Tags/folksonomies and Topic Maps

<< 2006-04-06 21:37 >>

Tagging is an interesting and useful technique, but the way it's currently used lacks a lot, and I think Topic Maps are the perfect way to fill the gap between what tagging is and what it should be.

How tagging works

The principle is of course extremely simple: you put tags on a piece of content to classify the content, usually by saying what the content is about, or something about what the content is. So if I used tagging in this blog, probably I'd add tags like "folksonomies", "tags", and "topic maps" to this posting. This would let me easily find all postings related to folksonomies, tags, or Topic Maps. A folksonomy is just a set of tags. And that's it, really.

Some web sites, like Flickr and del.icio.us, aggregate pieces of content tagged by different users, so that the folksonomy emerges from the activities of many different users who know nothing about each other. This is the use of tagging that really has people excited, of course, and these folksonomies tend to add an additional feature: a syntax for allowed tags. Typically, all characters must be lower-case and users are not allowed to put spaces in their tags (the spaces are replaced either with underscores or simply taken out altogether). This is of course done to reduce the number of different tags you can create for the same name.

Tagging compared to other things

If you know taxonomies a folksonomy is really just a taxonomy without the hierarchy. People will usually tell you that another difference is that taxonomies are created by professionals while folksonomies are created by normal users. I don't really agree, as I've yet to find a folksonomy that has a hierarchy (or a taxonomy without one), and, really, who creates the classification has nothing to do with what kind of classification it is. If you want to distinguish different kinds of classification systems from each other the only useful criterion is really their structure.

If you know Topic Maps it's even simpler. A folksonomy is basically a topic map with one topic per tag (and each topic has only a single name), and each tagging of a piece of content turns into an external occurrence of a single type. That's it.

So what's missing?

Well, actually lots of things are missing from simple tagging. The good thing about it is that you can quite effectively organize stuff simply by attaching tags to it. The downside is that there are lots of things you'll eventually want that simple tagging just won't do. Here's a simple list of missing things:

• You can't have multiple names for the same thing. So if people search for "XTM" but you called it "Topic Maps", well, bad luck to them. In Topic Maps you'd solve that by attaching "XTM" as an additional name for the same topic.
• You can't have two different tags with the same name. So you can't have "Paris" and "Paris"; instead, you have to create tags like "paris" and "paris_ontario". Not too pretty. In Topic Maps this just isn't an issue, since you can have two topics with the same name without problems.
• You can't relate the tags to each other. This may not sound like much of an issue, but if you look at pictures from Flickr you'll quickly see that this means people have to use more tags than they really need to. A picture of the building I live in, for example, would probably get tagged with "schweigaardsgate", "old_oslo", "oslo", "norway", "europe", "the_earth", "solar_system", ... Well, maybe not all of those, but you get the idea. Associations between topics saying that each of these is contained in the next would solve this neatly, because you'd then know that any picture from Oslo would necessarily also be from Norway.

Some people also talk about "meta noise" in that tags as used on Flickr and del.icio.us inevitably get used differently by different people. This is true, but really has more to do with how the classification is managed (or, in the case of Flickr and del.icio.us, not managed) than with the technology itself. Of course, with Topic Maps you'd be able to merge tags that mean the same, and you have more ways of dealing with the problem, but at heart this is a process/control problem, and much less a technical one.

So?

The benefit and attraction of tagging is that it's simple: you just write a list of tags, and you're done with classifying. The benefit of Topic Maps in this scenario is that you don't have to give that up, but if you treat your tags as topics you can get a lot more out of the same approach without necessarily having to put that much more in.

Similar posts

Comments

Richard Volpato - 2006-06-07 05:12:00

This is a very good summary! Keep up this kind of work.

While looking 'over the fence' (from the garden of Topic Maps) you might notice the new RDFa initiative.

I (and many others) would greatly appreciate your considered view of this deployment of RDF as embedded mark-up in web pages (generalising micro-formats) and yet, it seems to me, opening up a new resource for Topic Maps, a kind of 'grove' in new garb :^) See RDFa.info.

Nils - 2007-03-07 08:55:08

What you write here is very much in line with my own thoughts, Lars Marius. I am actually doing a presentation of this on a seminar tomorrow (http://www.jbi.hio.no/bibin/nils/imageintermsseminar.htm thus I am interested in learning about user-friendly applications that combine folksonomies and tagging, i.e. folksonomy topic maps applications with a very low technical treshold for tagging. Do you know of any?

Lars Marius - 2007-03-07 09:13:25

I think Fuzzzy.com has to be the best example of a Topic Maps-based social bookmarking system out there.

I recently did a presentation on Web 2.0 and Topic Maps myself at the Topic Maps Users' Group. I'll try to post the slides soon.

Lars Marius - 2007-03-07 09:36:38

The Web 2.0 slides are posted now, at http://www.ontopia.net/topicmaps/materials/web20.pdf

Add a comment

Name	required
Email	optional, not published
URL	optional, published
Comment
Spam	don't check this if you want to be posted
Not spam	do check this if you want to be posted