Does the price include
VAT or not? Is that price for wholesale or retail dealers? Which department has
established it? When was it modified last? These, and many other questions, may
stay behind data, simple and clear only in appearance.
In our mind, data always exist together with their metadata: when we look at prices
of goods at the grocery store, we know the answers to all the questions, because
we are aware of the context (goods in a little retail store). But when we look at the
column ???price??? in a table and nothing else, the context is unknown, and data may
have several different meanings. In the case of ???enterprise data warehouse??? there
could be a lot of information/metadata to manage, so the need for metadata management
is clearly visible. One must know the exact meaning of any single data item
contained in DW, otherwise one cannot do any reliable analysis (that is the goal of
a DW). In other contexts, where the domain is limited and well known, the need
2 Adzic, Fiore, & Sisto
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of
Idea Group Inc. is prohibited.
of metadata is less relevant because the operators are familiar with their own data
(e.g., in a system that collects Telco measures, it is useless to specify what column
???erlang??? means). In these cases, the data are stored in the DBMS and metadata could
continue to stay in the mind of the persons.
Pages:
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193