Man is an animal suspended in webs of significance he himself has spun, I take culture to be those webs, says the American social anthropologist Clifford Geertz in his famous book The Interpretation of Cultures from 1973. Classification is ways of placing knowledge into meaningful and representative categories. What is thought of as meaningful is closely related to one specific culture and a certain set of values. Looking through a semiotic magnifying glass, cultures represent unique sets of symbols, values and experience, and it is clearly no possible way to make one language or way as a default truth to classification or universality as such.
Found this somewhere at the net but can't remember where! Apologize to whoever took this great picture!
Instead of trying to capture the whole world as a main goal, I believe it’s essential to define who you target group is and what your classification scheme attend to represent before one starts building any representation as is slices of reality.
By putting classification and categorization in relation to the user’s context the categories will be more meaningful for those who will use it because the categories will reflect established and recognizable ways of seeing the world. By context I mean both the users actual situation and their cultural framework. Cultural is here not related necessary to geography, but to what Benedict Anderson refers to as imagined communities as well as to subcultures – or as your target group so to speak.
By exploring the context of the users it is necessary to use a qualitative methodology with its informal face-to-face interaction with informants, open ended questions and observation in natural settings where your target group most likely will use what ever you are developing. In this ethnography you will receive, you might uncover people’s prototypes – as in the cognitive psychologist’s Eleanor Rosch’ prototype categorizations sense. Prototypes are typical representations of concepts people have– and is closely related to how you see the world. For instances the concept ‘circus’ might bring associations to tent, clowns, elephants, acrobats and so on – and certain concepts will not be mentioned because they are not relevant. It could be the concept ‘rabbit’ – but not if it is represented together with the concept ‘magicians’, or to ‘shark’ as one would probably place as fearful as ‘tiger’ but seldom associated with the concept ‘circus’.
The point is – we need to, in my point of view, explore other aspects of classification based on how actual users experience their surrounding world, more than the other way around, where technology determine the perceivers options/classes of categorizations. And then we need to build our categories or faceted search on behalf of these findings – and not on the type of element that lies and sorts underneath – ‘excel documents’ or by elements that are based on similarities to ‘her/him/child’ – as we often see as categories at e-stores of clothing. I think we need to ask – with good old Gregory Bateson in mind – What is the difference that makes a difference? By exploring this, meaningful categories might emerge. And this is what we need to ask the people that are likely to use the system. What is the relationship between the ‘genders’ in for example an e-clothing store? Is ‘age’ an important issue? Or ‘prize’? Or is it the whole context all at once? Could one think of classifying products based on typical situations the user/customer will wear a certain product? Do I want to explore, say trousers as such, or am I looking for trousers I can wear at work, at home, in my spare time, at holiday, at Christmas Eve or during mountain climbing? These situations will most likely differ culturally, and also sub-culturally, but still.
I agree with Vegard that there isn’t much extraordinary – at least not what I’ve seen, and I admit I haven’t been seeking the whole internet – about how categories and options to choose between are visually and interaction designed represented. One reason for this could be related to this aspect of the often lack of context. Are categories based on qualitative or quantitative differences? Lets say, for instances if we take the already mentioned e-store clothing example you can use visual metaphors’ your target group has a known relationship to. Categorized by, say apparel more than the products attribute as such – on situations where I will wear this object - more than the object it self – because there is qualitative differences in the products attributes.
If you classify your stuff within different occasions the buyer are likely to identify – work, sports, party, weddings, relaxed – and giving the user an option of what amount of money she expects to use represented by a visual metaphor by skinny, thick and loaded wallets representing the amount of money the customer sees herself to spend for shopping and at final giving the customer different visual roles she will identify herself as to represent age – she can then cross-choose between
- product based on situation and style
- money expected to spend
- age (and style) of the customer
- gender of the customer
In this example you have introduced context – namely mine – to your categories and by doing this you add at least meaning to your classification scheme and it will have value for a specific group of people that identifies themselves within this set of representations.
Within other cultural settings your categories might have no significance at all – people might have other expectations to context and place things related to work rather to the sports context and so forth.
I guess my comment was one more of the why’s more than drawing up a how as Christopher Kelty announces we are starving for in his interesting book Two Bits – The cultural significance of free software.
The iPod is an excellent example
I guess you’ve seen Jared Spools key note, Journey to the center at design, at this years IA summit as well – I throw in the link for those that has’nt: http://www.slideshare.net/jmspool/journey-to-the-center-of-design
I think you’re making a very good point, Lene. Faceted search (and navigation) should be based on basic level categories and prototypical category members. Everybody knows what to expect from the basic level category MP3 players (as opposed to eg. portable multimedia devices), and the iPod is the instantly recognizable prototype. So when you’re searching through consumer electronics, it’s obvious what you’ll find behind the MP3 player link with the iPod icon/thumbnail.
Another example is datetime navigation. Some time intervals, like hour, day, week and month, are prototypical examples and easily understood. Other time intervals, like 25 minutes, 1 1/2 day and 8 months, are more vague. It depends on context, of course, which intervals constitue basic level categories. My point is that it makes little sense to navigate search results on arbitrary time intervals (which happen to be the default configuration of FAST ESP).
You are spot on about the need to observe users and how they experience the world around them. You can’t rely on technical know-how when it comes to modelling meaningful representations in the real world. User studies and interaction design first, technical problem solving thereafter.
Sometimes a simple phrase search might be useful, at least when it is one specific object which you want to search for, and not the class of it – to use a term from object programming.
The reason for pointing this out in this context is displaye when googling the phrase:
“compliance and inertness clause”
This will return one of the famous one-hits in Google, and points out the page 321 of “Having Thought” by John Haugeland” with the subtitle of “Essays in the Metaphysics of Mind”.
Even though this is not the reference to the photographer, as Lene was missing from her attached illustration, the content of this book might possibly touch the topic of the article?