Re: Queries on Single Sourcing

["Mark Baker" <listsub -at- analecta -dot- com>]

kk wrote:

> The places I've worked that tried the chunking
> approach in a single-source procedure, they abandoned
> the attempt when it became apparent the overhead of
> cataloging the chunks was just too much of a problem.
...
> So how do you use information chunking without having
> to spend hours each day doing nothing but categorizing
> and indexing chunks?

If you design proper information components, you do not spend very much time
labeling them. A well designed information component has certain discrete
and consistent properties that naturally identify it. These properties are
both part of its data and part of its metadata. You spend your time creating
components. Most of the metadata takes care of itself.

The situation you describe sounds like the typical attempt to turn fragments
into usable components by adding metadata too them. The reason that the
cataloging becomes overwhelming is because no matter how much metadata you
add, you can't change a fragment into a component. It's like trying to turn
straw into gold by adding margarine. Adding more margarine doesn't help and
you soon get tired of carrying it.

Mark Baker
Analecta Communications