LYDB: What advice about stored procedures and other server side code?

I am putting together some advice for developers about getting the most out of SQL servers in general and Postgres in particular. I have in mind the likes of most web developers, who through ignorance or a strange cultural preference that has emerged, tend to treat their database server as a dumb data bucket.

I call the project Love Your Database (LYDB). It is starting as a series of blog posts:

https://medium.com/@gisborne/love-your-database-lydb-23c69f480a1d#.4jngp2rcb <https://medium.com/@gisborne/love-your-database-lydb-23c69f480a1d#.4jngp2rcb>
https://medium.com/@gisborne/love-your-database-simple-validations-68d5d6d0bbf3#.az4o2s152 <https://medium.com/@gisborne/love-your-database-simple-validations-68d5d6d0bbf3#.az4o2s152>

I would next like to cover server-side code such as stored procedures and triggers.

I am inclined to advise folks to use PL/V8 on Postgres, because it is a reasonable language, everyone knows it, it has good string functions, decent performance and it tends to be installed everywhere (in particular, Amazon RDF offers it).

Broadly, what advice should I offer that isn’t obvious? Not just about PL/V8 but server side code in general.

TIA