| Re: Is it an IDL bug or feature? [message #5239] |
Thu, 26 October 1995 00:00 |
hahn
Messages: 108
Registered: November 1993
|
Senior Member |
|
|
Liyun Wang <Liyun.Wang.1@gsfc.nasa.gov> wrote:
> Hi,
> In a process of debugging our IDL codes, we found an interesting case which you
> may want to be aware of. Consider this:
[snip]
> IDL> b = 1
> IDL> a = b([2],[1],[2])
[snip]
> What do you think IDL will respond? Will it issue a syntax error message?
> Wrong!
IDL works as described in the manual: User's manual p 5-5, formula
beginning with V(S) =
> As long as the subscript indices of a *scalar* variable, like b here,
> are arrays with the same dimension, IDL does not care nor complain, no matter
> how
> many array elements you give. Do you consider this to be a potential bug or a
> feature of IDL?
The reason for this feature is to save time when doing image
processing: Imagine a 512x512 array of luminance values and
use these to adress a color lookup table. Some luminance
values may be missing or out of range in this example. Thus
you save lots of processing time (for subscript checking)
when the top and the bottom element of the lut are set to the color of
missing value.
> --
> ============================================================ =========
> Liyun Wang | Voice: 301-286-3126
> G1 Building 26, SOHO EAF | FAX: 301-286-0264
> NASA, GSFC, Code 682.3 | Liyun.Wang.1@gsfc.nasa.gov
> Greenbelt, MD 20771 |http://orpheus.nascom.nasa.gov/~lwang
> ============================================================ =========
Norbert Hahn
|
|
|
|
| Re: Is it an IDL bug or feature? [message #5243 is a reply to message #5239] |
Thu, 26 October 1995 00:00  |
chase
Messages: 62
Registered: May 1993
|
Member |
|
|
>>>> > "Liyun" == Liyun Wang <Liyun.Wang.1@gsfc.nasa.gov> writes:
Liyun> In a process of debugging our IDL codes, we found an
Liyun> interesting case which you may want to be aware of. Consider
Liyun> this:
IDL> b = 1
IDL> a = b(1,2,1)
Liyun> IDL will complain about subscript being out of range. But watch this:
IDL> b = 1
IDL> a = b([2],[1],[2])
Liyun> or
IDL> a = b([[2],[1],[10],[30]])
Liyun> or even
IDL> a = b([[2],[,1],[10],[30]], [[20],[12],[70],[20]])
Liyun> What do you think IDL will respond? Will it issue a syntax
Liyun> error message? Wrong! As long as the subscript indices of a
Liyun> *scalar* variable, like b here, are arrays with the same
Liyun> dimension, IDL does not care nor complain, no matter how many
Liyun> array elements you give. Do you consider this to be a potential
Liyun> bug or a feature of IDL?
It is not a bug. The specific behaviors are clearly documented
features in Chapter 5 of the "IDL User's Guide" (version 3.5 and
version 4, perhaps earlier too) under the subsections "Subscript
Examples" and "Array Subscripts".
It does seem unfortunate the behavior seems inconsistent. I find that
both behaviors are useful.
Chris
--
===============================
Bldg 24-E188
The Applied Physics Laboratory
The Johns Hopkins University
Laurel, MD 20723-6099
(301)953-6000 x8529
chris.chase@jhuapl.edu
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
