dolfin team mailing list archive

["Martin Sandve Alnæs" <martinal@xxxxxxxxx>]

2008/11/30 Anders Logg <logg@xxxxxxxxx>:
> On Sun, Nov 30, 2008 at 10:43:00PM +0100, Martin Sandve Alnæs wrote:
>> 2008/11/29 Anders Logg <logg@xxxxxxxxx>:
>> > On Sat, Nov 29, 2008 at 06:46:58PM +0100, Martin Sandve Alnæs wrote:
>> >> <Paranoid stuff>
>> ><Sensible answer>
>>
>> You're right, there's no problem here.
>
> I'm happy to hear that. I'm just guessing as usual so if you think it
> works it's a good indication it might. :-)

Well, "Functions" won't work as I initially argued, but we don't need
it and ufl split does the job. And a T*Function is never used as a
dolfin.Function, so there are no complications there.

Also, it's clearer when the user writes

fg = Function(mixed_space)
f, g = split(fg)
a = ...f...g...*dx

since it's easier to see that fg is a single coefficient while f and g are not
dolfin.Function objects (and thus cannot be plotted etc) than if the
user writes:

f, g = Functions(mixed_space)
a = ...f...g...*dx

--
Martin