Hi Bob,
Interesting, I've been interested in indexing in HDF5 for a while now, but
I wouldn't necessarily have thought it would be needed for the NC4 work. Are
you thinking about the "get the n'th object" calls in netCDF?
Quincey
> As a result of some discussion, I have a revised list of major goals
> for NCSA.
>
> In the web page:
>
> http://my.unidata.ucar.edu/content/software/netcdf/netcdf-4/index.html
> there is one item under 'New Features for HDF5', 'Explicit support for
> coordinate variables/axes'. Item 2 below is a refinement of that
> bullet.
>
> The other 3 can be added.
>
> Consider these bullets a proposal, subject to discussion. When we get
> consensus on these, we'll set some firm milestones.
>
> 1. Update and correct the HDF5 specification document.
>
> 2. A dimension scales API sufficient to support coordinate
> variables/axes.
> See the proposal
> <http://hdf.ncsa.uiuc.edu/apps/netcdfh5/h5shared_dims.html>
> for an initial idea. We need to review and gain consensus on
> the design, and then implement it.
>
> 3. An index object that permits specifying an order for iterating
> over
> a set of objects, and perhaps also permits efficient insertion into
> and deletion from the ordered set it indexes.
>
> 4. MPI-IO enhancements to improve performance, as suggested by the
> Northwestern/Argonne work.
>
> --
> Robert E. McGrath (HDF Java Team Leader)
> National Center for Supercomputing Applications
> University of Illinois, Urbana-Champaign
> Champaign, Illinois 61820
> (217)-333-6549
>
> mcgrath@xxxxxxxxxxxxx
>
