This was said years ago:

 ”Table (n.) – a collection of information (data?) describing a population of entities which possess some common characteristics, called attributes. -itis – “suffix denoting diseases characterized by inflammation, itself often caused by an infection.”  ---------- from the Wikipedia Wiktionary.”

Tables are the building block of relational databases. Tables must generally be “normalized,” at least to 1NF. That may be an appropriate way to think of databases when implemented in a modern day DBMS. However, it is not the way the world thinks logically. People have no problem with commonly occurring phenomena such as:

·         A multi-valued attribute, e.g., an Employee possesses multiple Skills.

·         Many-to-many (M:N) relationships, e.g., as between Employees and Projects

·         A relationship with attributes  

     This just now, on LinkedIn (check out my comments).

“Putting to one side the argument that your data almost certainly didn't start out broken out in to tables, and it almost certainly isn't consumed that way either, here's the thing; MongoDB, if you squint, is essentially a relational database with an unorthodox take on first normal form and some great high availability and scalability features.” -- Graeme Robinson