From bug-request at octave dot org Fri Mar 17 08:17:20 2006
Subject: Re: 2.1.69: load/save -hdf5: matrix are transposed
From: Bill Denney <denney at seas dot upenn dot edu>
To: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
cc: David Bateman <David dot Bateman at motorola dot com>,         Cyril Humbert <cyril.humbert@univ-mlv.fr>, bug@octave.org
Date: Fri, 17 Mar 2006 09:16:38 -0500 (EST)

On Fri, 17 Mar 2006, John W. Eaton wrote:

> Is that because there is no way to determine that the HDF5 file was
> written one way or the other?  If so, could we add a file version in
> the header of the file that would allow us to process the old files
> (that might not have the header) using the old method, and those with
> the version info using the new method?
>
> At some point we might also consider doing that for the other formats
> so we could clean up the code that checks mdims < 0 to determine
> whether we have an ND array.

I'm planning on having versioning in the figures for the Object Graphics 
for this specific reason.  I personally have seen enough file formats that 
didn't have a version number that then became very hard to work with when 
someone needed to keep the original format mostly, but needed to change a 
few small things.

I think that any file that octave outputs should have a file version 
number.  To me, this should be a feature by 3.0 to prevent problems by the 
time we get to 3.1 and beyond.

Bill

-- 
Guil : Shall I read us a story before we go to sleep?
Ros : You can't read.
Guil : No, you can't read. I can't tie shoes. Keep it straight.
   -- _Rosencrantz and Guildenstern are Idiots_ by Jeffrey Gaulding



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------