[OSGeo-Discuss] scale of FOSS projects

[Tim Bowden]

On Thu, 2008-05-08 at 21:28 -0400, Miles Fidelman wrote:
> Michael P. Gerlek wrote:
> > Or, to quote the IETF, "rough consensus and running code".
> >   
> Except that the reference is to the informal criteria for when one might 
> even beginning to firm up a standard.  In the IETF community - unlike 
> pretty much every other standards body on the planet - there's a pretty 
> strong insistence that there are multiple implementations of something, 
> that  an talk to each other, before even thinking about pinning down 
> anything that looks like a standard.
> 

IMHO standards are just a fancy way of documenting the solution.  Until
you've build the solution, you don't understand the problem properly
[1].  If you try and write your standard while your understanding of the
solution space is underdeveloped, you'll end up with a pile of shite.

Development is relatively fast and cheap, whilst standards are slow and
expensive.  Start with required outcomes, develop the solution, then
document or "standardise" the solution.  Put them in the wrong order,
and you'll cripple both the solution and standard.

[1] ESR explains it better than I can in catb: lesson 3 in
http://catb.org/~esr/writings/cathedral-bazaar/cathedral-bazaar/ar01s02.html. 
 
Regards,
Tim Bowden

> Pretty much everybody associated with the IETF is funded by nice, large 
> government contracts or has nice positions at large corporations, or 
> both.  And pretty much all of the early code in and around the Internet 
> (and the ARPANET) was written by people with DARPA and NSF grants (when 
> they defined the TCP/IP protocol, Bob Kahn was either at BBN, my old 
> stomping grounds, or at DARPA, and Vint Cerf was a professor at 
> Stanford).  The original reference implementation of TCP/IP - which 
> found it's way into an awful lot of different Unix variants - was 
> written by folks at BBN, again, funded by DARPA.  Just read through the 
> library of RFCs at www.ietf.org and you'll find that most of the authors 
> have fairly serious organizational affiliations - they're doing the work 
> as part of their day jobs.
> 
> Not that I'm complaining, mind you.  Simply pointing out that leading 
> edge software tends to be written by folks with solid institutional 
> bases, and salaries, supporting them. 
> 
> Miles
>