Hi Folks
Sorry about the long silence. Those who know me, will know I have strong
opinions about a number of the things being discussed recently, so you'll
also know that the fact I haven't been replying is due to other pressures.
Anyway, I'm out from under for a while ...
I'll comment on a number of other threads shortly, but firstly I wanted to
expose some ideas some of us have about the future of CF.
Firstly, I think we'll all agree that CF is evolving.
Secondly, I suspect most of us are having trouble holding the various threads
of activity in our heads, and the current status of those threads.
Thirdly, the CF authors, who are our benevolent dictators (using a phrase
beloved of the programme language community), have all at one time or another
expressed that they can't cope (quickly enough) either ...
Fourthly, anyone attempting to write code based on CF is faced with a standard
that is getting harder and harder to interpret
Fifthly (stop numbering everything Bryan), the standard name vocabulary is
becoming unwieldy in and of itself to manage
Sixthly (I failed to stop :-), CF is becoming so important it is having to
interoperate with other communities (e.g. GIS, of which more in a later
email) ...
You get the picture. I could have written more.
So what can we do about it?
One of the things I think we need is a methodology which is more sophisticated
than a mailing list to keep track of threads and discussions. Things need to
be marked as live, or forgotten, or under action, or going into a future
VERSION of CF. OK, one of my cards on the table: we need version control! We
also need a mechanism of making a decision about what's happening, and we
need more status information.
<We> needs to be defined. In my mind <we> is the entire CF community. Well, we
all know how efficient a committee is at building things, so a mailing list
in and of itself doesn't make decisions... arguably CF is successful because
we have a decision making process (the CF authors have final say). But for
how long can they keep that up? If we want to change things, how could we
change them? What happens when the CF authors interests start to diverge from
the community of CF users?
I would argue we need a community process, perhaps including an elected
committee, perhaps including the right of veto from the CF authors (or some
trusted group of folk). There are lots of options. This email isn't so much
about what we do, as about suggesting a strategy to coming up with a process.
(And, we need a standards track. We need CF to become more than a community
standard, we would like WMO to take it on board etc ... if that happened,
then we really do have an incentive for writing code which understands and
exploits CF.)
Anyway, back to the strategy. So what should that strategy be? Some of you
will know of the Global-Organisation for Earth System Science Portals
(GO-ESSP, see
http://go-essp.gfdl.noaa.gov/). We have a meeting in June, in
the UK, and on the agenda we will have a discussion of the future of CF. All
the CF authors except Brian can and will make it. (Meeting details at:
http://go-essp.gfdl.noaa.gov/news_meetings.html).
On the table will be some real support for CF activities rather than
best-effort activities. I expect (but wont be able to confirm til June) to
have 50% of an FTE on CF support at the BADC from April 2006. It would be
nice if we could find some other players who can commit some real effort to
making this work. Having done that we obviously need a mechanism to make sure
those of us managing and funding the effort are doing what the community
wants, rather than what we want alone (if we did the latter, we'd be wasting
the effort).
Obviously not everyone who cares can come to that meeting, and many of those
who care wont want to come because CF is only a small part of the agenda. So
it will be important for us to canvas opinion via this email list before that
meeting.
So, this long email is by way of foreshadowing (and perhaps engendering) what
I hope will be some vociferous discussion of how we take CF forward. I think
we all agree that it is important, and I hope we can find some consensus on a
mechanism for community maintenance of CF that maintains the existing rigor
(spelt like an American to show how International I can be :-), but is easier
for the community to use and evolve. (For evolve we must :-)
Discuss :-)
Bryan
--
Bryan Lawrence, Head NCAS/British Atmospheric Data Centre
Web: badc.nerc.ac.uk Phone: +44 1235 445012
CCLRC: Rutherford Appleton Laboratory, Chilton, Didcot, OX11 0QX
Received on Thu Feb 24 2005 - 02:05:18 GMT
