[Sigia-l] tagging versus taxonomy

[Linda Sutherland]

Hello,

I'm delurking (briefly, I promise!), so I'd better introduce myself, 
I suppose. I'm an ancient-ish librarian, trained in pre-computer 
days, now working as a book indexer.


At 08:49 16/10/2006 +0200, Zbigniew Lukasiak wrote:
>I am talking about fundamental teoretical
>differences between those two models of classifying thigs and their
>consequences in the cognitive process of both classifying and 
>finding
>something previously classified.  With tags you can specify compound
>terms like 'tag1 and tag2 and tag3' while with taxonomy, as I
>understand it, you would need a separate name for each of those 
>terms.

I think that what you're asking about is what's called, in 
librarians' jargon, 'pre-coordinate'and 'post-coordinate' indexing. 
If I'm right, then I fear you are indeed reinventing the wheel!

I suspect you may also be conflating what traditionally are regarded 
as two separate steps in the indexing process - choice of terms, and 
their combination. The first of these is always done by the indexer, 
but combination of terms can be done either by the indexer, or by the 
searcher.

'Pre-coordinate' indexing implies that terms are combined by the 
indexer (or classifier) to create a more-or-less specific description 
of the subject of a document. For example, 'science bibliographies' 
might be used to describe bibliographies of scientific literature.

In post-coordinate indexing, the indexer 'tags' a document with 
whatever individual terms are thought necessary to describe the 
subject fully (e.g. 'science' and 'bibliography'), but doesn't 
attempt to combine them. It's up to the searcher to put together a 
string of terms. In this case, the combination will describe 
information sought, which may or may not equate to information 
existing, in the database or whatever is being searched.

Pre-coordinate indexing usually implies a classification scheme or 
taxonomy, in the sense of a standard that tells the indexer the 
preferred term(s) and rules for combination. Post-coordinate indexing 
may draw on a standardised, structured vocabulary for choice of terms 
(but doesn't have to), and any rules governing combination are 
applied by the searcher, not the indexer.

That's the theory. I should add, though, that as far as choice of 
terms is concerned, the distinction between pre- and post-coordinate 
isn't always clear-cut. Post-coordinate indexing in particular is 
prone to making use of compound terms, implying some degree of 
pre-coordination.


 > In other words  taxonomy is a language where you need a word for
 > everything while tags are words plus some rules about how to 
combine
 > them.

In my mind, the distinction is rather that in taxonomies (or 
classifications), both vocabulary and rules are built into a 
pre-defined structure of concepts, relationships and terms. They're 
essentially pre-coordinate systems. Tags are more like terms in 
post-coordinate indexing, selectable and combinable in ways specific 
to the individual context.



Linda Sutherland
linda.sutherland at britishlibrary.net