the library-wide digital collections site. We track the use of the individual photographs in our physical collections through “Permission to Use” forms and through general use statistics, as well as through the high resolution download requests of CONTENTdm images. These statistics are useful; however, they only provide slices of detail for those images, and not enough for meaningful item-level analysis of use. Additionally, they only give us data for the images for which people have submitted direct requests. This doesn’t include the casual visitor to our digital collections who does not submit any formal reference or download requests. Unfortunately, we found that gathering specific information from CONTENTdm for images viewed proved problematic.

The level of granularity for general statistical reporting is not fine, nor is it easy to generate, which means that any queries need to be rather specific; moreover, when we started monitoring use statistics for this project, we realized that our version of CONTENTdm is currently configured to track information on the files that are uploaded or added to the system, rather than use data. In an archives world, this would be like tracking processing or accessioning, but not looking at reference statistics.

We turned to the Libraries’ Web analytics program, Urchin, looking for automatically gathered use data.¹⁹ Again, this proved to be only marginally informative. Because there are multiple entry points to the collections and multiple redirections occurring around the collection pages, the general collection page data is not accurate; however, we were able to view the number of times any given thumbnail image had been selected by visitors to the digital collections, as well as how many times the high-resolution TIFF image had been accessed. We were also able to filter the Urchin report by keyword, which allowed us to discern how many referrals to the digital collections came from sites like Flickr, search engines like Google or Yahoo!, or URLs with the words “osuarchives” or “osucommons” (one of the Flickr account pages) for any given time period. The results are shown in the table below, which has been divided into “pre-Commons launch” and “post-Commons launch”:

Table 1. Referrals to OSU digital collections

Referrals	Flickr	Google	Yahoo!	osucommons	osuarchives
July 1, 2008 to Apr. 14, 2009	532	8,830	1,104	255	230
July 1, 2008 to Feb. 13, 2009	94	6,766	252	6	101
Feb. 14, 2009 to Apr. 14, 2009	438	2,064	852	249	129

Beyond raw use data, what is interesting to point out is the proportion shown in these numbers. Compared with the period from the beginning of the fiscal year through the Flickr Commons launch, search engine and Flickr site referrals increased dramatically after the launch exposed us to a larger user base. One possible explanation is that by entering a more public space with our Flickr Commons, users knew who we were and looked for us. Another explanation is that we had tapped into an audience that expected and wanted to find historic photographs in Flickr, so they were more likely to click through the OSU Libraries digital collections.

Flickr: Assessing Use

In a side-by-side general comparison of functionality for gathering statistical data, despite some quirks (lack of a textual report, changed time periods for reporting use data), gathering use statistics from Flickr is much easier. From a staffing and workflow perspective, rather than contacting the CONTENTdm administrator or logging in to Urchin, a student worker can easily tally and track use data on Flickr. From an “ease of use” standpoint, we were able to login to Flickr and view real time statistics, complete with a graph. We were also able to see which photographs were most popular, which sites had referred users to our account, and how people had used the photographs after leaving Flickr (i.e., put them on a blog, or website). One potentially significant drawback is that Flickr only keeps statistics for the previous twenty-eight days; initially, we weren’t aware that there wasn’t a database of “archived” use data, so our recorded statistics start on February 11, 2009, just before our launch into the Commons. However, we have gathered interesting data during that period.

On February 11, there were 13,312 total views on our account. We predicted that there would be a massive spike in use once the Archives was a Commons member, though the actual increase was slightly less dramatic, as is shown in the table below.

Table 2. All-time views of images in OSU Flickr account

	Feb. 11, 2009	Feb. 16, 2009	Feb. 18, 2009	Mar. 4, 2009	Mar. 18, 2009	Apr. 1, 2009	Apr. 15, 2009
Views (Cumulative)	13,312	14,539	15,296	17,180	19,677	20,905	22,639

What is noteworthy is the steady bump in “total views.” In the seven and a half months before launching into Flickr Commons, there were 13,312 total views; however, in the eight weeks after the launch, that number grew by 9,327 views to 22,639. On one hand, this allows us to conclude that the launch into the Flickr Commons led to an increase in use of our osu.archives account; this is supported by increases in image views that correspond with new set launches in the osu.commons account. At the same time, it appears that users traverse the Flickr site, exploring and following links. More specifically, we know that most of the images on our osu.archives Flickr page had at least one view; however, beyond general view data, we wanted to know which images were viewed and what users did with them.

Which photographs were users viewing? Again looking at the time before the Flickr Commons launch, the top photographs viewed on the osu.archives account were those related to our microform tutorial, library events, or historic images of sporting events. There are many possible explanations for the popularity of these images. We worked to promote the tutorial in the library, which led to a direct link from various library subject guide pages; the library events photographs were promoted in-house and covered in local media outlets; and finally, these were “contemporary” images, the time period people expect to encounter when they visit Flickr. On the other hand, and this is key, while Flickr is known for being a site for personal photographs taken since the advent of color film, the most popular images on our osu.archives account were actually historic images. With an intentionally timed release when the OSU football team was flirting with the Rose Bowl, the images with the most views were those in a “football” set we created; while we promoted the set internally, the images were also mentioned on a highly visited sports blog. This led us to conclude that Flickr is an appropriate venue for contemporary images and tutorials, but it is also a good place for certain kinds of historic images.

Though this case study focuses on the osu.archives account, when we turn to an analysis of user interaction, it is interesting to provide the statistics for the osu.commons account for comparison’s sake. For the osu.archives account, we found that people weren’t likely to tag or comment; further, in looking at “all time” viewing statistics we were disappointed to see that the Williams images were not at the top, despite our marketing efforts during Oregon Archives Month and instruction sessions. In fact, the day we made our first set of Williams’ Collection photographs images “live” in the osu.commons account, which meant removing them from the osu.archives account, the first Williams Collection photograph to appear on the list of photographs viewed on the osu.archives account appeared at number one hundred eighty—with fifteen total visitors.

Tags

Flickr counts the tags we have assigned, so while these numbers are interesting, it may be misleading to draw any meaningful conclusions about user interaction.

• osu.archives: 939 tags on 737 photos, with 22,639 total views (April 15, 2009)
• osu.commons: 385 tags on 236 photos, with 60,657 total views (April 15, 2009)

When assessing the substantial difference in the number of tags assigned, it is important to note two things: first, there are more photographs in the osu.archives account, and second, we were consciously restrained in the tags we provided for the osu.commons account, wanting to encourage and allow users to categorize them without our influence. For the osu.archives account, of those items with tags that we did not assign, there were 155 photos with user-generated tags—with one person, a frequent Commons user, tagging eight images.

Comments

In general, most comments left on the osu.archives account were requests for photographs to be added to groups; interestingly, of the ten total requests for seven photos, six of those requests were on the Gerald Williams photographs now included in the osu.commons account. There were only two other comments: one on a Williams’ Collection photograph and the other on a Rose Bowl program cover.

• osu.archives: 10 comments on 737 photos (April 15, 2009)
• osu.commons: 85 comments on 236 photos (April 15, 2009)

As the above results show, it is obvious that the Flickr Commons community is much more likely to comment!

Flickr & CONTENTdm: Qualitative Analysis

The quantitative section, which contains statistics for views and use, naturally leads to a discussion of the types of photographs viewed, as well as an analysis of why some images are more appropriate for Flickr and/or more interesting to users than others. In essence, this question asks us to look at how user motivation and their intended purpose might influence the images viewed. Anecdotally, it seems that when people go to Flickr, they want to be entertained; though an archivist might find images of logging trestles in the 1930s entertaining, most people would be more amused by football rosters.

In this way, looking at the users’ projects is also quite important when looking at their inclination to tag or comment on photographs. Depending on their motivation for being on the site, they might be more or less interested in tagging or commenting. If they are casual users, interested in the community and social aspect of Flickr and social software in general, they are more likely to comment; on the other hand, if they are serious researchers, they are less likely to “waste time” talking to others and more likely to focus on their own project. It is also possible that, when researchers think about adding or using user-generated metadata, they are actually more likely to share their findings than read what others have written; again, this might be dependent on project and motivation. Another factor to consider is the generation or age of the user. This is an avenue for further exploration, but age may affect expectations and desires for interaction with historical materials.

In general, our Flickr projects have been a great success. They have increased visibility and access to our collections, given us avenues for further study or research, and given our users a unique opportunity to interact with the Archives and other users. While the easy access to statistics is also a pro, not having access to more than twenty-eight days of statistics and not having an easy mechanism for capturing use data is a con. The ability to upload photographs, the organizational and arrangement features, and the connection with the community of Flickr users all factor into a positive review.

Lessons Learned

We have learned many lessons during this process, and the general learning curve was shallow. One aspect of “social software” that we weren’t entirely prepared for was how “social” and “open” the interaction with our users would be. In a traditional setting, the archivist controls access to the historical information; however, in this online, off-site, out-of-the-reference-room setting, this is not the case. In many ways, the balance of power automatically shifts as soon as any image is placed online. Many of the immediate aspects of “control” are removed when a user is working with an item without any direct contact with an archivist. In Flickr, not only is work happening independent of this direct relationship, users are weaving their own research paths through commercial and “educational” sites, freely categorizing our images, questioning our rules and asserting their own, and commenting on what they see—all in a very public space.

This public evaluation and communication can leave an archives and archivist feeling exposed. One way we dealt with these shifting boundaries was to embrace them. As the Library of Congress status report on their project said

[c]oncerns about loss of control over content will continue to be discussed but can also be mitigated. Community practices and forums like the new Flickr Commons, where cultural organizations can now offer collections, help reduce the risks. Pilots like the Flickr project provide practical experience and concrete data on Web 2.0 risks and rewards, and help staff learn to operate in less formal environments that enhance recognition for a library’s valuable cultural roles.²⁰

We found that early and frequent interaction with the user community allowed us to become a part of the community; in Flickr, outreach and direct connections are necessary. We did this by making contacts, accepting contact offers, adding our photographs to groups, and joining or creating groups. Becoming a Flickr Commons member institution was also a means for connecting to our users, as membership came with a ready-made community and a dedicated group of users.

While the power of publicity is far from a new concept, the statistics give us concrete evidence of how critical outreach and outside site referral are for the success of projects like this. For instance, the football set was mentioned once on a sports blog, and those continue to be among the most-viewed images, with more referrals coming from that site than any other besides Flickr itself. Another example comes from our experiences with the Flickr Commons user group: if someone posts an image from our accounts on their group page or blog, it is always the top-viewed image, and we immediately see an overall rise in use statistics.

Again, as discussed in previous sections, we learned that there are “appropriate” collections for different accounts. Initially, we found the greatest success with sports-related images and image-based tutorials; however, once we joined Flickr Commons, thereby acquiring a user group with a more general interest in historical images, we began to include other images from our archival collections. For example, we released a set of First World War images taken in Ireland on Saint Patrick’s Day and a “Then & Now” set with historic and contemporary fashion related images timed with OSU’s Recycled Fashion Show.

Future Plans

In addition to general publicity and media outreach, we have many ideas for additions and uses for our accounts. We have a series of “How-to” tutorials for using and finding microforms, maps, government documents, and historical primary sources in the Libraries’ databases, and are working on a “How-to” tutorial for archival research, beginning with a tour of our reference room and introduction to our resources, to orient people to our space and demystify the process of doing research in an archive. We’d also like to continue our use of the geo tagging feature in Flickr, combining it with Flickr Maps to create guided campus or building tours. This provides an opportunity to showcase historical images, and also to document the current campus. Finally, we’d like to continue to partner with others in the Libraries or local community to foster relationships and promote our collections. We recently completed a large Flickr Commons project with the Oregon Explorer, a natural resources digital library, which celebrated Oregon’s 150th birthday with a collection of 150 historic images from Oregon’s fifteen river basins.

We also intend to explore a series of research projects, including how undergraduates and faculty might use a social image software management system in their own research, how social software might change the relationship between user and historical materials, how to enable “meaningful” user contribution of descriptive content or background information for archival collections (and how to work that user input into our workflow), how shifting boundaries and notions of authority might impact the relationship between the user and archivist, and how this inherently social technology might establish or strengthen a community of users.

Resources

Flickr Commons

indicommons. “indicommons.” indicommons. http://www.indicommons.org/ (accessed May 5, 2009).

Springer, Michelle, Beth Dulabahn, Phil Michel, Barbara Natanson, David Reser, David Woodward, and Helena Zinkham. “For the Common Good: The Library of Congress Flickr Pilot Project.” Library of Congress. http://www.loc.gov/rr/print/flickr_report_final.pdf (accessed March 6, 2009).

CONTENTdm

Reese, Terry. “Adding Social Tagging, User Commenting Support to CONTENTdm 4+.” Oregon State University. http://www.oregonstate.edu/~reeset/contentdm/downloads/CONTENTdm_Tagging.doc (accessed April 1, 2009).

Tagging

Golder, Scott, and Bernardo Huberman. “Usage patterns of collaborative tagging systems.” Journal of Information Science 32, no. 2 (2006): 198-208.

Rainie, Lee. “Tagging Play: Forget Dewey and His Decimals, Internet Users Are Revolutionizing the Way We Classify Information—and Make Sense of It.” Pew Research Center Publications. http://pewresearch.org/pubs/402/tagging-play (accessed April 23, 2009).

Folksonomies

Guy, Marieke, and Emma Tonkin. “Folksonomies: Tidying up Tags?” D-Lib Magazine 12, no. 1 (2006). http://www.dlib.org/dlib/january06/guy/01guy.html (accessed May 5, 2009).

Vander Wal, Thomas. “Folksonomy.” vanderwal.net. http://vanderwal.net/folksonomy.html (accessed May 5, 2009).

Sturtz, David. “Communal Categorization: The Folksonomy.” davidsturtz. http://www.davidsturtz.com/drexel/622/sturtz-folksonomy.pdf (accessed March 6, 2009).

Image Management, Archives, and User Studies

Chi, Ed. “Information Seeking Can Be Social.” Computer 42, no. 3 (2009): 42-46.

Craven, Louise. What are Archives? Cultural and Theoretical Perspectives: A Reader. Oxford: Ashgate, 2008.

Dijck, Jose van. “Users like you? Theorizing Agency in User-generated Content.” Media Culture Society 31, no. 1 (2009): 41-58.

Marchionini, Gary. “Exploratory Search: From Finding to Understanding.” Communications of the ACM 49, no. 4 (2006): 41-46.

Rege, Manjeet, Ming Dong, and Farshad Fotouhi. “Building a User-centered Semantic Hierarchy in Image Databases.” Multimedia Systems 12, no. 4-5 (2007): 325-338.

Rorissa, Abebe. “User-generated Descriptions of Individual Images Versus Labels of Groups of Images: A Comparison Using Basic Level Theory.” Information Processing & Management 44, no. 5 (2008): 1741-1753.

Stvilia, Besiki, and Corinne Jargensen. “End-user Collection Building Behavior in Flickr.” Proceedings of the American Society for Information Science and Technology 44, no. 1 (2007): 1-20.

Yakel, Elizabeth. “Inviting the User Into the Virtual Archives.” OCLC Systems & Services 22, no. 3 (2006): 159-163.

Zwol, Roelof van. “Flickr: Who is Looking?” Web Intelligence, IEEE/WIC/ACM International Conference on Web Intelligence (2007): 184-190.

Notes

1. This case study will consider and compare Flickr Commons, but will not discuss its implementation or impact at length.

2. The OSU Libraries digital collections are comprised of digitized documents, photographs, maps, and data.

3. Michelle Springer, Beth Dulabahn, Phil Michel, Barbara Natanson, David Reser, David Woodward, and Helena Zinkham, “For the Common Good: The Library of Congress Flickr Pilot Project,” Library of Congress, http://www.loc.gov/rr/print/flickr_report_final.pdf” target=”_blank”>http://www.loc.gov/rr/print/flickr_report_final.pdf (accessed May 18, 2009).

4. Flickr, “About Flickr,” Yahoo, Inc., http://www.flickr.com/about/ (accessed May 18, 2009).

5. Beth Kanter, “How Nonprofits Can Get the Most out of Flickr: Tips for Using the Online Photo-sharing Community Successfully,” Techsoup.org, posted on March 5, 2008, http://www.techsoup.org/learningcenter/internet/page8291.cfm (accessed May 18, 2009).

6. Wikipedia, “Tag Cloud,” Wikimedia Foundation, http://en.wikipedia.org/wiki/Tag_cloud (accessed May 19, 2009).

7. Bill Schaffer, “Cutting the Flow: Thinking Psycho,” Senses of Cinema, http://archive.sensesofcinema.com/contents/00/6/psycho.html (accessed May 19, 2009).

8. Thomas Vander Wal, “Folksonomy,” vanderwal.net, http://vanderwal.net/folksonomy.html (accessed May 5, 2009).

9. Ibid.

10. Geoffrey C. Bowker and Susan Leigh Star, Sorting Things Out: Classification and Its Consequences (Inside Technology), Cambridge, Mass.: The MIT Press, 1999.

11. David Sturtz, “Communal Categorization: The Folksonomy,” davidsturtz, http://www.davidsturtz.com/drexel/622/sturtz-folksonomy.pdf (accessed March 6, 2009).

12. You have to display either sets or collections; Flickr does not allow for a mix of the two levels.

13. Kanter.

14. OCLC, “CONTENTdm: Digital Collection Management Software,” OCLC, http://www.oclc.org/contentdm/default.htm (accessed May 5, 2009).

15. Terry Reese, “Adding Social Tagging, User Commenting Support to CONTENTdm 4+,” Oregon State University, http://www.oregonstate.edu/~reeset/contentdm/downloads/CONTENTdm_Tagging.doc (accessed April 1, 2009).

16. Interestingly, the addition of the tagging and commenting features on the Braceros collection was an experiment that emerged from user feedback on the “Rising Flood Waters: 1964 Corvallis” digital collection. This collection was created during another citywide flood and Archives’ users were keen to share their stories and give feedback on the 1964 images; in many ways, the desire for an online community was akin to the natural development of a physical community. Library and Archives staff assumed the same might happen around the Braceros images.

17. In the ACRL “best practices” document, one of the guidelines is that the teaching should respond to “multiple learning styles.” See Cecile Bianco, “Online Tutorials: Tips from the Literature,” Library Philosophy and Practice 8, no. 1 (2005), http://www.webpages.uidaho.edu/~mbolin/bianco2.htm (accessed May 5, 2009).

18. In January 2008, Flickr and the Library of Congress launched a new project aimed at increasing access to publicly held photograph collections in civic institutions. They called it “The Commons,” a space to showcase hidden historic treasures and gather the public’s knowledge about the pictures, creating even richer collections. The Commons opened its “doors” with nearly three thousand images from the Library of Congress; since then, twenty cultural heritage institutions from around the world have joined, releasing over fifteen thousand digital pictures.

19. Urchin Software from Google analyzes traffic for one or more websites and provides easy-to-understand reports on visitors. See Google, “Urchin Software from Google,” Google Inc., http://www.google.com/urchin/index.html (accessed May 5, 2009).

20. Springer et al.