Flickr brings tagging to vintage images
By Daniel Terdiman
Staff Writer, CNET News.com
Published: January 17, 2008, 1:34 PM PST
Scores of gorgeous historic photos--from shots of early 20th century baseball players to 1940s-era images of horse-drawn carts and factory workers--showed up on Flickr this week, and the public is busy tagging them in an effort to bring new context to the collection.
The labeling is part of a pilot project by the U.S. Library of Congress, which is making 3,115 of its archival photos available for public tagging in an attempt to bring a sort of "wisdom of the crowds" intelligence to the photos' metadata.
The project kicked off Wednesday in conjunction with the launch of a Flickr initiative dubbed "The Commons." That effort aims to open peoples' eyes to the "hidden treasures in the huge Library of Congress collection" and to show people how their "input of a tag or two can make the collection even richer."
For now, the library is making two distinct groupings available on the popular photo-sharing site: "1930s-40s in Color" and "News in the 1910s." The former is filled with images of World War II-era industrial scenes and military personnel posing in full-dress uniform. The latter contains equally stunning shots of random ball players, barns, yacht clubs, and more.
The Commons is Flickr's attempt to showcase and add context to the archives of public-facing institutions--first the Library of Congress and later potentially others, such as museums or other civic entities.
"I'm interested in whether this can establish the usefulness of the folksonomic approach, in combination with the expertise and curatorial skills that these institutions hold," said George Oates, program manager for The Commons at Flickr. "I don't know what the future looks like, but I do know that there's a lot of interest in the museum industry about this...approach."
For now, though, The Commons is going to focus exclusively on the library's archives. One notable aspect of the project is the addition of a new Flickr copyright category in which the photos are said to have "no known copyright restrictions."
It might be tempting to read that designation as equating to "public domain," but that's not the case, say those driving the initiative at the Library of Congress.
"It's always incumbent on a user of any work to do their own due diligence" about copyright, said Matt Raymond, a public-affairs officer at the library. (The public can visit the library's Prints and Photographs Web page for an FAQ on the specifics of copyright issues pertaining to the images in The Commons project, as well as to any of the millions of other images in the library's archives.)
It may eventually be possible for the general public to submit photos to The Commons project, Oates suggested, but for now, Flickr will limit participation to public or civic institutions.
As a result, the public will have to continue using the previously existing copyright designations Flickr offers, such as "some rights reserved," "all rights reserved" and assorted licenses available under Creative Commons.
Regardless of the copyright issues involved, the library and Flickr are pushing these initiatives because of what they see as a unique opportunity to bring to bear the knowledge of the photo-sharing site's millions of users. The images, which previously had very little metadata attached, will now be searchable by the countless tags being added.
In addition, said Raymond, the project is giving the library a chance to experiment with the latest interactive technologies.
"We're moving as aggressively as a government agency can to recognize the growing importance of Web 2.0," Raymond said. "This was a very low-cost opportunity to observe the tagging behavior and evaluate the quality we could get, and get our feet wet in a Web 2.0 community."
Already, just a day after the project debut, some individual photos have received more than 50 tags. Some might think so many tags for an individual photo could water down the usefulness of the metadata, but those involved in the project disagree.
"That's where the long tail comes in," Oates said. "If you compare a photo with 3 tags with one with 26, you've increased" searchability by orders of magnitude.
The Library of Congress initiative shares characteristics of projects like SETI@home and Amazon.com's Mechanical Turk, both of which harness the power of the masses to look for solutions.
Those projects, which both task individuals with contributing their time or computer processing power for the larger good, are in some ways the definition of what Wired magazine contributing editor Jeff Howe calls "crowdsourcing."
"The cost of five minutes of a user's time is so marginal, it's almost a why-not," said Howe, author of the forthcoming Crowdsourcing: How the power of the crowd is driving the future of business. "Are they asking for a day of my time? No way, I've got kids. Are they asking for a few minutes? (Then no problem)."
One might think the public would be willing to help the Library of Congress because the institution's mission to archive and collect public knowledge and information is benign. But Howe suggested that there would be those who would want to get involved no matter what the agency was.
"If it was the (National Security Agency) asking if I'd want to classify spy photos," Howe said, "even then, you'd get the wannabe spooks" to help.
That's exactly what Flickr and the Library of Congress are counting on as their side-by-side projects get off the ground.
Some observers view the projects as an admirable way of mixing public and private expertise.
"Except for my general nervousness about putting this stuff into a privately held, for-profit organization," blogger David Weinberger wrote Wednesday, "I think this is quite cool. It has the advantage of putting the data where the people already are. As a footnote to the posting says, it takes a photo of a grain elevator as an example 'because it helps illustrate that there are active Flickr user groups for even such diverse subjects as grain elevators.'"
Michelle Springer, project manager of digital initiatives at the Library of Congress, said the agency doesn't mind that its partner is a company that can leverage its participation for potential profit.
"The library's interest is in working with virtually any partner that will help us achieve our mission," Springer said, "and the key thing we keep in mind is that we avoid exclusive arrangements. So if other people wanted to work with us and do similar things, it's not unprecedented for us to do that."
She added that she hopes that the library's experience with the project might spur other government agencies and public institutions to follow its lead.
But for now, the library is just getting used to the fact that the public seems to have responded to its project with very open arms and with an unexpected amount of participation.
"We didn't have a sense of what all we would learn," Raymond said. "And by going through the exercise, we will reach undiscovered territory and uncharted countries."
vendredi, janvier 18, 2008
tagging our history
Information architecture 101 says that you ask the target audience what it calls things, in an effort to make them easier to find. That's the awesome power behind folksonomy. Now, the Library of Congress and Flickr are teaming up to categorize lots of vintage images. Awesome!