Metadata, Metadata, Metadata. Data about the data.
I once saw Inmon, the holy prophet of the data warehouse,
talk about metadata. The guy went on and on about the stuff
for two hours.
What he said was that metadata does not apply just to a data
warehouse. It is the guts of any database.
Successful projects I work on know what each
attribute and entity mean as a semantic definition,
how to create, update, delete and archive the
model elements. How elements change over time.
This is in addition to the language definition of
the model element and its relation to other elements.
There is a lot of modeling software on the market to
help track metadata. Use it or not, there is no excuse not
to have a reasonably complete model.
Dangerous projects are interfacing one application
to another or migrating a legacy application to an ERP
application. A first step for a DBA is to find and
secure the metadata of each model involved.
Having the metadata means that you can interface applications,
migrate legacy models to new models, plan resource allocation
and manage the data with no surprises.
Project managers and companies who do not understand
metadata and its role are always in some crisis mode.
Nothing they do will work or be delivered on schedule.
Avoid these people.