Hi all,
I'm a bit concerned about the number of version increments we have made
to the CF convention so far this year. A lot of software and systems
depend on CF (as is the nature of standards) and I'm worrying about
software keeping up with the spec. Even more worrisome to me is the
complexity and maintainability of software that tries to track and deal
with an increasing number of versions. I know the recent changes weren't
sweeping in nature, I'm just wondering if quarterly changes to the
version number aren't a bit excessive for a standards document.
The "Rules for CF Conventions Changes" document says that once a ticket
is closed and accepted, the changes are added to the specification
document and the version number is incremented. It also mentions that
the changes are provisional "until at least two applications have
successfully interpreted the test data". I'm wondering about two changes
to these change rules:
1) Decide on a schedule for version number increments (quarterly,
semi-annual, ... ?)
2) Don't add changes to the upcoming version of the specification
document "until at least two applications have successfully interpreted
the test data".
I know this might slow the process down but I'm not sure that isn't a
good thing.
On another note, how is the test data and successful interpretation by
applications being tracked? Is each ticket required to have test data?
How fully do the applications need to interpret the data? Did ticket #18
(the horizontal CRS ticket) have test data? And if so, are there
applications that successfully interpret and use the ellipsoidal earth
attributes? [The change rules document doesn't actually mention
applications _using_ any additional information, is that intentional?]
Ethan
Velimir Mlaker1 wrote:
> Dear community,
>
> Version 1.2 of the CF Conventions has been released (see
> http://cf-pcmdi.llnl.gov/documents).
>
> This update includes changes detailed in Trac tickets #18
> (http://cf-pcmdi.llnl.gov/trac/ticket/18) and #25
> (http://cf-pcmdi.llnl.gov/trac/ticket/25).
>
> Sincerely,
>
> Velimir Mlaker
> PCMDI
>
>
--
Ethan R. Davis Telephone: (303) 497-8155
Software Engineer Fax: (303) 497-8690
UCAR Unidata Program Center E-mail: edavis at ucar.edu
P.O. Box 3000
Boulder, CO 80307-3000 http://www.unidata.ucar.edu/
---------------------------------------------------------------------------
Received on Mon May 05 2008 - 13:39:57 BST
