-=:: FBSL ::=-

I tried to improve my Range() function inspired by Killrami.





I'm not very satisfied with my script. I think that it looks like a patched-up job. Without a doubt, one can do it better.

Dear Bertrand,

I am afraid, the code which you propose here IS NOT directly inspired by Killrami's question. His question was about Python's version of the Range() function while your code behaves more like the PHP v4/5 version of Range().

IMO Python's function expects its parameters in the following generic order:

1. input: ascending order (start, stop, +step) whereby stop >= start. Also, the console printout of a special case of 1 parameter e.g. Range(10) will be "0,1,2,3,4,5,6,7,8,9" but not "1,2,3,4,5,6,7,8,9,10" as your function does;

2. input: descending order (start, stop, -step) whereby stop <= start.

In other words, the enumeration order (ascending or descending) is explicitly determined by the sign of the step parameter and if the magnitude of the start and stop parameters does not match the step sign used, the function will fail. This algo, of course, will have a simpler (perhaps also faster) implementation than you suggest.

Your function behaves like its PHP's counterpart: step is positive and its equivalent sign is determined by whether start is smaller or greater than stop. Clearly enough, this algo requires more checks and substitutions and is perhaps always slower than the Python-like implementation.

So which of the two implementations are we going to optimize?

Mike

Yes, I didn't really go further into the question. It was just a first draft
I didn't remember how works the Python's function Range(). In fact I hesitated over the decision to start with 0 or 1 (my Logo experience decided in favour of 1, probably wrongly).
I admit to having struggled to find a solution for the descending order and I lazily chose the easier one for me.
And what is more, I don't know anything about PHP.


Personaly, I would opt for the simpler and faster implementation or for a solution which satisfies Killrami's request in a better way (so, not my solution) ...

OK, with all this in mind, first we'll work out the design rules for our more-or-less optimized Python-like implementation of Range() function:

1. Macros are generally handy but not here: PrintItem(item) would require an additional Print to go to a new line after the entire array contents are printed out while Array_Print(vector) would imply a call to an extra function which means additional time overhead for setting up Map() calls;

2. The implementation should provide for a special case with only 1 parameter given which is essentially the Range(end - start) parameter with a step = 1;

3. The generic parameter format is Range(start, end, (+/-)step) with start <= end for a positive step and start >= end for a negative step while any other parameter set (barring case 2 above) is illegal.

4. Array_Print is essentially an application-specific feature. Other tasks may require other ways to intercept/handle/respond to invalid parameter values or combinations so "If Count(params) = 0 Then" may need to be logically Or'ed accordingly with any other invalid conditions to produce the desired result for early error detection in the caller code.

Following these guidelines, I can suggest the following implementation:


Does this implementation still produce a "patched-up job" feeling?

Mike

Hey Mike !!!
Simply A M A Z I N G !!!

Absolutely not ! ... really clearer.

Thanks for the lesson.
I like your 'Array_Print' better than mine. I was too much tempted to play with the Map function (one more conditionning by Logo or Lisp-like languages)