On Feb 06, 2014, at 3:03 PM, Cory Casanave wrote:
I really don't care much how people persist and exchange their data (or metadata) as long as the model is sound.
Assuming we're talking about systems here, what happens when there are not models? Or the models are incomplete/inaccurate?
Is there any sort of guesstimate as to what percentage of systems actually have worthwhile, accurate models?
I think I'm on firm ground that very, very, very few firms have any sort of easily accessible, accurate, complete, current metadata.
Since metadata is pretty slippery stuff, I'm more than happy to quote that over the past 50 years, the survival rate of data dictionary/metadata repository efforts in Fortune 500 organizations has been under 5%.