>With modern software design and especially OOP the risk of data inconsistency is less of a problem This sounds mighty hand-wavy. How does "modern software des

>With modern software design and especially OOP the risk of data inconsistency is less of a problem This sounds mighty hand-wavy. How does "modern software design" reduce the risk of data inconsistency? >On the other hand duplication of information is a great way to scale an application Most relational systems have facilities for duplicating information to scale. Materialized views, for instance, generate vile offenses to all normal forms but perhaps the first, hyper optimized for consumers, but it is guaranteed coherent and consistent, and happens with barely any work. Seriously, we've been solving these performance issues for years. Every time some, failing a better word, noob writes up their big internet paper on why the relational model fails (with "modern" software, which is chuckleworthy), the world gets just a little bit dumber. The bit about checking data for corruption is just disturbing.