Re: long variable name problem inside IDV

Hi Kent:

I think Joe is mistaken, there is no 15 char limit either in CF or in the 
IDV/netcdf-java stack. Im guessing that he has unnecessarily truncated to 15 
chars in the server.

He is right that a file must follow a coordinate convention like CF. Opendap 
grids alone do not have sufficient semantics to allow geolocation.

If you can send me a sample HDF5-EOS file i can comment further.

MuQun Yang wrote:
> Hi John,
> 
> We found a limitation inside IDV that doesn't allow the variable name to
> be longer than 15 characters. This is a problem for EOS data whose
> variable name may be longer than 15 characters. We sent an email to one
> of IDV developer but have never received a reply(See the email below).
> Could you help us build up the connection? IDV is a cool tool. If it can
> support long variable name, that will be perfect.
> 
> Thanks,
> 
> Kent
> 
> The name of the first contact is Jeff McWhirter.
> We got his name from this list:
> 
> http://www.unidata.ucar.edu/about/staff.jsp
> 
> ---------- Forwarded message ----------
> From: H. Joe Lee <address@hidden>
> Date: Wed, Aug 27, 2008 at 11:25 AM
> Subject: [Q] IDV and OPeNDAP Grid variable name length?
> To: address@hidden
> 
> 
> Hi, Jeff!
> 
> My name is Joe and I'm working on OPeNDAP HDF5 handler project at The
> HDF Group. I got your e-mail from Unidata's staff list web page.
> 
> In the OPeNDAP HDF5 handler, I've implemented a configuration option
> to support IDV client for NASA HDF-EOS5 HDF5 files. During the handler
> testing, I noticed that IDV couldn't handle OPeNDAP Grids that have
> variables whose names are longer than 15 characters. Is my observation
> true? If so, do you know why such restriction is enforced in IDV or
> know how to relax such restriction?
> 
> You may also be interested in reading my technical note with the
> OPeNDAP HDF5 handler:
> 
>   http://hdfdap.hdfgroup.uiuc.edu:8080/cf.html
> 
> Regards,
> 
> -- 
> -- Joe
> 
> 
>