[netCDF #MSV-188900]: C++ API for NETCDF version 4.0

Lynton,

> I thought I should drop you a note to say that I am back working on the
> C++ API again and hope to have it completed very soon. I hope that the
> results will have the possibility of being incorporated in the NETCDF
> distribution, assuming that the API is successful.

That's great to hear!  I considered asking you about the status of that 
effort, but didn't want to appear to be pressuring you for results when
this is a volunteer contribution to the netCDF community :-).  I would be
happy to incorporate your results into the netCDF distribution when it's
ready.  We have a beta version of netCDF 4.1 that would be a good place
to deploy a test version of the new C++ API implementation.

> I have spent quite a bit of time playing around with ideas. In the end I
> have elected not to use the blitz array library and stick with what the
> STL can offer. I have also moved away from the original C++ API quite a
> bit. Hopefully the result is simpler and more transparent. One of my
> rationals is to store as little data in each object as possible and rely
> on the C API calls to deliver everything else. This minimizes the
> possibility that objects may not be holding the most current information .

That's a philosophy I agree with whole-heartedly.  Keeping extra copies of 
the state of things in the C++ layer is undesirable for exactly the reason 
you have stated.

> Please let me know if there are any changes regarding the C++ API on
> your end.

No, we have been telling users who ask about it that the netCDF-4 interface
is "under development" and we have made no changes to the partial 
implementation that's in the current distribution.

Thanks again for your valuable work on a new C++ API.

--Russ

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: MSV-188900
Department: Support netCDF
Priority: Normal
Status: Closed