Why this is in my collection

From the publisher:

Data models are the main medium used to communicate data requirements from business to IT, and within IT from analysts, modelers, and architects, to database designers and developers. Therefore it's essential to get the data model right. But how do you determine right? That's where the Data Model Scorecard(R) comes in.The Data Model Scorecard is a data model quality scoring tool containing ten categories aimed at improving the quality of your organization's data models. Many of my consulting assignments are dedicated to applying the Data Model Scorecard to my client's data models - I will show

Highlights

  • Model quality can be scored, so it can be governed — Hoberman's ten-category scorecard turns "is this a good model?" from a matter of taste into an inspectable rubric; rules applied to the model layer itself, one level up from data-quality rules on the data.
  • Correctness and completeness are different failure modes — a model can follow every notation rule and still not capture the business, or capture the business in a structurally broken way; the scorecard forces both checks explicitly instead of letting one masquerade as the other.
  • "Does the model match the requirements" is category one for a reason — the most expensive modeling defects are fidelity defects, not notation defects; a reminder that generation amplifies whatever the model says, wrong or right, so fidelity review must come before compilation.
  • Naming and definition standards are load-bearing — the scorecard treats consistent naming and rigorous definitions as scored quality dimensions, not cosmetics; in a model-driven pipeline the name and definition become the generated artifact, so sloppiness propagates mechanically.
  • A shared rubric depersonalizes model review — scoring against published categories converts review from opinion-clash into evidence discussion, the same move a validation framework makes for data checks; structure defuses the ego problem.
  • Scorecard categories are automation candidates — several of the ten categories (structural soundness, standards adherence, readability) are mechanically checkable, which makes this book a requirements list for model-linting inside MDDE tooling.

Highlights on this page are generated with the help of AI.