For a big joint research project we want
to transfer data between LabVIEW and GNU-Octave.
In this case data means structures,
as we intend to have complex sets of measuring data and meta information
in a single file.
On the octave side a structure seems
to be the way, on LabVIEW it could be represented as a cluster.
There is even a library available to
write labview clusters (and other types) to a mat file "http://lavag.org/topic/10976-a-mat-file-io-library/page__st__40".
When I use write a mat-file with this
library (c.f. attached file "test.labview.mat" ) the file can
be loaded by MatLab, but not by octave:
-----------
octave:4> load "test.labview.mat"
error: load: invalid array name subelement
error: load: trouble reading binary
file `test.labview.mat'
-----------
However, when I save the structure again
with MatLab (c.f "test.labview.matlab.mat") I __can__ read the
resulting file in octave.
------------
octave:1> load "test.labview.matlab.mat"
octave:2> who
Variables in the current scope:
}
-----------
I don't use any special flags when I
save from MatLab it's simple load+save.
Is there I way I can investigate the
two files for the crucial differences in order to either improve octave
or the labview library (both open source).
I'm at a loss on how to get down to
the root of this problem.