topic Re: Just testing you in Visual LISP, AutoLISP and General Customization Forum

Just testing you

john.uhden — Fri, 11 Jun 2021 17:38:01 GMT

What's wrong with this code?

(defun factorial (n / product) (setq i 0 product 1) (repeat n (setq i (1+ i) product (* i product)) ) product )

Re: Just testing you

pbejse — Fri, 11 Jun 2021 18:06:10 GMT

@john.uhden wrote:

What's wrong with this code?

It works as expected. <--- it does have a limit, like most things in lisp that deals with integers

(defun factorial (n / product) (setq i 0.0 product 1) (repeat n (setq i (1+ i) product (* i product)) ) (fix product) )

Re: Just testing you

john.uhden — Fri, 11 Jun 2021 19:16:22 GMT

@pbejse
You sorta beat around the bush with that answer, but You deserve an A+ for
demonstrating that you totally understood. Yes, I saw that you changed one
of the operands to a 'REAL.
Seems like even long integers have their limit in AutoLisp.

Anyone else?

Re: Just testing you

hak_vz — Fri, 11 Jun 2021 19:25:35 GMT

(defun factorial (n / product) (setq i 0 product 1.0) (repeat n (setq i (1+ i) product (* i product)) ) product )

Yes. Integers have limits in autolisp too.

Re: Just testing you

dbroad — Fri, 11 Jun 2021 22:49:42 GMT

IMO, the ideal factorial function should be recursive, not iterative. It is the paradigm for a recursive definition. If the user wants a real result, then, he should provide a real argument, else, he should accept the limitations of integers.

(defun factorial (n) (if (= n 1) 1 (* n (factorial (1- n)))))

Since its only defined for positive whole numbers(integers in real math), that is all you should need to do.

Re: Just testing you

john.uhden — Fri, 11 Jun 2021 23:24:22 GMT

@Anonymous
Whoa! I ain't never been too good with recursion.
Gotta study that and probably give you an A++.
I will do a time test, of course.

Re: Just testing you

dbroad — Sat, 12 Jun 2021 12:21:19 GMT

Thanks but time tests are probably irrelevant. Factorial results increase faster than exponentially. You probably won't want a number larger than 64 as input.

Re: Just testing you

john.uhden — Sat, 12 Jun 2021 21:20:05 GMT

Next question:

What's the largest integer that AutoLisp can handle?

Re: Just testing you

Jeff_M — Sun, 13 Jun 2021 16:02:46 GMT

@john.uhden wrote:

Next question:

What's the largest integer that AutoLisp can handle?

2147483647

https://help.autodesk.com/view/ACD/2015/ENU/?guid=GUID-EF6114FC-F1E4-4C71-91CC-07D01E6C8ABB

Re: Just testing you

john.uhden — Sun, 13 Jun 2021 16:19:22 GMT

Hi, Jeff.
Well done, even though you cheated.
Didja notice it's one shy of (expt 2 31)?
which turns negative and (abs) can do nothing to it.

Re: Just testing you

martti.halminen — Tue, 15 Jun 2021 11:08:30 GMT

This one also works in Common Lisp:

CLO 17 > (factorial 92)
12438414054641307255475324325873553077577991715875414356840239582938137710983519518443046123837041347353107486982656753664000000000000000000000

- The trick here is that CL has bignums when it runs out of space on integers.

- Also demonstrates the problem with recursive functions, on 93 I get:

Stack overflow (stack size 15998).
[Condition of type CONDITIONS:STACK-OVERFLOW]

Restarts:
0: [CONTINUE] Extend stack by 50%.
1: [ABORT] Return to top loop level 0.
2: [ABORT] Quit process.

So if I really needed larger ones, I'd need to change to another algorithm (iterative or recursive with memoizing or using an accumulator).