[PEP] Checking available() on an unavailable solver should not throw an exception by default

[blnicho]

This PEP proposes changing the default behavior of calling the solver available() method. As was pointed out in #1064, the default behavior is to raise an exception if the solver is not found. This behavior can be changed by setting the exception_flag=False option. I propose that exception_flag=False become the default for the available method. This method is only called when there is uncertainty if the solver is available and I think it's counter intuitive for an exception to be raised when a programmer is calling something as a defensive measure to avoid future exceptions.

[jsiirola]

As clarification, are you proposing that we change the default value for the exception argument, or both change the default behavior and deprecate/remove the argument entirely?

[blnicho]

I'm proposing changing the default value for the argument

[jsiirola]

I am definitely in favor of changing the default, and would go even farther and advocate complete removal of the argument.

[ghackebeil]

I support changing the default, but I support deprecating the argument for a while before removing it (e.g., check if it is in **kwds). I know it’s not something that is in the book, but I am aware of a number of external projects that explicitly set the argument to its non-default value.