Re: Inconsistent results with the examples in Data Analysis Tutorial

(Sorry, I messed up again by double mailing. Should have replied to
the mailing list.)

2007/7/17, Steven G. Johnson <[email protected]>:
> On Tue, 10 Jul 2007, Chia-Hung Lin wrote:
> > However, what I got looks like the other attached file,
> > ubuntu_result.png, which is a 90-degree-rotated version of the example
> > in the tutorial, the original distorted image is the same with that of
> > the example, though.
>
> I think that h5topng may have changed conventions (rotating its output
> image by 90 degrees) since that tutorial was written.
>
> Steven
>
> _______________________________________________
> mpb-discuss mailing list
> [email protected]
> http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss
>
> On Tue, 10 Jul 2007, Chia-Hung Lin wrote:
> > However, what I got looks like the other attached file,
> > ubuntu_result.png, which is a 90-degree-rotated version of the example
> > in the tutorial, the original distorted image is the same with that of
> > the example, though.
>
> I think that h5topng may have changed conventions (rotating its output
> image by 90 degrees) since that tutorial was written.
>
> Steven
>

Dear Prof. Johnson,

Thanks for the kind reply.

But if the changing of conventions of h5topng is the case,  the same
will be true when I use rectangular coordinate system, in which I did
not observe the rotating effect, though.

For example, if I use the sq-rods.ctl provided with the MPB package,
and change the center of the rod to, for example, (0.25 0 0), i.e. to
shift the rod horizontally by +0.25a,   (a part of the ctl file reads
as follows:)

(set! geometry-lattice (make lattice (size 1 1 no-size)))

(set! geometry
      (list
       (make cylinder
         (material GaAs)
         (center 0.25 0) (radius r) (height infinity)))),

then according to the changing of conventions of h5topng,  the
dielectric function should look like the attached file, named
"vertical.png", namely, the rod should be shifted downward.

However, the result looks like the other attached file, named
"horizontal.png", in which the rod was exactly shifted horizontally in
+x direction.  And apparently this is an inconsistent result.

Therefore, is there any possibility that the rotating effect was not
caused by the changing in conventions in h5topng, but the data
processing of mpb-data in converting a non-orghogonal coordinate
system data to a rectangular one, instead?

Best regards,

Chia-Hung

--
Chia-Hung Lin
Institute of Photonics and Optoelectronics, National Taiwan University

_______________________________________________
mpb-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss