Inmon vs Kimball
This is a question that receives a significant amount of resources, driven
almost in an ideological fashion in the backroom. One can spend hours arguing
the merits of normalised versus de–normalised design, and other differences
between these approaches.
The fact that there has been no winner in this debate tells us that there are
merits in both methodologies, and that a horses–for–courses approach is
preferable. If this theological war does occur within your project consider:
-
whether the products you use drive the design approach e.g. some products work
better with normalised or de–normalised data
-
whether the differences matter for your project
-
whether the team has experience in one approach
-
what the key business drivers, departmental or corporate are
-
what the delivery timescales are.
Read next »
