[Sigia-l] Redundant information and content

[Paul Ford]

Peter Merholz wrote:

> However, even for smaller content stores, I've thought that faceted
> is the way to go. Let's say you're a financial services site, and of
> the 40 product types you have, there are car loans. Should such a
> thing be found under "car" or "loans"? Why not both? Why not let
> some people find things by "car" and others by "loans"?

Hi, Peter.

What I have been trying to do lately is have a default hierarchy which
provides a basic taxonomy, then create alternative taxonomies which
connect to the default hierarchy. Thus you have infinite paths through
a data set, but there is a single hierarchy to start with.

Why do this? Because it makes it much easier to keep things in mind
when you're creating and editing content, particularly as a group, and
because you're arranging things in a hierarchy you inherit a lot of
metadata accordingly, which is handy; the "car" node "knows" it's a
form of transportation. This can save you a great deal of work as you
build alternative taxonomies.

The danger of the "every node is at equal value, just add metadata" is
seen in Wikis: a lot of Wikis have pages without inlinks--the
connection is gone and thus the page is adrift. Having a
simple-as-possible hierarchy to hold all of the content and then
rearranging it as you see fit via metadata links to other structured
taxonomies means no page lack context outright. It's less "pure" but
it is probably easier to manage and maintain than node soup. As long
as every node has a unique ID it's easy enough to move things around
if you want to change your faceting approach, both in the default
hierarchy and the alternative taxonomies.

Paul E