| Re: Yet another bug?? [message #4253] |
Fri, 19 May 1995 00:00 |
lmudge
Messages: 9
Registered: October 1994
|
Junior Member |
|
|
In article 95May17132737@peace.med.ohio-state.edu, phil@peace.med.ohio-state.edu (Phil) writes:
> Sorry to keep posting but I've just started developing some useful
> tools in IDL and it seems I run across a new 'bug' every day. Could
> someone tell me the sense behind the following:
>
> IDL> print, not 0
> -1
> IDL> print, not 1
> -2
>
> Shouldn't (not 0) = 1 and (not 1) = 0 or am i just missing something?
> Is there somehting in the manuals that describes this logic?
>
> Just wondering.
>
>
> --
> /*********************************************************** ****************/
> Phil Williams
> Postdoctoral Researcher "One man gathers what
> MRI Facility another man spills..."
> The Ohio State University -The Grateful Dead
> email: phil@peace.med.ohio-state.edu
> URL: http://justice.med.ohio-state.edu:1525
> /*********************************************************** ****************/
One must thing in binary terms to understand what is happening here.
Integer variables in IDL uses twos complement binary arithmetic thus the
binary representation of 0 is 0000000000000001. When we perform a not on this
we get 1111111111111110 which is the twos complement representation for -1.
Try doing this:
not float(1)
and
not float(2)
and see what you get.
Regards,
Leith Mudge
------------------------------------------------------------ --------------------
Leith Mudge
lmudge@awadi.com.au
AWA Defence Industries
PO Box 161
Elizabeth SA 5112
Australia
Phone: +61 8 256 0952
Fax: +61 8 255 9117
Microsoft Network is prohibited from redistributing this work in any form,
in whole or in part. Copyright, Leith Mudge, 1995
License to distribute this post is available to Microsoft for $1000.
Posting without permission constitutes an agreement to these terms.
|
|
|
|
| Re: Yet another bug?? [message #4282 is a reply to message #4253] |
Wed, 17 May 1995 00:00  |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
phil@peace.med.ohio-state.edu (Phil) writes:
> Sorry to keep posting but I've just started developing some useful
> tools in IDL and it seems I run across a new 'bug' every day. Could
> someone tell me the sense behind the following:
> IDL> print, not 0
> -1
> IDL> print, not 1
> -2
> Shouldn't (not 0) = 1 and (not 1) = 0 or am i just missing something?
> Is there somehting in the manuals that describes this logic?
> Just wondering.
What the NOT function really does is to reverse all the bits in a number. Thus
the bit pattern representing 0, which not surprising is all 0s, turns into all
1s which is the bit pattern for the integer -1. Such behavior is really useful
when one wants to get down to the bit level of a number, and I think that other
languages behave this way as well. Apparently, any even integer is treated as
false and any odd integer is treated as true. For example,
IDL> for i = -5,5 do if i then print,i
-5
-3
-1
1
3
5
But the function KEYWORD_SET, which is also associated with boolean values,
seems to behave differently. It only treats the value 0 as false, and anything
else as true.
IDL> for i = -5,5 do if keyword_set(i) then print,i
-5
-4
-3
-2
-1
1
2
3
4
5
IDL> print, keyword_set(not 1)
1
I consider that a bug.
Also, floating point numbers work differently then integer numbers, and match
the behavior of KEYWORD_SET.
IDL> print, not 0.
1.00000
IDL> print, not 1.
0.00000
and
IDL> for i = -5,5 do if float(i) then print,i
-5
-4
-3
-2
-1
1
2
3
4
5
Basically, it's a mess. I'm not too upset about the difference between
floating point and integer behavior, but KEYWORD_SET really worries me now. If
I use a statement like
MYPROG, TEST_KEY=(NOT A)
and I'm using KEYWORD_SET inside the routine to test whether or TEST_KEY was
set, I'll get the wrong answer.
Bill Thompson
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
