[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: HDF, netCDF, etc question

Aaron Birenboim wrote:
> Randall Skelton <rhskelto@atm.ox.ac.uk> wrote:
> : I suggest the use of an SQL database :-)
> Frankly, I agree.
> I have used many formats, including a customer that
> demanded a customized HDF based representation.
> All a big mess.
> I highly recommend storing all your meta-data in an SQL database.
> If you do not want to use BLOB's, store path names to the raw binary data.
> Store data type, size, etc in the DB.
> HDF is very flexible.  So flexible that its not very useful
> as a standard.  

I agree HDF or netCDF is very flexible. And if you don't give rules
to the people using it you get very different files for the same data
by each people. 

How to give rules?

Our routines are using definition files e.g.


Only attributes which are defined will be read or written. (case
If they are global or parameter attributes is defined 
and the resulting structure is defined by this rules too.

David Stern told me last year that they don't upgrade
HDF till the problem of the versions by HDF is fixed.
So idl user should not get into this problems.

>In fact, we implemented on HDF 4, then HDF 5 came out,
> and its not backward compatible.  So this major customer
> is sitting on top of TONS of data stored in an OBSOLETE format.
> --
> Aaron Birenboim
> Albuquerque, NM
> aaron@boim.com
> http://boim.com/~aaron

Reimar Bauer 

Institut fuer Stratosphaerische Chemie (ICG-1)
Forschungszentrum Juelich
email: R.Bauer@fz-juelich.de
a IDL library at ForschungsZentrum Juelich


read something about linux / windows