28% of Online Americans Have Used the Internet to Tag Content Forget Dewey and His Decimals, Internet Users are Revolutionizing the Way We Classify Information – and Make Sense of It
9 pages; PDF.

The report mentions Dewey Decimal Classification and includes an interview with the always interesting David Weinberger.

Quick comments from Gary. Comments from ResourceShelf Contributing Editor, Dan Giancaterino, are included at the bottom of this post.

1) We’ve always believed that tagging could potentially be VERY useful for individuals (personal information management) and small groups (e.g., a group working on a class project). It might also become useful for groups who use the same language/terminology to describe an item. For example, scientists and the use of tagging service like Connotea.

However, it might be a different story for the masses. Can or will tagging help the masses do a thorough search, potentially unearth useful content and, most importantly, save the end user time, effort, and aggravation?

Today, if you look at the most popular tags on del.icio.us a majority are technology-based topics. Are non-techies not tagging, at least publicly? Will these people tag and share publicly or only tag for personal info management? It’s a wait-and-see situation.

2) Why is it that whenever tagging is discussed, only the Dewey Decimal Classification is mentioned. Someone should point out that other classification systems exist and, as we’ve said before, there are differences between verbal subject headings (LCSH) and classification schemes like Dewey, UDC, and Library of Congress Classification.

3) Since tagging is done without structure, some users might describe an item with one tag, while others might use 20 or more. Do we need some basic requirements for public tag sharing? Again, what you do for yourself or in a small group is of course, your call.

4) Where does proper name authority control fit into this mix? Same name, different people? Same name, various spellings? Various names of a company? Again, all of this depends on the person doing the tagging (skills, time to get it done, etc.). What about location? If a image or document or image is tagged “Portland” does that it’s about Portland (for example the school system) or of Portland (a skyline image). Is it Portland, OR or Portland, ME?

5) From the report:

The act of tagging is likely to be embraced by a more mainstream population in the future because many organizations are making it easier and easier to tag internet content. For instance, Gmail users can label their email content and Amazon users can apply the labels of their choosing to books and other published material.

Again, this might be the case (only time will tell) but also reads like personal information management than mass sharing.

6) Vivisimo’s/Clusty’s management talks about letting a document speak for itself; clustering dynamically might be of more value than tagging (for the masses). Vivisimo calls it post-retrieval clustering vs. pre-retrieval tagging. Can clustering do a better job (letting the document speak for itself) of bringing like things together (for a successful, complete, and fast search) than tagging? Don’t forget that clustering using a controlled database is possible and offers more options, as seen with ClusterMed, where you can cluster on specific PubMed fields.

Finally. Ask.com* offers (with Zoom) related search suggestions to help save end users time finding related items and perhaps items they might not know about. In other words, using clustering and related search technologies as an info discovery tool

7) The Pew report makes no mention of spamming and spammers. How difficult is it for spammers to tag all items with the most popular tags? If/when tagging becomes even more mainstream as predicted in the report, will spammers become even more of an issue? It seems that as soon as someone solves the problem, another way of spamming is always found and exploited. Danny Sullivan has written on tag spam. Let’s also not forget that back in the early days of web search webmasters were allowed to use uncontrolled meta “keywords.” That feature, after a couple of years, was abused so much that it’s no longer useful. We also posted few weeks agotwo new research showing how a very useful feature on eBay (reputation management) is being gamed.

Subject headings and classification are supposed to work to bring like things together. Again, for personal use, tagging can be great. But, if tagging is going to grow (and it certainly can) spamming and other problems will have to be addressed. For example, application of tags to an item.

9) Of course, tagging will play a role in image and video retrieval. And spamming will also be an issue here. At the same time, content-based image retrieval will also grow to automatically describe what’s in an image or the “aboutness” of an image. To see a combination of a controlled vocabulary and CBIR used in synch, check out this demo from LTU of royalty-free images from Corbis.

10) How much time do people have to tag an item and is that reflected in the quality of the tagging? Will the masses be willing to give up this time? Again, let’s be clear, we’re talking about tagging for large group usage. What you do on your own, for yourself, is another matter.

10) Bottom Lines:
+ Tagging for personal or small group use? Potentially useful and valuable.
+ Tagging for large group (the masses) usage? Jury is still out in our opinion. A larger, more diverse group is needed with the understanding that they are tagging for use by a large group. Spam, spammers, and related issues (tagging skills, how tags are applied) must also be considered and kept in check. Right now, it’s hard to tell if its tagging to assist in bringing like things together, offering easier and more complete searches, turning up useful content, and saving the searcher time (most important) will be of value.

Source: Pew Internet and American Life Project

See Also: Newsweek Cover Story Tackles Social Search, Tagging and Related Issues (March, 2006)

* In addition to working on ResourceShelf, Gary is the Director of Online Info Resources at Ask.com

———————————————————————————————————-

Philadelphia-based law librarian and ResourceShelf Contributing Editor, Dan Giancaterino, has shared a few experiences and comments about tagging.

I’m not a big fan of tagging. Having started in the info biz at ISI, I had “controlled vocabulary” drummed into my head. And while I understand that tags can reflect an evolving terminology better than controlled vocabulary, they’re just too free-form for me.

Last month our Webmaster decided to post the Jenkins photo archive — about 1,700 images — on Flickr. She organized 2 training sessions for the staff. After that she asked for volunteers to help tag the photos. I didn’t participate because I was on vacation that week.

Today I browsed through the archive. I’ve included a sample photo which I think is typical. It was titled “Joe & Ken Moving Servers” and tagged with:

+ jenkins law library
+ library
+ servers

Now we’re a library with a staff of 30 — I think that qualifies as a “small group”. At least half of us have MLS degrees. And yet we did a poor job of tagging this particular image:

+ Who’s “Joe”? (And a couple of years from now, will we remember that he’s an IT contractor who comes in 1-2 days a week?)
+ Ken’s not really in the picture — he’s behind Joe.
+ Why are they moving servers?
+ Our Assistant Director, Kathy Coon, is in the lower left-hand corner
of the photo. Why isn’t she tagged?

Here’s how I would have (at the very least) tagged the image for both long-term internal and external retrieval:
+ Joe O’Connor, IT Contractor (perhaps a second tag for Information Technology Contractor)
+ Philadelphia, PA Law Libraries
+ Kathy Coon, Assistant Library Director
+ Ken Snyder, Systems Administrator
+ Jenkins Law Library
+ Server room
+ Servers
+ Library renovation
+ Information Technology

I know I’m being all anal and picky, but my point is this: if a bunch of librarians can’t do a good job of tagging, how can we expect users to do any better?

– Dan Giancaterino

This entry was posted on Wednesday, January 31st, 2007 at 12:45 am and is filed under Cataloging and Metadata, Info Management and Retrieval, Information Industry. You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.