We've recently been working through a similar question on a project and are
coming to the conclusion that keywords are not be a good way to manage
internally-used categorizations like this.
It basically came down to this: keyword searches tend to be more liberal,
returning more things rather than less and when building an internally used
categorization of content, our users want to be able to control quite
precisely what content appears on what page.
For example, say I have keywords:
"administrative"
"administrative assistant"
and I search on "administrative" -- I'll get back content associated with
BOTH "administrative assistant" and "administrative." In other words, it
doesn't do exact string matches. The above is a bit hokey as an example, but
it means you have to be very careful about building keyword hierarchies
because you may end up getting back stuff that you don't expect.
This is understandable if you are doing a public search and want to return
more rather than less to the user.
However, when you have internally-used categories that are being used to
make things show up on pages or in areas of a page, you often want to be
very precise.
Also, you can also get into some ambiguities when specifying more than one
category in a search that can lead to you getting back more than you expect.
Again, there is a lack of precision that may be fine for public searching
for content but doesn't work when you are talking internal functions that
need to clearly identify which content is going to appear on which page(s).
In short, the Verity metadata search has proven to be a bit to liberal for
us when trying to build internal categories of content that control exactly
where the content appears -- in which case, the option of building options
and using UUIDs that point to those objects is probably a better way to go
david aden
-----Original Message-----
From: Raymond Camden [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, April 10, 2001 12:16 PM
To: Spectra-Talk
Subject: RE: follow up: Metadata or embedded objects
> 1: create a organisation unit object and store the UUID of the
> organisation
> unit in the news object. I make the UUID sql searchable to find objects
> which belong to an organisation unit.
>
> 2: I create a metadata hierarchy with the organisation units and
> assign the
> news object to the particular unit. I do a metadataobjectkeyword search to
> find objects which belong to an organisation unit.
>
> Can you give me comments on which approach to use and why one of there
> approaches is better?
I am assuming that you may have more info about the unit. If that is true,
then you probably want to go w/ method 1, where the UUID of the unit points
to a real object.
In general it comes down to this - you want to associate the article with a
unit. If you plan on having a Unit content object type, then method 1 makes
sense, because units will have more info.
If your units only exist as simple strings, then metadata would work fine.
Does that make sense?
=======================================================================
Raymond Camden, Principal Spectra Compliance Engineer for Macromedia
Email : [EMAIL PROTECTED]
ICQ UIN : 3679482
"My ally is the Force, and a powerful ally it is." - Yoda
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Structure your ColdFusion code with Fusebox. Get the official book at
http://www.fusionauthority.com/bkinfo.cfm
------------------------------------------------------------------------------
To Unsubscribe visit
http://www.houseoffusion.com/index.cfm?sidebar=lists&body=lists/spectra_talk or send a
message to [EMAIL PROTECTED] with 'unsubscribe' in the body.
