The get*() methods that are used to access SmartDashboard can fail silently

[brettle]

I'm referring to methods like this one:

RobotCode2018/Robot2018/src/org/usfirst/frc/team199/Robot2018/OI.java

Lines 56 to 61 in 1b3045f

	public int getButton(String key, int def) {
	if (!SmartDashboard.containsKey(key)) {
	SmartDashboard.putNumber(key, def);
	}
	return (int) SmartDashboard.getNumber(key, def);
	}

Consider what happens if there is already an entry with the requested key but it is of a different type.

[brettle]

The get*() methods that are used to access SmartDashboard can fail silently

[dfang314]

It would just return the default. However, this should never be the case since the convention I already set (and tbh will most likely change, this is one great reason why it shouldn't change) is greatly specific ("such as 'Button/Shift Drive' or 'Port/LeftTalonSRXDrive' [the booleans are a separate matter that will be dealt with (hopefully) tomorrow to the existing standard]) therefore imo at least there is no worry about overlap of keys. (Sorry for so many parentheses so here's some more)

[brettle]

"this should never be the case" + "convention" + Murphy's Law = "this is bound to happen at some point"

My concern here is with the fact that there is no indication that this unexpected condition has occurred. Simply logging the error is sufficient.

[dfang314]

Added printing to Standard Error in last commit (currently in a pull request)