On Tue, Feb 04, 2014 at 06:24:48PM +0100, Felix Höfling wrote:
Am 04.02.2014, 16:05 Uhr, schrieb Pierre de Buyl
<address@hidden>:
I always felt uncomfortable with specialising to a certain String type.
I
suggest to drop "VL" and allow _any_ kind of HDF5 String type for the
writer. The reader can easily handle the different cases by querying
H5Tis_variable_str
http://www.hdfgroup.org/HDF5/doc/RM/RM_H5T.html#Datatype-IsVariableString
Actually, I think it will avoid a lot of problems to define what kind of
strings
we want. Aside from the purely practical aspect of writing less code, it
also
avoids the question from a new user: "what kind of strings should I
use?".
Removing this question makes it easier, in my opinion. Also, it
decreases the
amount of documentation in, say, the "implementation" page of the h5md
site.
I really didn't want to bring the debate back, but a solution in Fortran
should
be available :-)
If, as it is discussed on the other sub-thread, you cannot use VL
strings in
Fortran for the attributes (very specific case, but still), I'd prefer
that we
mandate fixed-length.
Cheers,
Pierre