Folksonomy
February 02, 2005 | 
22 Comments
Summary: A very, very simple overview of Folksonomies.
Update (02/03/05): Folksonomy defined. “A Folksonomy is a collection of metadata created by users.”
I’ve been thinking about folksonomies lately and it seems like so have many of you. Since I mentioned them a few weeks back I’ve had lots of questions about them and there are been much talk about how useful they are (or can be) and where and when they’d provide their greatest use. I’ve not had time to put too much thought into it, but I’d like to open it up for discussion and see what many of you think about this stuff.
As well, I’ve spoken to quite a few people who just don’t know what the term “Folksonomy” means. I went looking and found a few definitions and descriptions, but not one I found really easy to grasp right away.
So, for those who don’t know, and because I couldn’t find one to share with you, I’ll try and give you a simple, practical, jargon-less, and hopefully accurate definition of my own:
The term “folksonomy” was coined by Thomas VanderWal and it’s used to describe the tagging and classification systems used by many social networking applications (Flickr, 43 Things, Del.icio.us, etc.) wherein individual users tag data or content (Web pages, photos, links, lists, etc.) however they choose and the system aggregates those tags and (usually) groups them into a weighted list of some kind.
I’m not expert, so if you want more detail try the Wikipedia.
A great, and often used, example is Flickr. If 10,000 people use the tag “cameraphone” and 2,000 people use the tag “moblog” than, when the weighted list of tags is displayed to a user, “cameraphone” is going to be given more weight. For more, here is a blog entry on how Flickr uses Folksonomies.
This is great because it allows for organic social classification of data. It’s not perfect for everything, of course, but you can see how, for applications like Flickr, it can be really great. It’s another way we can address how users search for, think about, and digest information. The one-size fits all way of doing things never really worked and we’re starting to figure out all sorts of great new ways to address the problems we’ve been having with information on the Web.
Having said all of that, I’ve not yet found any sort of practical use for them in my own work. Not yet anyway. If anyone reading this has, I’d love to hear about it.
For more on Folksonomies:
Filed under: IA and Usability
Comments
1. Geof F. Morris said:
Actually, the term refers to the creation of the metadata [the taxonomy] and its creators [the folks], not to any weighting or grouping done with said metadata. That’s more of an aggregation technique, although it certainly has lots of value.
GFM
Posted on February 2, 2005 01:48 PM | #
2. Ste Grainer said:
(Not to be picky, but you forgot a zero in the 10,000 above. It just looks sorta weird with the comma where it is. ;)
Anyway, I’m still working on my own implementation, but I’ve a rudimentary “tagging” system set up on my new website. Eventually, I’d like to build in sorting by popularity or content amount as well as tying related content into technorati’s tags somehow. Right now, my system is basically just a huge mess of categories, but from that I should be able to cull more advanced info (category count and such). Of course, there’s still sparse content so building and testing an overall system will be difficult for a few weeks/months.
Posted on February 2, 2005 02:03 PM | #
3. Keith said:
Geof – That didn’t take long…Well, sure, I guess it’s probably more accurate to give that distinction, but I added that bit about the lists to help illustrate how they’re used.
Trying for a practical explanation, like the kind I’d give my mom.
As well, I’ve heard those actual wieghted lists called “folksonomies” as well. I’m not much for semantic debate however, so call ‘em what you like.
I expected a few IA folks to have problems with that (no surprises at all) and that’s why I added the Wikipedia link.
Posted on February 2, 2005 02:08 PM | #
4. Joshua Porter said:
Keith, I wrote a post on applying folksonomies to a site just yesterday.
In short, I suggest that the usefulness of folksonomies comes from the aggregation of the data, not the act of tagging. I doubt that many other sites will be able to leverage tagging the way that flickr, del.icio.us and 43 things have.
That said, there are all sorts of ways to leverage what people do on a web site, as long as it’s something easy to keep track of (like tagging is).
Posted on February 2, 2005 06:21 PM | #
5. vanderwal said:
A slight point of correction, folksonomy does not cover classification as the current Wikipedia entry states. Classification is heirarchial, which a folksonomy definately is not. The folksonomy is, as Geoff stated regular folks using their own language to tag objects (on del.icio.us it is URLs and Flickr it is photos). The weighting is just an information visualization applied to the how many terms are used to describe “X”.
The value in folksonomies is derived from many people adding their own tags (often so they can find the information again, but easily). The more people tagging one object the better the value. Patterns start showing easily as to what the norm of the group that is tagging the information will call something. But, the more people tagging also allows people that have alternative vocabularies to find the information in the long-tail of the power curve.
Posted on February 3, 2005 05:08 AM | #
6. Tanya said:
Thomas said “Classification is heirarchial…”
Classification can result in a hierarchical scheme, however, it does not by definition have to. Other types of relationships can be expressed instead of whole - part, or it can consist of facets each containing a list of flat values, etc. etc.
Posted on February 3, 2005 07:17 AM | #
7. Nick Finck said:
Keith, Geof is correct, a folksonomy is not something specifically identified as a tagging system, but instead the ability for a community to metadata for a given object.
For some reason a lot of people believe Tagging and folksonomy are synonymous and I believe it is because of articles like this that try to show a tangible example yet sometimes add an incorrect perception in the process.
Not that you did anything wrong here.. it’s just.. well, I hope someone develops a non tagging version of a folksonomy in use soon… very soon, because tagging is really a whole different problem within itself.
Posted on February 3, 2005 07:40 AM | #
8. Keith said:
Joshua, Geof, Thomas, Tanya, Nick – Why is it always so hard to get people (IA’s in particular) to come up with (and agree on) easy definitions for things? Here I try to offer a simple (and mostly correct) definition an you all come in here and really confuse it all up.
Funny thing is, you can’t even seem to agree even when you say you do.
Aren’t IA’s supposed to *add* clarity? ;)
If someone could have offered up an easily explainable definition for this I’d have just posted that, but (no big surprise considering the short discussion here) I couldn’t find one and now it turns out that what I did find and tried my best to illustrate, isn’t even agreed upon by everyone.
So…if what I’m pulling out of what y’all said is correct. A folksonomy is simply a taxonomy (metadata) created by users?
This taxonomy (the folksonomy) could be created by users tagging things (like photos, ala Flickr) and could then be subsequently weighted and or classified.
If that is correct then my understanding is correct as well, however my illustration may have been off. Thing is, I’m a visual thinker when it comes to this stuff, and I need examples that I can visualize to get my mind around things. So using Flickr as an example help me to understand the usefulness of a thing.
Hopefully I didn’t bend the folksonomy box too out of shape. ;)
Posted on February 3, 2005 08:50 AM | #
9. vanderwal said:
One thing a folksonomy is definately not is a taxonomy, it is the inverse of a taxonomy. A folksonomy has not formal structure it is a collection of words, which are tags (sorry Nick folksonomies use tags, which are used for other things that are not folksonomies) to provide metadata.
In short, a folksonomy is a collection of individual people’s tags using their vocabulary to mark an object so that they can find it more easily in the future. Folksonomies are bottom-up, user definced terms for objects (as opposed to top down formal taxonomies). Folksonomies do not have heirarchy, but associative inference can be derived.
What people do with the folksonomies, dirive contorlled vocabularies, search, build visualization tools (weighed display), formalize into classifications, etc. some places where value can be added. Part of the problem with the concept, which the term folksonomy is used for is people tying derived use to the concept and practice.
Posted on February 3, 2005 10:13 AM | #
10. nick santilli said:
Folksonomies aside, I see “tagging” as metadata for the people.
IMO, Folksonomies is clever, but somewhat of a misnomer. Taxonomies are generally structured - information/metadata architectures. (CMS’ have largely brought these into the limelate as of late)
Tagging allows Joe Schmoe to apply metadata to “stuff” that is relevant to him. As these social web-accessible apps allow communities to tag their stuff, the collections of similarly tagged stuff creates an ordering taxonomy based on popularity rather than on a logical hierarchy.
So yes, the folks are creating a new breed of ad-hoc metadata architecture. It’s not a logically structured taxonomy (that’s a little redundant I guess), but Folksonomy is catchy and relates the idea anyway.
my comments aren’t in opposition to anyone else, but I guess more for me to sort out my thought process on the topic…
Posted on February 3, 2005 10:37 AM | #
11. nick santilli said:
ah - You got the same stuff I was writing as I was writing it keith. :)
I think everyone’s saying GENERALLY the same thing here, but the terms are not all agreed-upon.
This thread is making my braing hurt…
Posted on February 3, 2005 10:40 AM | #
12. Keith said:
Ok, so a simple definition would be that a Folksonomy is a collection of metadata created by users.
Wow. Sure seems like it’s been made much more complicated out there then it should be.
Posted on February 3, 2005 10:52 AM | #
13. vanderwal said:
Keith, you have it.
What separates folksonomy from tagging in general is the folksonomy is people tagging information they do not own or control and more often than not the tagging is done by more than one person. This gets broken into broad and narrow folksonomies (see the Wired article on Folksonomies for clarity on this distinction). The distinction is mostly what can be done with the tags.
Posted on February 3, 2005 12:19 PM | #
14. Keith said:
Ah, at last!
Maybe I should title this post “Folksonomy - Learn along with Keith.”
Posted on February 3, 2005 12:25 PM | #
15. Nathan Logan said:
Perhaps a petty comment, but I think I’d be more apt to take this concept more seriously if it was labeled something other than “folksonomy”.
But thanks for the head’s up on it. I haven’t read anything about it before now.
Posted on February 3, 2005 12:38 PM | #
16. Nick Finck said:
Keith, that’s perfect and to the point. How about a followup post with your new definition? :P
Thomas, it seems what you call tags, I call metadata… adjectives grouped together to define something. To me a folsonomy doesn’t have to rely on a spcific action or process, much less a technology. “Tagging” seems to be action specific, and in some cases technology specific. I think that’s where my confusion lies.
Posted on February 3, 2005 01:53 PM | #
17. vanderwal said:
Nick, tags and metadata are slightly different. This is what Lou discussed in his widely absorbed Folksonomies? How about Metadata Ecologies? entry. Metadata has a more controlled aspect in its tag selection. Folksonomies are often called free-tagging, in that there is no learned metadata skill involved. Metadata is a subset of tags. How tags are applied, who applies the tags, how the tags are used and harvested is where it all comes together. This is where things get more difficult to understand.
Posted on February 3, 2005 07:51 PM | #
18. Ryan Brooks said:
Great! What a fabulous term! Huzzah!
…
In fact, it’s a great word, if only phonetically.
-Ryan
Posted on February 4, 2005 07:15 AM | #
19. Kevin Tamura said:
Salon today has an article on this very topic today, or tagging as they call it. Of course you’ll need a subscription or watch some ads to read the article.
Posted on February 8, 2005 09:05 AM | #
20. Zachary said:
If anyone out there has an extra XBox Live subscription card laying around, please email me the code. I am in desprite need and any codes would be greatly appreciated. Thanks ahead of time.
Posted on February 20, 2005 01:39 PM | #
21. Tom Smith said:
I think folksonomies could be a magic bullet.. if they are combined with 2 other falied magic bullets from the past, MetaData and AI.
Posted on May 9, 2005 07:15 AM | #
22. Kent said:
I have just implemented a folksonomy, it is still pretty much in its infancy and potentially buggy, however if anyone wants to try it, it is at http://www.bigblogzoo.com.
You use the tree to navigate to one of the blogs or feeds and then you press ‘animalise me’, which brings you to the folksonomy.
If anyone has any comments I would be very interested to hear them:
http://www.syndicatescape.com/bigblogzoo/
Posted on June 15, 2005 05:48 PM | #
Comments are now closed