Joho the Blog
|
|
|
September 21, 2005
NPR's "All Things Considered" last night ran a three minute commentary of mine on tagging. You can hear it here, thanks to Michael Shook. [Later: Here's NPR's own version.] Here's a near-transcript: Google is smart, but here's a tough problem for it. Let's say you type in "africa," "agriculture" and "grains" because that's what you're researching. You'll get lots of results, but you may miss pages about "couscous" because Google is searching for the word "grain" and doesn't know that that's what couscous is made of. Google knows the words on the pages, but doesn't know what the pages are about. That's much harder for computers because what something is about really depends on what you're looking for. That same page on couscous that to you is about economics could be about healthy eating to me or about words that repeat syllables to someone else. And that's the problem with all attempts by experts and authorities to come up with neat organizations of knowledge: What something is about depends on whose looking. Now a new tool is emerging on the Internet that helps us find things based on what we think they're about. It's called tagging, and without intending to, it's shaking up our ideas about how knowledge is organized. For example, take a look at the site that kicked off the new wave of tagging, It's called Delicious and its web address is http://del.icio.us. Let's say you come across the Moroccan couscous web page and you want to remember it. So you upload its Web address to your free page at delicious that lists all the pages you've saved. Then delicious asks you to enter a word or two as tags so you can find the Moroccan page later. You might tag it with Morocco, recipe, couscous, and main course, and then later you can see all the pages you've tagged with any of those words. That's a handy way to organize a large list of pages, but tagging at delicious really took off because it's a social activity: Everyone can see all the pages anyone has tagged with say, Morocco or main course or agriculture. This is a great research tool because just by checking the tag "agriculture" now and then, you'll see every page everyone else at delicious has tagged that way. Some of those pages will be irrelevant to you, of course, but many won't be. It's like having the world of people who care about a topic tell you everything they've found of interest.And unlike at Google, you'll find the pages that other humans have decided are ABOUT your topic. That's the real change in about-ness. Consider another tagging site, Flickr — that's f-l-i-c-k-r without the e — where you can upload photos you want to share with friends or the world. You might tag the snapshot you took of the guards at Buckingham Palace as "London" and "Buckingham." But I might come across it and tag it as "Big Funny Hats" because I'm working on a paper about fashion mistakes. You and I don't have to agree on what your photo is about. This takes classification and about-ness out of the hands of authors and experts. Now it's up to us readers to decide what something is about. Not only does this let us organize stuff in ways that make more sense to us, but we no longer have to act as if there's only one right way of understanding everything, or that authors and other authorities are the best judges of what things are about. And that's a big lesson. [Tags: EverythingIsMiscellaneous taxonomy tags delicious flickr] Posted
by D. Weinberger at September 21, 2005 10:29 AM
TrackBackListed below are links to weblogs that reference Me on Tagging on All Things Considered:
» Weinberger on Tags from Modern Marketing - Collaborate Marketing Services Tracked on September 21, 2005 03:17 PM
» Weinberger im Radio über Tagging from netbib weblog Tracked on September 21, 2005 04:16 PM
» Cool listening comprehension content from English360 Tracked on September 25, 2005 09:58 AM |
Comments
I heard you yesterday on my drive home, David. Terrific stuff. It made me think I need to get out of the dark ages and start tagging my blog entries and/or start using services like del.icio.us and flickr.
Posted by: scott | September 21, 2005 11:00 AM
First time I've been able to join together those small pieces you have been talking about for months, David. Sometimes there's something about the sound of a human voice that helps comprehension.
Posted by: johne | September 21, 2005 03:24 PM
Nice considering!
Has the definition of tagging settled around the social features you talked about (e.g., del.icio.us, and the "all public tags" on Flickr)?
I think the social aspect of tagging is a great topic to look at.
But, I'm not sure that what you're talking about is truly "classification", rather than more practical "categorization". (There is a lot of argument about the difference/confusion between the two, but, just looking at what you said, you're talking a lot about finding/organizing stuff and then seem to be suggesting that a new way to find/organize things establishes, or is based on?, a new way to define things.)
Are you saying "about-ness" to mean "is-ness" in some kind of true classification sense? Or is about-ness just locate-ness?
Given what you said, I'm not sure I could see tags as a way for people to classify things without depending on examples of people interacting with tags as a way to find things?
Posted by: Jay Fienberg | September 21, 2005 09:08 PM
Information on paper has been tagged for centuries, of course (if you're particuarly nerdy, see Taylor, Archer. "General subject-indexes since 1548." Publisher: Philadelphia : University of Pennsylvania Press, [c1966]. OCLC: (OCoLC)ocm00575660)
What's transformative about current developments is that they're being created socially in the information marketplace instead of behind-the-scenes by professionals. Creating universal thesaurii used to be the monopoly of librarians, and information seekers either employed the mandated terms or didn't find what they wanted. Today, sites like Flickr let a decentralized marketplace of searchers determine, through their behavior, which terms are most appropriate. We all had better tag with the most frequently used vocabulary or people won't find what we offer.
The central control of the professionals is gone -- a mixed blessing since so, too, is their unique expertise in such things.
Posted by: Michael | September 21, 2005 09:34 PM
Hi David;
I like how you point out this emerging folksonomy "helps us find things based on what we think they're about". That hit the nail on the head.
I've been watching the converging power of del.icio.us, Flickr and RSS and here are a few of my observations as well.
It seems to me that the Del.icio.us bookmarks are of higher quality than the links Google used to build an empire around. It's the next step Google might have done with links.
This simple platform, the sound of a thousands fingers pointing, soon became a crucial key to the new Mass Media by and for the Masses.
Del.icio.us and Flickr became channels for democratized real time reporting during the London bombings. Bloglines and RSS connected everything seamlessly, essentially turning the entire universe of Blogs and media into one stream.
Completely spontaneous emergent grass roots mass media, by and for the masses. Amazing.
Thanks for your excellent contributions.
Posted by: Ted Holmes | September 21, 2005 11:20 PM
Thanks to y'all for such helpful comments.
Jay, the All Things Considered piece focuses on social tagging as a way of finding, but hints at the bigger topic of what this does to meaning and to "is-ness." I talk about that latter topic in a piece I posted a couple of days ago.
I've been using "categorization" and "classification" synonymously because I've been sloppy and ignorant. Can you point me to a discussion of the difference? Thanks!
Posted by: David Weinberger | September 22, 2005 07:56 AM
Thanks for pointing me to "The New Is"--very nice!
Relative to categorization vs classification: AFAIK / IMHO, I think even the most serious library science literature can be sloppy in the usage of these terms.
There was some good discussion about this recently on the IA insititue mailing list. The list archive is password protected, but since you're a member, you should be able to see the thread on "question for the libsci wonks".
I recommend a blog post from Chris Dent on Classifcation vs Categorization.
And, I'm currently reading this interesting scholarly article, Classification and categorization: a difference that makes a difference. That's significant.
Finally, you've probably seen the new online social library catalog, Library Thing, which has social tagging. I think Library Thing's what are tags FAQ suggests how tags are an excellent way to create and maintain ad hoc categories. But, when I look at this, I find it hard to understand tags themselves as "classes".
Posted by: Jay Fienberg | September 22, 2005 01:46 PM
My biggest question about the usefulness of tagging comes from how does one know what to use for tagging? If we all tag something with some nouns that are close, does that still work? Doesn't there have to be some agreement on the tags to use? Do I tag it "Katrina", "HurricaneKatrina", "FEMAIneptitude"? Can I use spaces? If I do, will "Hurricane Katrina" and "HurricaneKatrina" still match?
Posted by: Jonathan Arnold | September 22, 2005 09:50 PM
Let's say you come across the Moroccan couscous web page
Waitaminit... we'd just established that you couldn't find the couscous page, hadn't we? (More here.)
Jonathan - that's the problem of controlled vocabulary, which I don't think tagging is ever going to do away with completely; at least, not in areas where it's important to hear what people are saying right now. Tagging has limits, and I think it's important to be aware of what they are (which is pretty much what I'm banging on about in the piece linked to above). See also here.
Posted by: Phil | September 23, 2005 05:01 AM
Phil and Jonathan, yes, controlled vocabularies take care of the problem. But so do two other approaches:
1. A folksonomy is, in essence, a bottom-up, voluntary controlled vocabulary.
2. Synonymns can be derived algorithically through cluster analysis or by using a thesaurus. (There are related services such as stemming that figures out that a search for blogging probably also should return hits on blog.)
Each of these approaches has its uses, IMO.
Posted by: David Weinberger | September 23, 2005 10:55 AM
A folksonomy is a slow bottom-up controlled vocabulary - and sometimes that's not enough. Hence the link I posted to tagsonomy.com, where a forum which exists to promote tagging was used to propose... a predefined controlled vocabulary.
Posted by: Phil | September 23, 2005 11:48 AM
Hmm, isn't a folksonomy, by definition, an uncontrolled vocabulary? My understanding of controlled vocabularies (at least in the library science sense) is that they are predicated on externalized authority (an institution, a cataloger, or a set of rules); and typically involve more than just synonym relationships, but also things like related terms ("see also" type references) and hierarchies (categories and sub-categories), which are more difficult to parse programmatically, absent something like the Semantic Web. This is not to argue whether one kind of system is better than another; just trying to clarify terms. Perhaps a more accurate term would be something like "self-organizing vocabulary"?
Posted by: Alex | September 23, 2005 01:54 PM
There is usually one strict controlled vocabulary within most popular tagging systems: there is only one domain of things called "tags". (Technorati has two, and del.icio.us actually has several domains of tags now).
The system controls this domain of the term "tags" and keeps it separate from other domains like (using upcoming.org as an example) events, friends, metros, users, groups, and venues.
Otherwise, I'd say tagging systems might better be said to have externally controlled vocabularies (e.g., people agree to use the same terms, but the system doesn't enforce it), or socially controlled vocabularies (e.g., control through: peer pressure, status seeking, tribe defining, etc.), or open vocabularies (any word or symbol goes).
Posted by: Jay Fienberg | September 23, 2005 02:36 PM
As far as tagging recipes goes, check out BigOven at http://www.bigoven.com. It lets you add arbitrary tags to any recipes you post, along with any other BigOven user. You can find, for instance, all Main Dish recipes that are spicy in the Chinese cuisine that have mushroom and chicken in them.
Posted by: Steve | September 25, 2005 11:09 AM
Jay, "socially controlled vocabulary" was precisely what I was getting at. Nice phrase, btw - I think it encapsulates most of what 'folksonomy' says, but without any added flummery.
Posted by: Phil | September 26, 2005 05:47 AM
http://www.macpimps.com/chat_mac/index.php?act=Print&client=printer&f=4&t=442&
Posted by: free sex | September 30, 2005 08:56 PM
Wow! Guess I'll speak for the minority on this one.
(Disclaimer: I'm an investor in a company named Attensa who just announced del.icio.us tagging functionality to great applause... so I have a real interest in seeing tagging become useful. That said, I have to be intellectually honest about my opinions.)
Gotta say that the notion of TAGGING pretty damn specious thinking. What many forget and Google understands is that TIME plays an increasingly influencial factor on how we recall information. We "tag" all the time when we name Word docs, Excel files, etc... We even tag in our daily offline world.
But as time passes, we forget what we tagged (named) our files -- hense the need for increasingly intelligent search. Creating yet another modality for recalling information will prove just to be a massive detour for most users. Improving the power of search is the key... services like Google and Apple's Spotlight have it right.
I'm not sure what to make of the couscous example sited above. If you're looking for couscous - then google couscous; not Africa+agriculture+grain.
Posted by: Scott Quick | October 2, 2005 07:55 PM