[Tutor] functions first?

[Alex Kleider]

On 2015-01-26 22:30, Ben Finney wrote:
> Alex Kleider <akleider at sonic.net> writes:
> 
>> Please correct me if I am wrong, but I've assumed that it is proper to
>> define all functions before embarking on the main body of a program.
> 
> I would say rather that as much code as possible should be in small
> well-defined functions, with the “main body” a tiny and trivial call to
> a function.
> 
> That way, all the code is easily tested by unit testing tools.
> 
>> I find myself breaking this rule because I want to set the default
>> values of some named function parameters based on a configuration file
>> which I have to first read, hence the need to break the rule.
> 
> Module-level constants are fine, and they obviously need to be bound
> before the definition of the function which uses them for parameter
> defaults.
> 
> But if they're not constants – as implied by your statement you read
> them from a configuration file – then they should not be in the 
> function
> definition, because reading the configuration file should itself be
> encapsulated in a well-tested function.

Thanks for your well reasoned advice.
You are of course correct in that if they are read from a configuration
file then they are not constants but in a sense they are, at least for
the life of the current program run.  I'm probably going to go with the
way Steve recommended which is to use global dictionaries to contain
the command line arguments (which docopt already does) and configuration
values (which are also placed into a dictionary by configparser) and 
then
use those dictionaries within functions that need them
rather than setting default named parameters for the functions.