Gmane
From: <mike.steckel <at> sematech.org>
Subject: RE: Re: Mutually exclusive?
Newsgroups: gmane.comp.infodesign.facetedclassification
Date: 2002-12-19 18:46:06 GMT (5 years, 37 weeks, 3 days, 23 hours and 20 minutes ago)

I think this is really getting to the heart of the matter. I say Yes, you need mutual exclusivity all the way down the tree. You can, however, use more than one term from a facet to describe an object. Remember that creating the facets (facet analysis) is not the same as assigning the terms from the classification to an object. Create your classification with mutually exclusive facets and assign both "New York" and "New Jersey" as the content warrants.
 
Mutual exclusivity for facets is very difficult, as was noted earlier today, but extremely important for the development of good facets. What you are creating is a "concept map" with facet analysis, or an understanding of the content that makes it easier to describe the concepts within it. Facets are post-enumerative, not pre-enumerative, so you are not putting an object inside a structure, but pulling terms, and the concepts they represent, from the structure to describe the object.
 
I would also say that what constitutes a good or bad facet in the abstract is difficult, though most likely your examples are good ones. "Places" and "Things to do" seem to be good facets because there can't be overlap between them.
 
Also -- Love the new name Peter!
-----Original Message-----
From: petertheman <peter <at> poorbuthappy.com> [mailto:peter <at> poorbuthappy.com]
Sent: Thursday, December 19, 2002 12:13 PM
To: facetedclassification <at> yahoogroups.com
Subject: [facetedclassification] Re: Mutually exclusive?


> > "The sorting of terms in a given field into homogeneous, mutually
> > exclusive facets, each derived from the parent universe by a
single
> > characteristic of division."
> This is one that bothers me. 

Me too, and I have always wondered if I haven't misunderstood that
requirement. Maybe a library person can help me out here? In XFML
(http://xfml.org), the mutually exclusive idea only applies to the
top level of a tree, for example "places" and "things to do". They
have to be mutually exclusive, so two facets called "places to live
in" and "places to visit" would make bad facets.

Within those top level nodes, things do *not* have to be exclusive,
because indeed you can classify something to belong to New York *and*
Jersey city for example. (Maybe an article that compares them)

So my question is:
- is there a need in facet analysis to have exclusivity all the way
through the tree (with a tree I mean something like: "places>USA>New
York state>New york)?
- if so, isn't that the wrong approach?

PeterV



Thanks for playing. To unsubscribe from this group, send an email to:
facetedclassification-unsubscribe <at> yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.


Thanks for playing. To unsubscribe from this group, send an email to:
facetedclassification-unsubscribe <at> yahoogroups.com



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.