Phil Michel and Michelle Springer from the Library of Congress presented on the LOC’s Flickr Pilot Project. The Library of Congress was the first cultural heritage institution to partner with Flickr to share photographic content and invite user participation and comments. With 15 institutions participating in what is now the Flickr Commons, it is an idea that caught on quickly and has been quite successful. I’ve been very excited about this project since its launch, and so I was motivated to clean up and blog my rather extensive notes on the session. For more information about the project, check out this LOC webcast.

Background

The motivation for the project came from a desire to explore including user generated content (UGC) in LOC descriptive processes. Photos seemed like a good place to start because there is no language barrier, there was already a big collection of photos online, and because they’re fun.

Initial investigations showed that bringing tagging to LOC collections would have had high technical barriers if handled in-house. There was a desire to keep initial expenditures low, and so they started looking around for existing web 2.0 sites that were doing the things they wanted to do.

The project had three goals: Increase awareness of LOC collections; gain better understanding of social tagging; gain experience participating in the kinds of web communities that are interested in LOC materials

There were a number of principles that guided the development of the pilot project: The involved content must already be available on the LOC site; the agreement with the third party site must be non-exclusive; access to the content must must be free; there must be an option to control or exclude advertising on the account; LOC should be clearly identified as the source of the images; must allow LOC to remove and moderate user-supplied content to prevent inappropriate tags and comments; UGC must be clearly distinguishable from Library generated content; must be possible to accurately convey copyright status.

Flickr had a great deal of appeal as a partner: It recently announced the upload of its 3 billionth picture, and has an active user community of over 23 million members. It had a pre-existing, vibrant community built around photography and a conversation that included notes, comments, and tags. From a technical standpoint, it also had APIs that allow for batch uploads and batch downloads of UGC, and a history of dealing with alternative copyright status (Creative Commons licenses).

Getting it off the ground

Flickr programmed the “No Known Restrictions” option especially for the LOC partnership, and it is now used by most of the institutions participating in The Commons. Every institution has its own page in its own webspace the explains exactly what they mean by the statement.

Some time and effort was required on the part of the General Counsel’s office to work with Flickr to create a modified Terms and Conditions agreement that could deal appropriately with the Library’s status as a government institution.

Technical process: Someone (I missed who – Flickr, LOC, or both) built a Java(?) app called Flickrj to push and pull content between the LOC databases and Flickr’s. They chose selected MARC fields whose content would go to Flickr along with the photos: The MARC 856 field was used as unique machine tag value, and so was the DublinCore identifier field.

All together, getting the project off the ground took about 100 hours of work for technical staff.

The photos all went to Flickr in their rough state. LOC folks didn’t do any cropping, color fixing, or clean up of dust or scratches. Part of the curiosity was to see how the public would respond to the images in this rough form.

Startup investments: The Library of Congress purchased a $24.95 Flickr Pro account, which offers members unlimited uploads and stats about traffic to photos. The Pro account is an annual expense that will go on as long as the project does. All Commons member institutions have pro accounts. There was no full time staff assigned to the project, but it required General Counsel involvement, some big conference calls, and eight staffers who contributed about 20 hours each to collaborate with Flickr on development.

Launching and maintaining

This was the first project that LOC ever announced without a press release. There were announcements on the Library of Congress and Flickr blogs, and the organizers considered it a soft launch. Though it involved no mainstream press, there was an enormous initial response, totally out of proportion to what was expected. The result was some near-immediate revisions of plans for maintenance and direct staff involvement; the scale was too big to be as involved as they’d planned.

A number of LOC staff share responsibility for monitoring all new comments, notes, and tags. They use the Flickrj app to pull all the new UGC at once. It takes about 2 hours a week to moderate comments/notes/tags for spam and inappropriateness. Sometimes users call attention to these things before staff find them. There are very few problems with inappropriate tags or comments; the Flickr community is quite well-behaved. LOC staff don’t correct spelling or syntax or remove seemingly useless tags. Staff do accept group invitations from public group administrators, but they only join public, nudity/vulgarity-free groups, so monitoring the group invites also takes time.

Updates to the images themselves take 15-20 hours a week. These involve corrections to descriptive information, fixes in the LOC catalog, and occasionally image fixes. Sometimes the orientation of images is wrong. First they fix it on the LOC server, then they generate new derivatives, and then send corrected versions to Flickr. In general, they limit edits to very basic changes and real errors. Sometimes they’ll point people from the LOC catalog back to Flickr when large amounts of conversation, updating, and information-sharing are taking place for a particular photo.

During the Q&A, someone asked about how have time pressures changed over the course of the project. Turns out, they haven’t exactly gone down, though they have shifted. When the project first launched, staff was checking the new comments and tags every 24 hours, and it was totally overwhelming. Efficiencies have come from the technical solutions, like the ability to batch download all new comments, notes, and tags. However, as the number of photos keeps going up, time demands on moderators continue to go up. Part of time demand comes from level of participation in the community, which is a steady stream; activity doesn’t stop on the older photos, so the rising total number of images leads to a rising total amount of new user generated content.

Outcomes

One of the main goals of the project was to drive more traffic to the LOC photo collections website, and it worked. People visit the LOC pages for higher resolution images, to get additional information, and to browse related collections. The organizers feel that the pilot has definitely achieved goal of raising awareness of LOC photo collections.

An unexpected outcome: Major search engines are finding, exposing, and weighting LOC’s Flickr images in search results. Many of the photos rank very high in images searches. It’s an unforseen way to further expose the content to the world.

Many of LOC photos are also being embedded in blogs all over the web (including this one). When it happens via “Blog this” function in Flickr, it’s easy for LOC to track it (and I imagine it’s trackable even when it happens in other ways).

The user involvement has been very interesting as a source of further study. There is a core group of about 20 commenters who provide historical research, fixes, comments, notes, etc. They’ll often support the information with citations, links to NYTimes archives and other external sites and archives. There are also 10 “power taggers” who have applied more than 3,000 tags each. One person was responsible for over 5,000 tags. The people at LOC did some work examining the different types of tags that people apply, and identified nine different categories: LC description based, new descriptive words, new subject words, emotional/aesthetic responses, personal knowledge/research, machine tags, variant forms, foreign language, and miscellaneous.

Users frequently post modern photos in the comments to show what the featured locations look like now. Sometimes people will go to the featured location and reenact the photos. There is quite a bit of playfulness and humor in much of the user involvement. Notes are a useful way to identify people in crowd shots and to transcribe text that appears in the photos. Some people also use notes to make jokes or silly comments, and while some people in the Flickr community have objected to the proliferation of notes, LOC has decided that for now the value of the function outweighs the irritation.

Conclusion: There has been a great response to the pilot, and great user participation, learned a lot. It stimulated conversation both between users & librarians and also between librarians. The project tapped into expertise in that resides in communities of interest. It brought up issues related to presentation and engagement that can inform decisions about how materials are presented on the Library’s own web site. While there are some risks associated with jumping into the web 2.0 world, and you have to be willing to cede some control, the benefits and rewards have been terrific.

• Sessions attended: 15
• Parties attended: 3
• Business cards collected: 16 (it seemed like more – I probably lost some)
• Moo cards collected: 5
• iPhone sightings: I lost count
• Minutes spent waiting in line for the bathroom: 0 (the best part of attending a heavily male-dominated conference)
• Most professionally relevant panel: Textbooks of the Future: Free and Collaborative
• Most endearing panel: The I Can Has Cheezburger guys
• All around favorite session: Kathy Sierra’s 20 Ways to Woo Users

It was an exhausting, exciting, inspiring, overwhelming experience. I can’t wait to come back next year.

I was surprised to learn that Warner Brothers has (or claims to have) a very liberal policy when it comes to no-budget and student filmmakers requesting licenses to use clips; while WB can’t legally grant free permissions because of agreements with unions and guilds, they do issue “no objection” letters guaranteeing that the student can use the clip without fear of a lawsuit. I can’t wait to tell my Communications and Art & Design faculty.

There were also two filmmakers on the panel, and I was particularly charmed by M dot Strange, an independent, pro-piracy animator. His philosophy is, “You should be happy if anybody wants to watch your movie in any way, and you should try to figure out how to make money from it after.” He suggests making the DVD value-added – his offers 8 alternate soundtracks to the film, along with very cool cover art – in order to encourage your fans to buy your movie after they’ve already watched it for free online. He also argues that “If you don’t have a P&A [Prints and Advertising] budget, piracy is your promotion.” The torrent seeders essentially worked on dot Strange’s behalf, not only providing free distribution, but also promoting the film on their websites and creating subtitled versions in several languages, instantly expanding his potential audience. He’s my new poster child for the power of unauthorized distribution.

The follow-up question, “What do we help our users kick ass at?” was challenging for me, because my copyright specialist job involves a lot of different things, and I’m still trying to figure out the best way to do them all. Do I help my users kick ass at negotiating with publishers? Maybe, but it’s not what I focus on. Do I help them kick ass at understanding copyright? Maybe, but “understanding” is not a particularly active or ass-kicking verb. Do I help them kick ass at advocating for their rights, as creators and as users? I hope so. Enough blogobrainstorming, back to wooing users. More on this in the future.

A lot of the material in the talk was stuff that Sierra has covered before, but most of it was new to me. She spent some time discussing stress, and how we should help our users manage stress and do our very best not to create more of it. One way to do that is to give people patterns and shortcuts that will help them do things faster. “Best practices” are not motivating; shortcuts are.

She also brought Gary Vaynerchuk up on stage, as an exemplar of someone who does all 20 user-wooing things. He seemed very charismatic, and his advice to novice wine drinkers sounded good to me: 1) Try different stuff (“Stop drinking Yellow Tail people!”), and 2) Respect your own palate.

Baraniuk and Moeller saw the textbooks of the future coming out of print on demand technologies, while Klein believes that POD is all wrong for updatable fact-based works. He argued that the web is superior for textbooks because our understanding of science and lots of other things is constantly changing, and those changes can be incorporated into a networked electronic text instantaneously, while paper is static, and therefore instantly outdated. I don’t agree; print books are still a very useful technology, even for fact-based textbook-type things. Customizable, cheap, print on demand books have the potential to be even more useful.

I thought the most interesting part of the session was hearing Richard Baraniuk from Rice University talk about Connexions, a nonprofit that provides “a place to view and share educational material made of small knowledge chunks called modules that can be organized as courses, books, reports, etc.”

Connexions offers a collaborative medium for quickly creating and sharing scholarly work, and they’re working to provide some sort of peer review-esque credentialing system. My understanding is that while it’s a collaborative platform, it’s also closed. They aim to have experts creating and developing content, and have partnered with groups like IEEE to recruit authors and reviewers.

They use XML to turn all their pieces of content into recombinable building blocks that can look like a single, unified textbook. The final products end up costing many times less than conventional textbooks – the sample engineering textbook that Baraniuk mentioned would have cost $130 from a traditional publisher, and cost $20 from Connexions. It seems like a very promising model.

According to Baraniuk, three big changes made Connexions a viable project:

1. New technology (XML)
2. New intellectual property regime (Creative Commons)
3. New quality control mechanisms.

The panel also spoke a bit about the copyright regime that open educational resources require. Creative Commons has been a boon to Connexions, while the OLPC folks prefer works without any licensing restrictions at all. According to Klein, once the XO laptops are widely available, “The only barrier… to getting textbooks to the third world [will be] the licensing barrier.” I’m not sure if I buy this, given the other major barriers to education in developing countries, but certainly in an online world, I agree that licensing is the biggest barrier to access.

For me the most exciting part of the session was to see that J Strother Moore, the Chair of UT Austin’s Computer Science Department (one of the 10 biggest CS departments in the country), is an old white guy who is very invested in recruiting women to his department, improving their experiences once they get there, and helping them go on to have satisfying careers in Computer Science after they leave. He clearly has a good grasp of the problems and challenges that women face in his field, and his ability to speak frankly about those things, and his evident dedication to fixing them, gave me hope.

Favorite moment: During the Q&A section, a young CS professor asked what he can do to support the only two girls in his programming class.

Dr. Moore’s answer: “Don’t hit on them.”

Exactly.

Jenkins talked about studies showing that the current generation of young people tend to speak in “we”, while the older generation speaks in “I… you…” The Language of We is the language of social networks and collective intelligence, the language of collaboration and the Internet. Jenkins suggested that Hillary Clinton and Barack Obama are an example of this split. Though Obama isn’t in Generation First Person Plural himself, part of his ability to appeal to it is perfectly exemplified by his slogan, “Yes, we can.”

Jenkins mentioned a new non-profit called Organization for Transformative Works, that seeks to establish that all creative works arising out of fan culture (“fanworks”) are “legal and transformative.” Rock on.

Also regarding fan culture, there was an exchange between Jenkins and Johnson in which the whole, “Who are these people, and why do they have so much time on their hands, (and what’s wrong with them)?” question came up. And Jenkins said that most people participating in creative fan culture are “pink collar workers – teachers and librarians and nurses,” and then basically flipped the question on it’s head by asking, “What’s wrong with American culture that these bright and creative people are getting so little intellectual challenge at work?” I was sort of flabbergasted – it sounded to me like Henry Jenkins just said that America is wasting the brains of its women. He never said the words “woman” or “women” or “female” or “girl”, but he did say “pink collar” and “teacher” and “librarian” and “nurse”. The conversation moved on, and it never came up again, but if I’d been gutsier I would have gotten up during the question period and said, “Henry Jenkins, it sounds like you just said that America is wasting the brains of its women and that’s why we have such vibrantly creative fan cultures. Is that what you just said? Would you like to elaborate?”

I learned some interesting things:

1. All but one teen “has a MySpace”, and the one that doesn’t sounded like she was actively rejecting it. Only two are on Facebook, and one of those has a sister in college with whom she communicates on Facebook.
2. They don’t email each other. They use MySpace or text messages or they talk face to face at school. Email is for website registrations and contacting teachers.
3. They all hate ads on websites. They especially hate pop-ups, ads with audio files, video ads, and ads that tell you you’ve won something when you haven’t. Even so, one of the kids said that when he sees ads that have little games in them, he can’t help but play them. “Even when I know it’s going to take me to a page I don’t want to go to, I still have to shoot the monkey.”
4. While only two of the kids identified as “gamers,” all the kids said they play games, on their phones and online. They like Tetris and PacMan, and little Flash games.
5. The teens time shift most of their TV watching using the web. They all watch YouTube. Most interesting to me, they make no distinction between watching a full episode of a show on the CBS website, watching it on the licensed third-party site Veoh, and watching it on any one of the many sites that stream pirated shows from China. Wherever they can find the episode they want to watch, that’s where they’ll watch it.
6. Some of the teens seemed pretty savvy about being advertised to, and they understand that the ads are what pay the bills. One commented that he likes to watch sports clips on NBA.com, but hates that before every single clip, there’s the same commercial for Miller Lite. It’s annoying, and he’s not interested in beer. “The advertising should go with what’s on the site,” he said. “If it were for basketball shoes, that would be okay.”

It was a great panel. These kids will be going to college in a few years, and they are going to bring their expectations for technology and communication with them to campus. They may not want their professors or librarians on Facebook, or sending them text messages, but they may respond a whole lot better to the things we try to teach them if we put it in the shape of a little Flash game.