Anyone still reading? David Weinberger (of Cluetrain Manifesto fame) responds to a piece critical of the folksonomy, tagging approach to classification by Elaine Peterson in D-Lib magazine. I’ll paraphrase loosely.

Peterson gives several objections to using tags as a way of organising information, as opposed to the sharp distinctions and uncrossable boundaries maintained by pupils of Aristotle. The strongest criticism she levels at folksonomies is that:

Because tags are relativized, personal, idiosyncratic views can coexist and thrive in the form of tags, in spite of their inconsistencies. Readers of texts on the Internet become individual interpreters, despite the document author’s intent.

Weinberger points out that, uhm, yeah — that’s the point. The author’s original intent is not the end of what a particular artefact, web page, blog post, photograph, movie means.

You upload a picture of your car to flickr and tag it ‘car’. I come along and think ‘mmm purple’, and tag it as such. Now when someone else comes along looking for a purple car, they have an easier time of it. Maybe someone else arrives and thinks, ‘that’s not purple, it’s indigo’, and they add that too.

Contradictions, idiosyncracies… and it’s much easier to find that picture of an indigo car.

Weinberger finishes:

I’ll take one step further toward the metaphysical: Folksonomies are not only frequently more useful than top-down taxonomies; they better reflect the bottom-up, messy, ambiguous, inconsistent, social nature of meaning—despite Aristotle and the tradition his genius spawned.

(found via Euan Semple)

For the purposes of this post, the distinction between Web 1.0 photo sites and Web 2.0 is a focus on their online presence and sharing on other sites and with other users, as opposed to the traditional approach of storage and printing. So what about flickr? People keep talking about them as the coolest thing on the block. What are they doing down there at the bottom of the graph? Well, actually, they are doing OK, with a 49% increase in market share between March and September of 2006.

Apparently, their rise is due to search engine traffic, though, rather than any social buzz malarkey. Because flickr encourages tagging and descriptions to a greater extent than other sites, its images are rising in the search engines. Knowing about that helps image searchers. When I’m looking for images for this site, for example, I always use flickr, not least because the ability to limit your searches to images with a CC licence really helps.

Is a relatively low market share for flickr bad news for Web 2.0? Not really. If you haven’t looked at it lately, Photobucket has actually adopted a lot of the features that drive the 2.0 trend. It hasn’t got groups or comments like flickr, but it does offer lots of options for sharing your photos elsewhere, such as your blog or facebook account. Strange that there’s no MySpace widget, but I know News International are pretty protective about third parties tapping into their audience. It’s the Web 1.0 sites that you need to feel sorry for, though:

Video sharing tomorrow…

a new wave of innovation on the internet, an innovation focused not so much on new technology as on the way people are beginning to use existing technology…

Quite a reasonable way to begin to describe these new sites and services, I would say. A certain degree of vagueness is almost inevitable given the breadth of quite different services that are described with the 2.0 label.

There’s also a certain amount of conventional wisdom in place, I felt. The idea, for example, that because certain properties have raised a lot of money then we are definitely in bubble 2.0 conditions. The ‘huge amounts of money’ ‘thrown at’ web startups nowadays are often fairly small compared to the hundreds of millions raised for dotcoms in the late nineties:

From the business point of view, the defining feature of this new goldrush is that established companies are throwing huge amounts of money at upstarts who have three things in common: they have grown from nowhere with astonishing speed; they have no revenue stream to speak of; and most of their content is provided by their users.

He goes on to divide this new wave into two rough categories. There are collective sites — such as digg and Wikipedia — and personal sites, focusing on ‘me media’, such as MySpace, del.icio.us and flickr. He allows that there is a lot of blurring between the two. Flickr, for example, is not just a gallery of your photos, but of everybody else’s. The distinction is reasonably useful, though, and allows for an excellent gag:

One way of putting it is to say that collective sites are useful (except when they’re not) and personal sites are interesting (except when they’re not).

The piece continues to describe the ‘800-pound gorilla’ that is MySpace. I got the feeling that Lanchester fundamentally dislikes MySpace and other social networks, though its size means that it’s certainly a subject of some awe: “if it were a country it would be the 10th biggest in the world, just behind Mexico”.

The piece ends on a melancholy note. For Lanchester, the social networking phenomenon is symptomatic of loneliness rather than the celebration of connection that others might see:

Sit someone at a computer screen and let it sink in that they are fully, definitively alone; then watch what happens. They will reach out for other people; but only part of the way. They will have “friends”, which are not the same thing as friends, and a lively online life, which is not the same thing as a social life; they will feel more connected, but they will be just as alone. Everybody sitting at a computer screen is alone. Everybody sitting at a computer screen is at the centre of the world. Everybody sitting at a computer screen, increasingly, wants everything to be all about them.

If you’ve got the morning off, check out the interviews and profiles with some key players: Jimmy Wales (Wikipedia), Craig Newmark & Jim Buckmaster (Craigslist), David Sifry (Technorati), Caterina Fake & Stuart Butterfield (flickr), Evan Williams (Blogger/Odeo), Joshua Schacter (del.icio.us), Tariq Krim (Netvibes), Martin Stiksel (last.fm), Kevin Rose (digg), Sam Schillace (Writely) and Michael and Xochi Birch (bebo).

Tags and tagging are a big part of the Web 2.0 ethos. Instead of sorting items into folders, you describe them with a series of words. The words you use, the ‘tags’, are up to you. Some people refer to this as ‘folksonomy’ in the sense that tags are home-grown and created by users, as opposed to putting things into folders in a tree structure decided by other people, ‘taxonomy’.

This can be useful for lots of reasons. Some of these are obvious:

The same item can be tagged with several terms. This post, for example, is about tags, but it’s also about web 2.0, del.icio.us, YouTube and flickr. It then occupies several locations in your filing system, and can thus be found in several different ways.

If you decide to bookmark it to del.icio.us or similar, then it’s up to you to decide how might you want to classify it. Your filing system comes to match your way of classifying things rather than one imposed by another person.

If your existing mental filing system can’t fit a new item, then you can simply invent new terms to accommodate it.

Multimedia, in particular, is very hard to classify using other means. The recently launched Google image labeler game challenges pairs of users to find matching descriptive words for pictures. The matching words are, of course, the tags that best describe the picture.

The big criticism of tagging is that it doesn’t work for finding things unless you think the same way as the person who tagged the item in the first place. If idiots do the tagging (viz. people who don’t think the same as me), then I’ll never find the item.

One example to show why this matters. I’ve been introduced to The Daily Show by YouTube. The YouTube filing system only goes as far as Videos>Comedy, but the clips I crave might be labelled Daily Show, The Daily Show, DailyShow, TDS or TheDailyShow. Sometimes, I might need to add Jon Stewart and (ahem) John Stewart for good measure. That’s potentially a lot of time wasted searching for clips that I could be wasting by watching them were there a more efficient way to find them in the first place.

Tom Morris states the problem neatly:

I’m a tag skeptic — I use tags on some services — del.icio.us for instance — but I think that their use is limited. A lot of people don’t grok tags — and the complete lack of tagging standards is annoying. I prefer an extensible ontology — that has outlining and multiple terms etc. Quite how you make it practical to do so is something I’m not sure about, but we can do better than tagging.

There are a couple of correctives to this, though.

Over time, standards do actually emerge for tagging items. You see the way other people tag items and you copy their technique. When I first started using del.icio.us, I used underscores a lot for my tags, such as social_media and web_2.0. Over time, I noticed that a lot of other people just shove the two words together to create socialmedia and web2.0. That started to seem a bit less messy, so I adopted that as my norm. (There is also the plus-sign brigade, who use social+media and web+2.0 — but they are just wrong ;))

Second, and thanks to Simon Collister for this one, tags help me find like-minded people. I said before that if an item has been tagged by an idiot then I’ll never find it. The point is, that by-and-large, I don’t want to find it! There’s so much information out there that what I really need to find are trusted sources. People who will recommend things that I am into.

Think about the distinction between someone who uses the tag ‘semanticweb’ as opposed to ‘web3.0′. In my little world, the first tagger is someone whose recommendation I’m more likely to trust than the latter one. The first tagger speaks my language. The lack of standards is actually an enabler to cutting through the noise and finding recommendations from like-minded people.

That doesn’t really work for my Daily Show cravings, though. But that’s not because the tagging principle is defective. It’s because of shortcomings in the way YouTube works. Only the submitter can tag items — if it worked like del.icio.us and items could be retagged according to users’ own tastes, then each relevant video could be found and tagged with the words I’d use to search for them.

Elsewhere:

Tagsonomy — a whole blog about tags and tagging

How tags change — from Tom Coates

I need to clarify that (before the divorce papers are filed). This is the top result for the search term ‘woman’, ranked by interestingness, that I found in a search on flickr this afternoon.

The picture was taken by the very talented Babeffe.

What makes for interestingness on flickr? It’s an aggregation of the number of notes, comments, favouritedness (sorry) and links to a submitted image.

The photo itself has been annotated a number of times by users. They point to the slight inequality between the eyes, the shape of the lips and the relationship between the woman, the photographer and the second woman in the picture. The comments are nearly all in Spanish, but my tourist-level translation skills suggest that she’s thought of as very beautiful by a lot of people.

But ‘interesting’? What does that word mean? (adjective 1. arousing curiosity or attention: arousing curiosity, attracting or holding attention, or provoking thought 2. not boring: enjoyable because of being varied, challenging, stimulating, or exciting). Thank you, Encarta.

Yes, she’s interesting. But the definition gives no idea of how to rank interesting things. In fact, it appears to be an entirely subjective quality, judging from that definition. That’s true in normal life too, of course. I tell people that I am interested in Web 2.0, and they tell me to grow-up and get a life. Does the fact that the vast majority of comments are in Spanish not suggest that there is a very strong cultural weighting to the idea of ‘interesting’?

I raise this because my new pal, Tim O’Reilly, has recently written on the subject:

Google made a breakthrough in web search with its original idea of links as citations (i.e. PageRank), and they are still the undisputed leader in general web search, but they haven’t done as well in searching rich media. I think they have some things to learn from Flickr. More specifically, web search innovators all need to think through what makes results “interesting” for a given domain. I like what flickr has done in calling out “interestingness” as a quality worth searching for, and leaving it as a playground for exploration.

I kind of agree. Interestingness is a quality worth searching for. I don’t want the most popular links on the subject I search for, say “mashups”, like Google gives me. I want the most interesting and informative one or two. Oh… hang on… that’s exactly how interestingness on flickr is calculated.

We don’t have agreement on the philosophical meaning of beauty, but we do have computer algorithms that will calculate it according to most people’s criteria. Again, we have a populist interpretation of very personal values. So by that scale…

Picasso is more interesting than Mondrian. That picture of dogs playing poker is more interesting than either. Still interested?