Hi Tom,

By default order, do you mean that you are using the AFNI_BYTEORDER
environment variable for these users? If so, it sounds like you don't
actually want that. Let me describe the programming logic:

Assume there is no AFNI_BYTEORDER variable set. If the user is
performing a computation (e.g. 3dcalc), then the data is read in, and
if the byte order does not match that of the CPU, it is reversed (to
match the CPU). A new dataset is created from whatever computation
the user requested, and data is written to disk in the byte order of
the CPU.

If the user is not performing a computation (e.g. 3drefit, like your
case), then the data on disk is not modified, and the byte order is left
as is, unless the user requests to change it.

So these two operations work the way most users will expect them to work.

-------------

If the AFNI_BYTEORDER variable is set, that will override the output
byteorder in _all_ cases where a HEAD file is written. So regardless of
whether a new dataset is created, an existing one is modified, or the
data of an existing one is not modified (but the HEAD file is, like with
3drefit), the environment variable will override the normal action of
AFNI, setting the order to what the user has specified with that variable.

This should be done only in very special cases, not in general.


Please let me know if this does not answer your question.

- rick