Just like using source control elsewhere, if you modify the code outside of source control - there's not going to be any tie-in/history in the Source Control system. That's true here too.

In this case - I've found that you can get around the human factor by simply putting up barriers to entry. Meaning- if a developer has access to everything, then it's altogether too easy for them to go in and "just update something" without following protocol; if, on the other hand, the only thing they see is their own local DEV instance, and all other environments are "limited permissions" to them, then the only way things gets moved into UAT, and then into Prod, is through checking in code.

Still - it's ultimately no different from using Source control for .NET code, or any other language. You have to respect the SCC solution's place, or it's worthless.