[Sigia-l] Classification is an essential skill

[Jonathan Broad]

Once more into the breach!

On Tue, 2003-02-04 at 01:30, Derek R wrote:
 	 
> Jonathan Broad wrote:
> >>| Why concentrate solely on the idiosyncratic?
> Christopher Gomez wrote:
> >| I'm not trying to drag anybody in mud
> 
> There's more truth to your statement then you may realize, Christopher,
> since 'nobody' is endorsing the idiosyncratic. My statement was:
> 
> >| Pay some attention to your users
> >| and not just the numbers 
> 
> which is not a dictum in favor of idiosyncratic properties, but a plea
> for common-sense. Try pinning your troubles on somebody else.

Yours is not a plea for common sense, but a straw man attack on an
illusory position (the 'categorical-cure-alls' as you say below).  

Unless--are you then saying that common sense includes some
non-idiosyncratic properties of things?  I.e. common properties?  You
can't have your cake and eat it too, here.

If you refuse to yield that categories (the 'non-idiosyncratic') are a
helpful way of organizing information for use, then you must admit to
holding a position of 'idiosyncratic-cure-all'.  Which is, as you
acknowledge as well, is idio-tic (etymology intended).  Otherwise, admit
you're wrong and join the rest of us died-in-the-wool 'common-sense
sometimes-categorizers sometimes-idiomizers'.

> >| We absolutely *cannot* cater 
> >| to everybody's quirks and fancy
> 
> Why would you want to? That's crazy and would never work. 

Quite.

> That's what
> the 'categorical cure-all' aims to do -- to cater to 'everyone's'
> understanding -- which is impossible without force. 

Not at all, not at all.  Neither 'everyone' nor 'just one' are the
proper targets of categorization, or at least, realistic ones.  

Categorization is a network of compromises between the aspects of
information (about things), and the needs/tasks/points of view of the
users of the categories.  Good categories don't pretend to be identical
to a user's unique point of view, but try very hard to be *pliant* to
that point of view.  Usable, not totally transparent.  

It's quite possible that a given set of categories can be satisfactory
for a large number of users, particularly in narrow problem domains.  

Put another way, just because a hammer doesn't conform perfectly to the
palm of your hand doesn't mean you can drive a nail with it.

> You should note the
> peer-to-peer (P2P) solution allows individuals to categorize themselves
> ('category of the person') and does not attempt to force commonality
> onto what is inherently a 'quirks and fancy' environment anyway.

I've already debunked this contention, but your insistence puzzles me. 
As a user of such systems, (albeit a librarianish one), the least usable
aspect of it is the truly ad-hoc naming conventions people use on their
files. A suprising number of people go through the effort of applying
good ID3 metadata to their files, though, which helps a lot.

It's actually a good case where more metadata would allow for more
productive serendipity, as you could expand searches to include artist
and genre more easily.  Now we make do with full-text searches and
browsing collections by IP when we find something interesting. 

> But -- you are right in that real 'architecture' is defeated by these
> dissections/reductions. Web users are very aware their purpose has been
> defeated (they are alienated). Now try and get the IAs on this list to
> realize that. The decline of information continues and IAs would prefer
> not to notice or take any responsibility...

Your usual bluster aside, did you catch the link I sent to PeterMo
yesterday morning?  

http://www.consumerwebwatch.org/news/report3_credibilityresearch/stanfordPTL_part1.htm

Sure, users are become less trusting of internet content.  Funny thing
is, what makes them more trusting is...good navigation!  How much of
that navigation involves categories, I wonder??

Jonathan