•Jones
proposed 2 changes to traditional approach
–Build
an EDW a subject area at a time, loading data into
the first subject area design as soon as technically feasible,
and iterate collection of feedback from users thereafter
(weeks, not months).
–Abandon
the mandate of building a large, comprehensive, data warehouse prior
to building star schemas for individual subject areas.
•Prove
value earlier, reconcile onerous issues early, leverage
advances in tools and software.
•Dimensional
EDW, done as prescribed, actually leads to less redundancy and improved consistency
without a 3NF data warehouse.
–