Re: [h5md-user] units module

[Pierre de Buyl]

On Wed, Nov 06, 2013 at 03:58:30PM -0500, Peter Colberg wrote:
> On Wed, Nov 06, 2013 at 04:47:27PM +0100, Felix Höfling wrote:
> > So what is the conclusion? That a reader has to be prepared for the
> > different encodings and define the memory datatype accordingly when
> > reading strings? What about reading a file datatype UTF8 to a memory
> > datatype ASCII (i.e., not caring about the encoding)?
> 
> My suggested conclusion is to test, and choose either, but not both.
> 
> If the tests yield that UTF-8 can be written and read easily and by
> all needed library interfaces, than we stick to UTF-8 encoding for the
> unit field as defined by the units module.
> 
> If the tests yield that UTF-8 cannot be handled easily by some
> clients, than we stick to ASCII encoding for the unit field.

Pragmatic :-)

Well, I tested writing from Fortran and that works nice :-)

Peter and I tested reading in Python and that is ok.

Konrad, you expressed concerns about UTF8, could you comment on whether that
would be a problem?

Olaf?

What about Lua, tcl, other?