[Sigia-l] Faceted approach applied to content

Travis Wilson trav at ciaheadquarters.com
Fri Nov 14 22:27:04 EST 2003


>Do the values of a facet need to be mutually exclusive?  Because I'm working
>with content and concepts, more complex ("fuzzy") than a bottle of wine, I 
>will
>allow an internal tagger to select multiple values from a facet to describe an
>entity.

It's not a no-no; you should do whatever most clearly connects your users 
to the resources. But consider that you'd be defeating the purpose of 
faceted classification if you assign multiple values.

What you're describing is a system where your resources (pieces of content) 
can be associated with a bunch of different headings ("values" or 
"attributes"). There would be no restriction on the mapping; a resource 
could connect to no headings or all of them. Then, on top of that, the 
headings themselves are classified as Facet A or Facet B or Attribute Type 
C or whatever. That's not exactly the same thing.

Topologically, the whole point of faceted classification is that for a 
particular facet, the resource can actually be *classified*. It can be put 
in a bucket that's exclusive of the other buckets. If you have an auto 
part, a "manufacturing location" facet makes sense, because you made the 
part at that place. You know where to go to get the part, and you know who 
to call for support, and so on. If the part falls under two different 
headings in that facet, though, you have a conflict and no obvious way to 
resolve it. You've simply associated the part, not classified it. Strict 
classification avoids conflicts like that.

So the question is, is faceted classification really the model you're 
looking for? If not, you should just use the model you've developed. It'll 
adapt well to some uses and adapt badly to others.

Travis Wilson
Development
http://facetmap.com




More information about the Sigia-l mailing list