comp.lang.idl-pvwave archive
Messages from Usenet group comp.lang.idl-pvwave, compiled by Paulo Penteado

Home » Public Forums » archive » Re: netCDF-4 support in IDL?
Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend 
Return to the default flat view Create a new topic Submit Reply
Re: netCDF-4 support in IDL? [message #66085 is a reply to message #66082] Sat, 04 April 2009 11:35 Go to previous messageGo to previous message
George White is currently offline  George White
Messages: 10
Registered: November 1998
Junior Member
On Fri, 3 Apr 2009, Paul van Delst wrote:

> Do any insiders out there know about the status of future support for
> netCDF-4 in IDL? I received third/fourth-hand news yesterday that there are
> no plans for IDL to support netCDF-4 and I wanted to get some more scoop
> before contacting ITTVIS directly. I searched the ng on google but didn't
> find anything.

I'm not an insider, but I was at the HDF meeting in Denver last fall
where there was considerable discussion about support for HDF-5 and
netCDF-4 support in commercial packages (IDL, Matlab, etc.). The
problem is that netCDF-4 depends on HDF-5, which has a very
big API that so far defies attempts to automate the task of writing
wrappers for other languages.

Commercial developers have no way to predict which parts of the API are
needed by user A, so they have no choice but to implement most of the API.
Meanwhile, open source (R, octave, gdl, ...) users have been writing
wrappers for the parts of the API needed for their work. So at present,
octave allows you to save vars to hdf5, but R won't read the file if one
of the variables is a range/sequence.

Add to this the issues that some hdf5 users are writing 64-bit programs,
so we have a 64-bit hdf5 library that lets us build program X, but need to
use the data in a program that uses a different library for which we don't
have the 64-bit version, so in the end one has to dump data in some binary
file using the 64-bit library and then read it into your 32-bit
application.

I recommend that people start playing with hdf-5 and then netCDF-4 using
gdl with the goal of working out the subsets of the API's needed to use
some important data sets. Then, rather than asking ITTVIS to implement
all of hdf-5+netCDF-4, just ask for the parts of the API's you need now.

--
George White <aa056@chebucto.ns.ca> <gnw3@acm.org>
189 Parklea Dr., Head of St. Margarets Bay, Nova Scotia B3Z 2G6
[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Re: reconstructing iraf spectra?
Next Topic: where to find Thin() Function's source code?

-=] Back to Top [=-
[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Sun Oct 12 00:01:29 PDT 2025

Total time taken to generate the page: 0.01067 seconds