[knowledge representation: tagging, folksonomy, content, information, literacy]

Claudia Cragg was going to interview me about tagging for her Creative Writing and New Media Master's project...sadly (and super annoyingly) my mic. didn't seem to work for skype today. So, we're going to go all social media with this interview and I'm going to post Claudia's questions here, with my answers, for any of you who might be interested in my views on tagging and folksonomy and digital literacy and and and...

So, here goes (caveat - my personal opinions!):


Who came up with the term Folksonomy and how is it defined?

Thomas Vander Wal came up with it during an e-mail list conversation in 2004.

As the name suggests, it's a taxonomy made by the folks – user generated definitions and information structures. But folksonomy is just a part of a larger idea: tagging. Tagging is the tying of words to objects. I think Vander Wall explains that this method of tagging has less "cognitive load" for users because it’s about key words rather than some kind of overlying systemic planning. I see it more of a free–form way of categorising information – personalising it.

Folksonomy is a subset of tagging – identifying/categorising for personal use, “re–finding” information

Has it caught on as a term?

Yes! Just do a google search for folksonomy; there are 1,620,000 hits (at 15:07 GMT). But then, I suppose just because it has "caught on" doesn't make it any less fractious. I'm thinking of "web 2.0" and how it is bandied around...still lots of problematising. I'm remembering Cory Doctorow's "metacrap" and I think a lot of people still don't quite "trust" the folks...that's why users concerned with retrieving the "right" kind of information might trust certain folks whose ideas they value...a kind of filtering through the (*wisdom* of the) masses.

What, in your academic opinion, makes a good or bad Tag Cloud? (i.e. your thoughts on Anatomy of a Tag Cloud vocalized)

Hrm...good question. Firstly, a caveat: there can be no *exact* laws or rules about good/bad tag clouds because the tags/vocab and value are constantly changing - a punctuated equilibrium.

For me, a good tag cloud makes information accessible to those who are interested in it. Tag clouds with a gazillion different terms look "messy" to me. A pet peeve is the inclusion of spam in tag clouds - that just changes the whole positive participatory idea behind folksonomy. Also, if taggers use a lot of similar words (as I did when I first started tagging) like: blogger, blog, blogging, blogs - that just adds to the mess. I guess rules are necessary, figure out if you're going to stick to uppercase or lowercase and whether you'll use singular or plural terms (blog or blogs? FirstName or firstname?)

Look at TechCrunch's tag cloud courtesy of technorati:

Most of the terms are of a similar size which makes spotting information trickier (at least for me) and there is some html included which shouldn't be there...so it seems messy.

A good tag cloud is "tidy" with (seemingly) transparent access to information. I don't want to be left wondering how the "blogger" and "blogging" tags are different and whether I should bother clicking both tags...I want the story (or most of it anyway) there in the cloud. I just want the general overview (I always look first for the tags that are weighed heaviest and then move to those tags least used), it's up to each tagger to make things more precise/personal to them. It's a vocabulary that's constantly evolving.


By the way, there are loads of tools out there to create tag clouds of your site (rather than of your delicious - or similar - bookmarks).

I've just used TagCrowd to make a cloud of my current blog which lists my last 10 posts:



How can Tag Clouds be used to drive traffic to a site?

In terms of general business use – there is huge potential here especially for smaller to medium size companies.

I think tagging can help with “findability” of company information although perhaps not so great with emergent vocab. which keeps changing. Also, I guess there might be a need to compare internal tags (tagged by employees) with external tags (tagged by customers) as each might have different words for the same or similar ideas.

But, as with peppering content with keywords, you can make sure you tag specific blog posts etc...with key words that you know your customers will search for.

Plus, the easier it is for customers to navigate a site, the more chance they’ll come back and using a tag cloud is, I think, a good way of making visible an overview of company info.


Tagging can also be a chance for any company (or organisation or university group etc...) to popularise their key word/s (or coin one) while simultaneously making data cohesive. Before we started using delicious there was no "nlab" as a bookmark, but now it is there and it means conference-goers and other interested parties can follow what NLab has been up to for the last 2.5 years. Shirky suggests that a refined approach to this kind of group classification is the next "big frontier."

I wonder how many tags there were for "longtail" before Shirky's article or for "web2.0" before frames of reference changed and people took to O'Reilly's coinage? (See Michael Wexler's 3 part series "I Hate Tags")

"In reality, our understanding of things changes and so do the terms we use to
describe them. How do I solve that in this open system? Do I have to go back and
change all my tags? What about other people’s tags? Do I have to keep in mind
all the variations on tags that reflect people’s different understanding of the
topics?"

If tagging is about naming/defining/narrativising content, then tag clouds aggregate content. Businesses can use this information in numerous ways, a few initial thoughts: establish a new market/audience, create a (new?) community interested in the same (or similar) things, get to know (on a deeper level) the needs of your customers and by having "tidy" tag cloud businesses are able to provide that much-called-for "transparency."

For example, a company can get an rss feed of a certain term and then track its usage (there are 190,688 photos tagged with "ipod" on flickr).

How are they best structured as 'jumping off points for dialogue'?

I think it's more about it's use-value. There isn't a "best structure" for dialogue but perhaps there are rules/strategies for certain kinds of dialogues. Two key words: tag clouds can refine conversations: they are "specialised" and can become (I'm optimistic) more "sophisticated."

As for jumping off points - tag clouds always already offer serendipity not structure (other than in the most general and probably ephemeral sense). I think when Will Richardson quotes Bruce Sterling who quotes Stowe Boyd (structure? what structure?) what he is saying can also apply to the role of tag clouds in dialogue:

"Basically, conversation is moving from a very static and slow form of
conversation — the comments thread on blog posts — to a more dynamic and fast
form of conversation: into the flow in Twitter, Friendfeed, and others. I think
this directionality may be like a law of the universe: conversation moves to
where is is most social…The way I am getting tugged to blog posts is
increasingly as a mention within a conversational bite in Twitter or Friendfeed.
I then click out of the flow to see the larger post, and offer my view in the
flow — not on the blog — and then I return to the flow, where I will be spending
most of my time. This makes sense: I want to talk about the blog post with the
person who brought it to my attention, more so that with some group of strangers
at the blog, or even the author, who I may not know at all. I also don’t think
we can expect the fragmentation of the social experience to slow down: it will
get a lot worse before it gets better."

Yes, tag clouds are dynamic and (should) reflect changing ideas and changing communication and people are probably drawn to tag clouds/taggers who offer valuable information but, in the end, it IS about communication.

What can be learned from their early use in Flickr for those wishing to use them in a more strictly narrative context - should there be distinctions between the types of clouds for predominantly textual content?

At the outset when tagging was new and etiquette hadn't yet entered the scene, I think people used as many tags as they could to classify something, trying to be as open as possible. but as the use has focused so have the terms and we see people (look at delicious) using a handful of tags to describe their bookmarks instead of trying to be all encompassing

I think clouds evolve according to the information so they’re kind self–aggregating or self–describing as the info changes so I don’t think you really need a distinction between types of clouds – i think that’ll be apparent to users.

Look at the flickr tag cloud from Jan. 2007 that I included in my "Anatomy of a Tag Cloud" post and look at the current flickr tag cloud (these are both for the "all time most popular tags"):

key tags for '07 were: wedding, party, japan, friends, family, travel, london

but today look how "France" has appeared as a tag and "band". Also, both flickr tag clouds have the term "girl" but neither have "boy."

However, both clouds have the tag "me." Isn't that an answer there - tagging isn't about structure; it's all about "me"!

Do Tag Clouds in any way alter the reception of text by a reader and if so how?

I think there is an interesting difference between people who tag for themselves and those who tag for others – when there’s a clearer idea of the subject the tagging is much more concise. Readers of tag clouds might judge a site by it's cloud (judge a book by its cover?)...and come to the site with the knowledge of the value of a site or of a tag (depending on weight and size of font etc...).

I wonder whether the tags imbue readers with a certain kind of passivity or...gee, not quite sure of the word...some kind of awareness of their role as reader rather than creator? I'm wondering this because the flickr tags seem to be mostly descriptors...where are the verbs? If readers were guided by tag clouds/tags that were active (running, reading, creating, see, listen, looking) that must surely change the perception/reception of any ensuing text/story/media? I think it's about different kinds of readers and contexts (what are you searching for and why) as well as different literacies.

See Sheryl Nussbaum-Beach's tactical or strategic view:

In what other ways have Tag Clouds evolved and how do you think they might evolve in the future?

Have a look at Philipp Keller's tag history:

and read Vander Wall's own rationalising of the state of tagging.

My thoughts on the future of tagging...hrm...I think there is going to be a visual tagging service. I’m thinking of the new visual search engine that I’m beta testing (searchme.com) and I think we’ll start seeing visual tags and maybe sonic tags. Besides the richer interfaces and applications

As an educator, I'd like to see tag clouds used as a mode of assessment. I'm thinking of Janet Harris's use of Tag Crowd to analyse the MSNBC Democratic debate:

(aside: isn't there loads of interesting stuff here...note who is the only person to mention women...hrm...also note the use of "America" but one candidate chooses only to say American, keeping it more personal?)

We could generate tag clouds (of work that is handed in electronically) of the student's most-used words. Wouldn't that be a good way of showing students why it's necessary to avoid repetition if they can actually *see* the repetition? We could also use tag clouds for our lecture notes or powerpoint presentations etc...to help students get an overview of the key points we're trying to share with them. What about generating tag clouds of 18th C. lit. and current lit. to see how vocabulary changes? When I taught a media module last term we looked at the supposed *neutrality* of reporters...but we could generate tag clouds for each reporter and compare how they write about different news items as well as compare what reporter A and report B say on news item C. hrm...seems lots of possibilities here. But, that age-old question arises... critical/digital/transliteracy: how do we *teach* students how to adequately *read* tag clouds.






buttons found at haveyouseenthisgirl on flickr.