Less burden on reviewers, less guess work on the contributing person (ergo me). Also automation in supporting IDEs with automatic fixing/adapting as needed.
It looks like the current code looks mostly like PSR-2. Is that intentional?
(Extracted out of #16)