On Thu, 15 Mar 2001, Darren R. Gitelman wrote:
> Dear Peter:
>
> We have the exact same setup at our center (HPUX + LINUX), and have not had
> any problems with endian issues. The HP machine can read volumes created on
> the linux machines and vice versa, both over NFS and on volumes moved from
> one machine to the other. The only caveat is that you can't let the *.hdr
> files and the *.img files get separated from one another. If you remake the
> headers on the opposite platform they will be wrong. Otherwise it's worked
> fine without modification. If you have specific questions about this please
> ask our programmer, address above.
>
Perhaps I did not make myself too clear when raising the original
question. Our problem is more or less that we are using other software
alongside of spm which requires that the data is and stays big-endian.
It seems, however that the changes in in spm_hwrite.m and
spm_write_plane.m proposed by John Ashburner is doing exactly what I
want. I will do a thorough test of this some time next week and return
with the result to the mailinglist.
Thank you all for your input.
Peter Willendrup
--
-------------------------------------------------------
Peter Kjaer Willendrup, Cand. Scient. (Physics)
Copenhagen University Hospital
Neurobiology Research Unit
Phone: +45 35456721
email: [log in to unmask]
-------------------------------------------------------
|