From bug-octave-request at bevo dot che dot wisc dot edu  Wed Sep 30 08:58:01 1998
Subject: Unidentified subject!
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: Ben Kao <bkao at laslo dot eecs dot uic dot edu>
Cc: bkao at ai dot uic dot edu, bug-octave@bevo.che.wisc.edu
Date: Wed, 30 Sep 1998 08:57:56 -0500 (CDT)

On 30-Sep-1998, Ben Kao <bkao at laslo dot eecs dot uic dot edu> wrote:

| To: bug-octave at bevo dot che dot wisc dot edu
| Cc: bkao at ai dot uic dot edu
| Subject: matrix indexing bug?
| 
| Bug report for Octave 2.0.13 configured for i686-pc-linux-gnulibc1
| 
| Description:
| -----------
| Octave 2.0.13:
| 
| octave:14> a(1) = 1
| a = 1
| octave:15> a(2) = 2
| a =
| 
|   1
|   2
| 
| Matlab 5:
| >> a(1) = 1
| a =
|      1
| >> a(2) = 2
| a =
|      1     2
| 
| As you can see the matrix grows in a different direction.
| Is this a bug?  At least this seems to be an inconsistency.
| 
| Fix:
| ---
| 
| If a matrix is explicitly defined, (i.e. matrix(x,y) = ...) there is no 
| problem since we know x denotes column number and y denotes row number.
| However, if the matrix is one dimensional (i.e. matrix(x)) then although
| it is more or less arbitrary whether this is a row or column matrix, it
| is ambiguous _when_the_matrix_is_first_defined_.  Thus there should be 
| (probably is) a rule or convention that addresses this matter.  I noticed
| that 2.0.9 and 2.0.13 are consistent with each other at least.  Has Matlab
| always done it the other way?

Yes, I think so, and I believe it doesn't offer you a choice about
which you prefer.

You can control Octave's behavior by setting the built-in variable
prefer_column_vectors to 1 or 0.

jwe