Re: advice needed for feedtype and product identifiers

Hi Joe,

Sorry to be so unresponsive, but we've got an all-consuming 2-day
Unidata Users Committee meeting going on right now.  I got your phone
message yesterday, but haven't had time to look into it yet.

> ATD is going to try to use LDM to send various data sets for the IHOP 
> project.
> 
> We need advice on what feedtypes and product identifiers to use for
> 
> 1) S-Pol radar data in DORADE sweepfile format
> 2) Composite gridded radar data in netCDF format.
> 
> I suppose I can start testing using the 'EXP' feed type, but if there is 
> a better naming convention, please let me know.

The current feedtype assignments are available at:

  http://www.unidata.ucar.edu/packages/ldm/feedtypes/index.html

but I don't believe there is anything better than "EXP" to use for
now.  Product identifiers are completely up to you, with the following
guidelines:

 - product IDs should be shorter than 255 bytes (KEYSIZE in
   protocol/ldm.h)
 - product IDs should be designed for regular expression pattern
   matching and extraction of information with substring matching
 - make sure the product ID contains all the info needed to route the
   product or to act on it with pqact.conf entries
 - don't use tabs or newlines in the product IDs (but spaces OK)
 - make your IDs unique enough that they are unlikely to clash with
   other products using the EXP feed type.  Typically this is done
   with a short unique prefix on product IDs.  (This may not be an
   issue if you are not on a network where other EXP products may
   appear.) 

There may be some better guidelines for product IDs; I'll look around
as soon as I get a little spare time.

--Russ