How is dielectric data in the .h5 'slices' stored?

MPB Users:

I would like to thank Mischa Megens and Steven for their helpful  
replies to my previous question.  Once I removed hdf5-1.8.1 and  
recompiled the older hdf5-1.6.7 libraries, both h5utils and the recent  
version of mpb installed without a hiccup.

My question today is about the epsilon.h5 files that MPB uses to store  
the dielectric data.  I have used an HDFviewer application and  
Matlab's hdf5read function to rip apart the file and pull out the  
dielectric 'slices'.  This is the first time that I have used the hdf5  
file format, but from what I can gather, data is stored in a  
hierarchical structure (similar to a UNIX directory tree) that allows  
each set of data to be separated and organized in a logical fashion.   
However, I am having some difficulty making sense of how the data is  
stored.  From what I can see, the data is stored like this:

epsilon.h5 -->	data
				description
				epsilon.xx
				epsilon.xy
				epsilon.xz
				epsilon.yy
				.
				.
				.
				epsilon_inverse.xx
				epsilon_inverse.xy
				.
				.
				.
				lattice vectors

The fields "description" and "lattice vectors" are straightforward  
enough.  My problem comes from making sense of the "data" and  
"epsilon" data sets.  Each of these sets forms a 1x64 row.  When I  
force the data into a 8x8 matrix, I can generate an image that roughly  
agrees with my dielectric function.

So my questions are:
1)	When I open the data from epsilon.nm , what data slice am I  
actually looking at?
2)	How does h5topng grab all of the data points for a given slice (say  
all points where x = 15)


True Merrill

251.648.4093
[email protected]