[Vis5d-users] Re: HDF5 complex datasets

Hi St=E9phane,

> before using vis5d, i am investigating the possibility to
> visualize an earth mesh made of irregular cells (all cells=20
> have 4 specific polar coordinates), that contain some physical
> parameter values.

Do you mean quadralateral cells in 2-D polar coordinates,
of tetrahedral cells (i.e. 4 vertices each) in 3-D spherical
coordinates?=20

> From HDF5 user's guide
> (http://hdf.ncsa.uiuc.edu/HDF5/doc/Dataspaces.html)
> i guess that "complex data space" may help me to represent my data,
> but the manual says "Currently only scalar and simple dataspaces are=20
> supported with this version of the H5S interface".
>=20
> I'd like to know if i'm right in the dataspace type i should use,
> and if yes are complex dataspace available today  ?
> (the document is dated 14 October 1999)

We are not experts in HDF. In any case Vis5D can only display
data with regular grids. However, our VisAD system can display
irregular data (go to the URL in my sig, and click on the
VisAD link). VisAD is for developers rather than end users,
and requires some programming in Java. Note that VisAD's
interface to HDF5 includes non-Java "native code" (C++) and
may not handle that data type you are contemplating (the HDF5
interface to VisAD was written by NCSA and I am not sure of
its exact capabilities).

Cheers,
Bill
----------------------------------------------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI  53706
hi...@fa...  608-263-4427  fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html