If I want readers to switch off (it can be useful; one can sometimes sneak something controversial past them), I usually chuck in a couple of things about metadata.
But 'data about data' is important, especially in a business context. I have no problem with risk-based analyses using lots and lots of comparatively low quality data, but unless you have some understanding of exactly what the data actually means and what it is trying to tell you, you risk making some pretty silly decisions. Now, Embarcadero has given us CONNECT, a new metadata governance platform that promises "to make database tools smarter and users experts".
Hmmm, not sure about the feasibility of that last one but I like the idea of a tool that gives your data a business context, as you use it, and adds collaboration facilities. What this all means is that CONNECT provides an "Enterprise Information Map" (a single, searchable registry of all available data sources). It promises you:
- an Enterprise Glossary - which views, stores and centrally manages authoritative business definitions in an extensible enterprise glossary.
- a Data Source Registry - which creates a single searchable registry of all available data sources to reduce the time you spend searching for data.
- Social Enterprise Collaboration - which applies features such as activity and discussion streams, notifications and the ability to "follow" and "like".
- Inline definitions - which allow viewing of definitions for registered data elements or business terms directly in widely-used database tools.
- Semantic Mapping - which relates business terms to critical data elements including tables, columns, entities and attributes.
With a bit of luck, when you are aggregating data across a million customers, this'll warn you that what Private Banking calls a "customer" has somewhat different characteristics to what your payday loans dept calls a "customer".
Now, I must admit that I haven't had a chance to look at this in detail; but to me, it sounds rather like the sort of active Data Dictionary, integrated with production systems, that I once found so useful when I was in DBA in the mainframe environment used by an Australian government department. Welcome back!
And please take that "Welcome" absolutely at face value. Way back then, we might have had similar functionality, but I'd have killed for the sort of user experience and social collaboration possible today; to say nothing of the potential for using data from outside of the corporate database.