Bug/Error in the execution of Range(3) function

#1 Опубликовано: 28.11.2012 20:11:03

0 сообщений из 9 понравились пользователям.

Группа: User

There's a problem with the Range(3) function.

Here's an example:

1. Create a vector for a variable n using the range(3) function such as:

n:=0.03, 0.0325.. 0.05

This should produce a vector with 9 elements ranging from 0.03 to 0.05 in 0.0025 increments.

However, it actually produces an 8 element vector from 0.03 to 0.0475. It appears the last element (0.05) is truncated.

Why?

Thanks in advance,
Brett

Bug in Range(3) function, SMath 0.95.sm (5 КиБ) скачан 46 раз(а).

#2 Опубликовано: 29.11.2012 04:50:25

Radovan Omorjan

325 сообщений из 2052 понравились пользователям.

Группа: Moderator

Hello Brett,

I think it is due to the machine floating point rounding error. Here is the possible explanation of it.

[MATH=eng]n:range(0.03,0.05+10^{-16},0.0325)[/MATH]

[MATH=eng]n=mat(0.03,0.0325,0.035,0.0375,0.04,0.0425,0.045,0.0475,0.05,9,1)[/MATH]

Adding a rather small number to the last number (order of magnitude of computing precision ~15 decimal place) might solve this.

Regards,
Radovan

When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"

#3 Опубликовано: 30.11.2012 19:51:51

btowler

0 сообщений из 9 понравились пользователям.

Группа: User

Thanks Radovan! That worked (though I'm somewhat concerned that an error like this popped up in the first place).

Brett

#4 Опубликовано: 30.11.2012 20:41:26

Martin Kraska

1222 сообщений из 2150 понравились пользователям.

Группа: Moderator

I agree with Radovan that this feature is not exactly a bug. However, testing the loop exit condition with some tolerance halfway between the iterator increment and machine precision would be a default that might perhaps meet the intent of most users. Then ugly workarounds could be avoided.

Martin Kraska Pre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx

#5 Опубликовано: 01.12.2012 01:46:06

btowler

0 сообщений из 9 понравились пользователям.

Группа: User

I think the easiest (and perhaps cleaner) workaround is the alternate method I included in the attachment. By using a For loop with a dummy (integer) index variable (e.g., i) run through a range of 1 to 9, then under the For loop create the vector n in an equation using the index variable.

For i:=1..9
n:=0.03+(i-1)*0.0025

#6 Опубликовано: 01.12.2012 06:51:28

Martin Kraska

1222 сообщений из 2150 понравились пользователям.

Группа: Moderator

The same without for loop:

[MATH lang=ENG]0.03+0.0025*(range(0,8))=mat(0.03,0.0325,0.035,0.0375,0.04,0.0425,0.045,0.0475,0.05,9,1)[/MATH]

Both versions of the workaround obfuscate the upper loop limit, that's why I would like the range(3) function to be modified. However, the workarounds make the iterator increment more obvious. Would perhaps be fine to have an alternative range(3) with range(start,stop,step) displaying as

start..stop by step

Martin Kraska Pre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx

1 пользователям понравился этот пост

Davide Carpi 01.12.2012 09:08:00

#7 Опубликовано: 01.12.2012 09:09:39

Davide Carpi

1415 сообщений из 2872 понравились пользователям.

Группа: Moderator

Wrote
The same without for loop:

[MATH lang=ENG]0.03+0.0025*(range(0,8))=mat(0.03,0.0325,0.035,0.0375,0.04,0.0425,0.045,0.0475,0.05,9,1)[/MATH]

Both versions of the workaround obfuscate the upper loop limit, that's why I would like the range(3) function to be modified. However, the workarounds make the iterator increment more obvious. Would perhaps be fine to have an alternative range(3) with range(start,stop,step) displaying as

start..stop by step

I agree with Martin :d

If you like my plugins please consider to support the program buying a license; for personal contributions to me: paypal.me/dcprojects

Форум SMath

Bug/Error in the execution of Range(3) function - brettwtowler@yahoo.com - Сообщения