| Re: what is Floating illegal operand question? [message #43676] |
Tue, 19 April 2005 04:51 |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
Mike and James,
Thanks for the clarification. I was thrown off by the emphasized on floating
numbers on the mathworld web site. Deep inside it did not make any sense, so
I blamed it on the lecture I mussed have missed in college. :-)
Mike's explanation makes perfect sense, so thanks for straightening me
out...
Cheers,
Haje
"Michael Wallace" <mwallace.no.spam@no.spam.swri.edu.invalid> wrote in
message news:1168aupboi2v87b@corp.supernews.com...
> (0 / 0) is just as undefined as (0.0 / 0.0) is. However, there is no way
> to express NaN or Infinity when working with integers. The IEEE floating
> point specification allows for NaN and Infinity, but the integer spec does
> not. Without a NaN option, IDL (and other languages as well), default to
> evaluating (0 / 0) as 0 even though this is technically incorrect.
>
> -Mike
>
>
> Haje Korth wrote:
>> Well, if I understand this right, 0/0 is defined and the result is 0, so
>> that the division by zero message is only informational. On the other
>> hand 0./0. is not defined for floating point numbers and results in an
>> error (see http://mathworld.wolfram.com/DivisionbyZero.html). Correct me
>> if I am wrong.
>>
>> Haje
>>
>>
>>> You're getting it from a floating point divide by zero. When using
>>> integers, IDL actually tells you that you tried to divide by zero. For
>>> some reason, they have a different message for floating point.
>>>
>>> IDL> x = 0 / 0
>>> % Program caused arithmetic error: Integer divide by 0
>>> IDL> x = 0.0 / 0.0
>>> % Program caused arithmetic error: Floating illegal operand
>>>
>>> -Mike
>>
>>
|
|
|
|
|
|
|
|
|
|
| Re: what is Floating illegal operand question? [message #43689 is a reply to message #43687] |
Mon, 18 April 2005 14:51  |
Michael Wallace
Messages: 409
Registered: December 2003
|
Senior Member |
|
|
(0 / 0) is just as undefined as (0.0 / 0.0) is. However, there is no
way to express NaN or Infinity when working with integers. The IEEE
floating point specification allows for NaN and Infinity, but the
integer spec does not. Without a NaN option, IDL (and other languages
as well), default to evaluating (0 / 0) as 0 even though this is
technically incorrect.
-Mike
Haje Korth wrote:
> Well, if I understand this right, 0/0 is defined and the result is 0, so
> that the division by zero message is only informational. On the other hand
> 0./0. is not defined for floating point numbers and results in an error (see
> http://mathworld.wolfram.com/DivisionbyZero.html). Correct me if I am wrong.
>
> Haje
>
>
>> You're getting it from a floating point divide by zero. When using
>> integers, IDL actually tells you that you tried to divide by zero. For
>> some reason, they have a different message for floating point.
>>
>> IDL> x = 0 / 0
>> % Program caused arithmetic error: Integer divide by 0
>> IDL> x = 0.0 / 0.0
>> % Program caused arithmetic error: Floating illegal operand
>>
>> -Mike
>
>
>
|
|
|
|
| Re: what is Floating illegal operand question? [message #43690 is a reply to message #43689] |
Mon, 18 April 2005 14:24 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
Haje Korth wrote:
> Well, if I understand this right, 0/0 is defined and the result is 0, so
> that the division by zero message is only informational. On the other hand
> 0./0. is not defined for floating point numbers and results in an error (see
> http://mathworld.wolfram.com/DivisionbyZero.html). Correct me if I am wrong.
0/0 is exactly as undefined for integers as 0.0/0.0 is for floating
point. To see why that is true, please try to identify a finite floating
point value for f that is NOT a solution to this equation:
0.0*f eq 0.0
Then try to find an integer value for i that is NOT a solution for
0*i eq 0
Is it any easier for the integer case than for the floating point case?
|
|
|
|
| Re: what is Floating illegal operand question? [message #43691 is a reply to message #43690] |
Mon, 18 April 2005 14:02 |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
Well, if I understand this right, 0/0 is defined and the result is 0, so
that the division by zero message is only informational. On the other hand
0./0. is not defined for floating point numbers and results in an error (see
http://mathworld.wolfram.com/DivisionbyZero.html). Correct me if I am wrong.
Haje
"Michael Wallace" <mwallace.no.spam@no.spam.swri.edu.invalid> wrote in
message news:11687a01efaaa29@corp.supernews.com...
> lixiaoyao wrote:
>> when I run this code in idl,it generated the following wrong
>> information,what is
>> wrong there?
>> % Compiled module: $MAIN$.
>> % Program caused arithmetic error: Floating illegal operand
>> window,xs=640,ys=480
>> x=10*!pi*findgen(1001)/1000
>> sn=sin(x)/x
>> plot,x,sn
>> end
>> thanks
>>
>
> You're getting it from a floating point divide by zero. When using
> integers, IDL actually tells you that you tried to divide by zero. For
> some reason, they have a different message for floating point.
>
> IDL> x = 0 / 0
> % Program caused arithmetic error: Integer divide by 0
> IDL> x = 0.0 / 0.0
> % Program caused arithmetic error: Floating illegal operand
>
> -Mike
|
|
|
|
| Re: what is Floating illegal operand question? [message #43692 is a reply to message #43691] |
Mon, 18 April 2005 13:49 |
Michael Wallace
Messages: 409
Registered: December 2003
|
Senior Member |
|
|
lixiaoyao wrote:
> when I run this code in idl,it generated the following wrong
> information,what is
> wrong there?
> % Compiled module: $MAIN$.
> % Program caused arithmetic error: Floating illegal operand
> window,xs=640,ys=480
> x=10*!pi*findgen(1001)/1000
> sn=sin(x)/x
> plot,x,sn
> end
> thanks
>
You're getting it from a floating point divide by zero. When using
integers, IDL actually tells you that you tried to divide by zero. For
some reason, they have a different message for floating point.
IDL> x = 0 / 0
% Program caused arithmetic error: Integer divide by 0
IDL> x = 0.0 / 0.0
% Program caused arithmetic error: Floating illegal operand
-Mike
|
|
|
|
| Re: what is Floating illegal operand question? [message #43693 is a reply to message #43692] |
Mon, 18 April 2005 13:30 |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
The first element of sn is 0./0., which is not defined. Haje
"lixiaoyao" <lixiaoyao5880@yahoo.com> wrote in message
news:1113855953.629412.311210@f14g2000cwb.googlegroups.com.. .
> when I run this code in idl,it generated the following wrong
> information,what is
> wrong there?
> % Compiled module: $MAIN$.
> % Program caused arithmetic error: Floating illegal operand
> window,xs=640,ys=480
> x=10*!pi*findgen(1001)/1000
> sn=sin(x)/x
> plot,x,sn
> end
> thanks
>
|
|
|
|
| Re: what is Floating illegal operand question? [message #43694 is a reply to message #43693] |
Mon, 18 April 2005 13:34 |
Haje Korth
Messages: 651
Registered: May 1997
|
Senior Member |
|
|
AFAIK, IDL does not evaluate any limits. Haje
"lixiaoyao" <lixiaoyao5880@yahoo.com> wrote in message
news:1113855953.629412.311210@f14g2000cwb.googlegroups.com.. .
> when I run this code in idl,it generated the following wrong
> information,what is
> wrong there?
> % Compiled module: $MAIN$.
> % Program caused arithmetic error: Floating illegal operand
> window,xs=640,ys=480
> x=10*!pi*findgen(1001)/1000
> sn=sin(x)/x
> plot,x,sn
> end
> thanks
>
|
|
|
|
| Re: what is Floating illegal operand question? [message #43695 is a reply to message #43693] |
Mon, 18 April 2005 13:39 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
lixiaoyao writes:
> when I run this code in idl,it generated the following wrong
> information,what is
> wrong there?
> % Compiled module: $MAIN$.
> % Program caused arithmetic error: Floating illegal operand
> window,xs=640,ys=480
> x=10*!pi*findgen(1001)/1000
> sn=sin(x)/x
> plot,x,sn
> end
IDL> Print, Sin(0)/0
-NaN
% Program caused arithmetic error: Floating illegal operand
Not sure you want to be dividing by zero. :-)
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
