My programming background was in file systems, not databases. I’ve done a bit of relational theory and SQL practice but I’m far from being a database expert. When I decided to start learning to write in a modern coding language I faced a huge set of technical hills that I needed to climb. I picked Python to avoid needing ‘to web’ straight away and when I discovered NoSQL databases, they sounded like a great way to dismiss the mountain of mapping objects onto the relational model. This morning I started to have doubts.
I see a database being in at least 2 levels, the database engine and the query language. This seems like having different shells atop a Unix/Linux kernel. Why hadn’t the NoSQL crowd built a new NoSQL front-end onto an existing DB rather than starting from scratch? Was there a fundamental mismatch at all levels? Did data have to be split between 2 different DBs? A wise man once told me that within an organisation, the word “database” had no plural. My personal experience is definitely that everything is connected, because human organisations are networks not trees.
I just found this:
Does anyone who is using NoSQL tech think this article is wrong? One of the comments says that Postgresql won’t scale to 30,000 machines. I’m not sure that I need to do that right now.
I’ll chuck this in for good measure http://nosql.mypopescu.com/post/5449745504/document-databases-and-the-impedance-mismatch-with