laba-4ic@WEB.berkeley.edu (byron c go) (07/23/90)
I got XScheme for the PC today and am relatively pleased with the following exception... >(fib 9) 2432 >(fib 10) -8448 (where fib is defined as (lambda (x) (if (= x 1) 1 (* x (fib (-1+ x)))))) it would seem that there is some kind of overflow error going on here. does anybody know of a patch to the source code that fixes this? perhaps a change in the representation of integers? thanks, byron ----- byron c go byron@QAL.berkeley.edu laba-4ic@web.berkeley.edu ab econ '92 ...!ucbvax!QAL!byron ...!ucbvax!web!laba-4ic "Help... I've become the anti-Dan_Kogai! (I've started writing Scheme code and translating it to C for my programming projects!)" - me
alfred@dutepp1.tudelft.nl (Alfred Kayser) (07/24/90)
Byron c go writes: >I got XScheme for the PC today and am relatively pleased with the following >exception... > (fib 9) => 2432 > (fib 10) => -8448 >(where fib is defined as (lambda (x) (if (= x 1) 1 (* x (fib (-1+ x)))))) ^^^ I know that fib and fac (faculty) are both used for testing the number system of Scheme implementations, but that is no reason to swap them. It should be: (define (fac n) (if (< n 2) 1 (* n (fac (- n 1))))) and (define (fib n) (if (< n 2) 1 (+ (fib (- n 1)) (fib (- n 2))))) Note the subtle difference. But returning to the original question: XScheme uses long integers (32 bits) and therefore will (fac 9) and (fac 10) overflow and return invalid results. They should be: 362880 and 3628800. >it would seem that there is some kind of overflow error going on here. Indeed. >does anybody know of a patch to the source code that fixes this? perhaps >a change in the representation of integers? This is not very simple. Big integers (unlimited length) are not simple to implement. >"Help... I've become the anti-Dan_Kogai! (I've started writing Scheme code > and translating it to C for my programming projects!)" - me The bignumber system as supported by most Scheme implementations (except XScheme) doesn't exist in C. Hope this helps. Alfred -- -- Ir. Alfred Kayser. PACS, OS/2, TCP/IP. --- Email: AKayser@et.tudelft.nl -- -- CARDIT, Delft University of Technology ------------ Tel: (31)-15-786179 -- -- P.O.Box 5031, 2600 GA Delft, The Netherlands ------ Fax: (31)-15-784898 --
grunwald@foobar.colorado.edu (Dirk Grunwald) (07/26/90)
>>>>> On 24 Jul 90 08:54:24 GMT, alfred@dutepp1.tudelft.nl (Alfred Kayser) said:
AK> The bignumber system as supported by most Scheme implementations (except
AK> XScheme) doesn't exist in C.
--
not true. libg++ for the Gnu G++ compiler provides classes Integer
(basically a bignum) and Rational. They're fast too.alfred@dutepp1.tudelft.nl (Alfred Kayser) (07/27/90)
grunwald@foobar.colorado.edu (Dirk Grunwald) writes: >>>>>> On 24 Jul 90 08:54:24 GMT, alfred@dutepp1.tudelft.nl (Alfred Kayser) said: >AK> The bignumber system as supported by most Scheme implementations (except >AK> XScheme) doesn't exist in C. >not true. libg++ for the Gnu G++ compiler provides classes Integer >(basically a bignum) and Rational. They're fast too. So it may be, but to my understanding is C++ not equal to C. No flames intended. -- -- Ir. Alfred Kayser. PACS, OS/2, TCP/IP. --- Email: AKayser@et.tudelft.nl -- -- CARDIT, Delft University of Technology ------------ Tel: (31)-15-786179 -- -- P.O.Box 5031, 2600 GA Delft, The Netherlands ------ Fax: (31)-15-784898 --