[Theforum] subcats vs. metadata/beginners vs. everyone

[Greg Sampson]

On Sun, Dec 16, 2001 at 10:34:48AM -0900, mccreath wrote:

> SUBCATS vs. METADATA

Just before I saw this post, I was thinking that a good metadata might
help out immensely in this situation. So it was good to find this;
makes me feel less crazy. But:

> It seems like we could create virtual subcategories using
> views. (You can create views on Oracle databases, right?) Say we run
> a query that gets all the articles in all categories with ASP in the
> body (as an extremely simple example) and build a view with
> that. Then just create a link on the 'Code' page that says 'ASP',
> which loads a page with just the 'Code'/'ASP' articles. That page
> would have a link that says something like 'See all ASP articles
> from all categories', which does exactly what it says, but separates
> the results by category (as opposed to the current search which only
> sorts by date).

Interesting idea. However, having the site's ASP category be defined
by a search of all the articles for the word ASP can get difficult, as
the search wouldn't be able to pick up any contextual meaning. So you
run the risk of not only returning an article on "here's why ASP is
inferior," (not much help for someone who's been told s/he has to
learn the scrpting language by tomorrow), but also one that might have
something to do with Application Services Providers (granted not many
people are talking about them any more, but I think you catch my
drift.)

So my point is this: Metadata might very well be a way to go, but
whoever implements the system will have to be very careful about
it. Otherwise it will wind up being a noisy, nonsensical mess.

--Greg