[galeon] A modular approach to CF-netCDF standardization

• To: CF-netCDF SWG <CF-NetCDF-1.0.SWG@xxxxxxxxxxxxxxxxxxxxxxxx>
• Subject: [galeon] A modular approach to CF-netCDF standardization
• From: Ben Domenico <Ben@xxxxxxxxxxxxxxxx>
• Date: Thu, 1 Apr 2010 12:06:46 -0600

Hello to all you CF-netCDFers:

Via Rich Signell,  Roy Mendelssohn recommended an article that espouses an
interesting philosopy, namely "Don't solve problems, copy success:"

http://www.fastcompany.com/magazine/142/switch-how-to-change-things-when-change-is-hard.html

I am taking that suggestion to heart so my proposed new approach to
establishing an OGC CF-netCDF standard is modeled on the highly praised and
apparently successful WCS 2.0 specifications drafted by Peter Baumann and on
Stefano Nativi's very useful clarifications regarding standardization
targets in his presentation to the CF-netCDF SWG at the recent OGC TC
meeting.

To that end, I have drafted a Best Practice Document that provides an
overview of the general CF-netCDF landscape.  It includes a list of possible
standardization targets and shows how they fit together in the long term as
a core standard, extensions to the core, and application profiles that
employ the core and some extensions.  A copy of the draft Best Practice
Overview is attached.

After mulling over the draft and having a few brief discussions, I'm
suggesting that the core be confined to a specification of the netCDF
Classic Data Model.  With that very limited core, it makes possible
extensions that include

- the Enhanced NetCDF Data Model,
- CF (and possibly other) conventions which can also be considered
extensions to the core data model
- several binary, xml, and text encoding options
- APIs

I submit that this Really Big Picture for netCDF standardization will be
very useful for discussion even if we end up taking a different avenue to
OGC standardization.

To provide a sense of what it would look like, I've also drafted a netCDF
core specification confined solely to the netCDF classic data model.  This
is a good illustration of the approach taken in WCS which results in a
number of concise and to-the-point standards documents tied together with a
Best Practice Overview.

Finally I've attached a diagram created by Stefano Nativi that provides an
alternative to my diagram in the Best Practice draft.  These are all
obviously very rough drafts and the new approach requires substantial
discussion so I'm sending them out now to get the discussion and adaptation
started sooner rather than later.

In the meantime, I'll be working on a draft specification for the netCDF
classic binary encoding that should fit into the context of the new
approach/structure.  If I can get that done in time, perhaps we can discuss
all this at a CF-netCDF SWG telecon next Thursday.

Happy Easter.
-- Ben

PS By proposing this on this date, I have a convenient out in case I have
completely missed the mark yet again. (smile face here)

Attachment: 10-NBP_CF-netCDF_overview_best_practice_2010-04-01.doc
Description: MS-Word document

Attachment: 10-NCD_NetCDF-core_2010-04-01.doc
Description: MS-Word document

Attachment: netCDF-extensions-planning(SN).png
Description: PNG image