Re: M_MAP tools

[David Bateman]

Rajith Padmanabhan wrote:

> Has anyone been able to use M_MAP toos in Octave?
> http://www.eos.ubc.ca/~rich/private/mapug.html
> <http://www.eos.ubc.ca/%7Erich/private/mapug.html>
> or is there somthing similar for octave?
>
> thank you in advance
>
> -- 
> Rajith Padmanabhan


I took a quick look. There are several problems preventing its use, but
these might not be show stoppers for you. I tried running m_demo and
found the following

1) Octave doesn't yet have private directorys, so you have to also do
"addpath('m_map/private')"
2) There seems to be an issue of logical indexing in private/mu_util.m
where octave doesn't recognize a vector of 1's and 0's as logical
indexing, where matlab does. This seems to be a subtile incompatiability
between matlan and octave, though in the assignment to a logical array
of a 1 or a 0 doesn't change the type, though in octave it does. Consider


>> i = logical(ones(1,4))

i =

     1     1     1     1

>> i(1) = 0

i =

     0     1     1     1

>> whos i
  Name      Size                    Bytes  Class

  i         1x4                         4  logical array

Grand total is 4 elements using 4 bytes

whereas in octave this gives


octave:2>  i = logical(ones(1,4))
i =

  1  1  1  1

octave:3> i(1) = 0
i =

  0  1  1  1

octave:4> whos i

*** local user variables:

  Prot Name        Size                     Bytes  Class
  ==== ====        ====                     =====  =====
   rwd i           1x4                         32  double

Total is 4 elements using 32 bytes

John what do you want to do with this incompatiability? My solution in
the context of m_map was to cast as a logical before using it. That is
add the line

indx = logical(indx)

at line 109 of private/mu_util.m


3) The patch function of octave-forge is incomplete. In particular, you
can't have a LHS and it doesn't accept x, y plus three string args. I
didn't go any further at this point, but this means that the graphics
themselves of m_map can't be used. Though the projection code, etc can
be and then manipulated by non m_map for the graphics. The recent work
in octave on graphics handles should address this in the future.

D.

-- 
David Bateman                                address@hidden
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 6 72 01 06 33 (Mob) 
91193 Gif-Sur-Yvette FRANCE                  +33 1 69 35 77 01 (Fax) 

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary