[Fwd: HDF5 prototype]

[Nick Maliszewskyj]

Ray Osborn wrote:

> In responding to this particular point, a key question is going to be the
> time scale over which HDF5 becomes available.  If sufficiently fast, we may
> not have enough legacy HDF4-based NeXus files to worry about backward
> compatibility.  Even if we do have a reasonable number of files, it may be
> possible to write simple conversion utilities from HDF4 to HDF5 using both
> sets of code.  It would be interesting to hear from Mark, who is probably
> further along than the rest of us in using NeXus, what he considers an
> appropriate time scale to be, and how easy it would be to write conversion
> programs.
> 
> If we reach a consensus on this point, will you carry on the correspondance
> with NCSA, Nick?  We should try and speak with a coherent voice.
> 
> Ray
> 

Sure, I'd be happy to continue corresponding with the NCSA. On the face of it,
HDF5 answers some of the concerns we had discussed some months back vis-a-vis
our overkill in using SDS's when all we really wanted was a dressed-up Vdata. 

I have downloaded the prototype API from the NCSA and will look into upgrading
the NeXuS API to suit it. I think that's really the best way to get an idea of
its abilities/limitations.

Nick


-- 
|_    |_  |_  |_________________   Nicholas C. Maliszewskyj
|__   |_  |_  |_         |_        NIST Center for Neutron Research 
|_\_  |_  |_  |_         |_        National Institute of 
|_ \_ |_  |_  |______    |_                  Standards & Technology
|_  \_|_  |_       |_    |_        Bldg 235, Rm E123         
|_   \__  |_       |_    |_        Gaithersburg, MD 20899
|_    \_  |__________    |_        nickm at rrdjazz.nist.gov