Coding guidelines are followed in safety critical systems, but to what extent is it necessary to adhere to them strictly? In particular if you want to use open source software.
“Guidelines” may mean something strict and official, but it may also mean a tool to help make decisions. So in code, how to streamline your code, make it more readable, avoid common mistakes. They are however often percieved as making your life more difficult and make no sense. Still, they can close language specification gaps (avoid undefined or implementation-defined behaviour), avoid obscure constructs, avoid language pitfalls, and reduce complexity. So the goal really is to make coding life easier and the result more reliable.
If you use coding guidelines for your project, you have to choose the right one. The guidelines must be applicable and help you achieve the above objectives. Preferably, you should adhere to a de-facto standard, perhaps adapt it to your needs. And finally check if there are (parts of) the guidelines that can be checked automatically.
Most common existing guidelines:
Blindly changing the code in order to follow the rules may actually break< it/em>. For example, a MISRA rule is that each if/elseif sequence must have a final else – if you add that final else later with an error handler, you may overlook that there was actually a valid case to arrive there.
[There was a part that I couldn’t follow here. Unfortunately it was probably a fairly important part.]
In conclusion, coding guidelines are needed but they don’t need to be followed 100%. It’s important to define upfront what will be the procedure for deviating from them. Also, if it turns out that some guidelines become too cumbersome, adapt them, maybe even add a new rule. Don’t forget to document all these decisions (why they were taken).