It's expensive to scale by adding more DB servers. It's cheap to scale by adding more webservers or 'middle tier' servers. You can't just completely ignore th

It's expensive to scale by adding more DB servers. It's cheap to scale by adding more webservers or 'middle tier' servers. You can't just completely ignore the scaling argument by saying "it's no problem, just spend money." Lots of other things are fairly hand waving. Stored procedures can be version controlled... if you write some code to do it yourself, or buy RedGate's tool for that. Still, I've never worked in a shop that did as good a job at version controlling their SQL as they did their code. It's harder. Stored procedures can be refactored... but I'm not sure what tool is being used, and I don't want to have to evaluate 20 different tools to find the good one. I know I currently have better tools for writing compiled code than I do for writing database sprocs. There are many valid reasons for doing as little as possible in the DB layer, most of which are listed here, and I just don't see this article as successfully debunking them.