On Sat, 2005-12-10 at 20:58 +0000, Andrew Brown wrote:

> I'm not a coder either. I have written some small macros that no one else 
> did first. But I do think it's clear that the open source model is not 
> universally applicable.

Did anyone ever think that it was? Well perhaps RMS and a few idealists
but I doubt most people think every single bit of code would become Open
Source code any time soon. The obvious candidates for Open Source are
large productivity tools where de facto standards lock people into
ineffecient economic production models.

>  It completely fails at some things -- show the Gimp 
> to a graphics professional some day. 

You mean like the people that did Shrek ;-)
You seem to assume GIMP is unsuccessful. The people that do use it
probably disagree. Success is not necessarily measured in market
dominance, it might simply be in providing a free alternative. Growth
rates are unpredictable, in fact the whole thing is relatively recent so
a lot is unpredictable.

> It is wholly successful at others -- 
> FreeBSD servers, Python, Perl, PHP. OOo itself seems to me to lie on the 
> margin. 

Like Firefox? Linux at the desktop?

The speed of adoption is related to a whole range of factors. FreeBSD
servers have been around for a very long time. I had BSD on a computer
back in 1989. OOo has been spectacularly successful when you take into
account that its being marketed against an entrenched monopoly that has
virtually unlimited resources and piracy on its side in terms of
maintenance of market share. OOo has only just got to its fifth birthday
and has 10-15% market share. In these circumstances this is nothing
short of miraculous. Do you think adoption of OOo is slowing? What do
you think it will be like 5 years from now? <15% or >15%? I'm willing to
bet $100 that OOo and its spawned variants will have more than 15% of
the market in 5 years time. Any takers?

So if the code continues to improve - no reason to suspect it won't the
only uncertainty is the rate at which it will improve - further growth
in market share will take place. Again the only real uncertainty is the
rate. If it drives adoption of open file formats that too will be a
success.

> Essentially, the project needs a political API -- a set of well-defined and 
> inviting rules which let interested parties help. 

The reasons that its difficult to participate in OOo are complex. I
doubt yet another set of rules will help. 

1. The code is massive and complex and you need a lot of coding skill
and experience just to get started.
2. The web site is difficult to navigate and there are barriers such as
JCAs and SSH tunnels to negotiate if you want to participate
3. While OOo is perceived as a Sun project feeding StarOffice, other
large companies are not going to be so inclined to put resources into
something they believe mainly benefits a competitor.

To get round these problems

Establish an OpenOffice.org foundation
GPL rather than LGPL 
Re-design the web site as a Wiki and only make the coding bits and
anything that absolutely needs it subject to high level security.
Provide tutorials and support for people who want to learn how to code
and contribute.
Tap into the formal education system to provide future coding resources.

While Sun dominates the project these things will not happen without Sun
taking the initiative. If Sun does not see it in its corporate interest
to do these things the status quo will remain. However, if things really
did start to stagnate badly I would expect a fork so I can't see OOo not
being successful in the long run no matter what the politics turn out to
be. Some routes will be easier and smoother than others and there could
be delays along the way but the long term future looks fairly bright
from here and certainly better than the previous monoculture.

-- 
Ian Lynch <[EMAIL PROTECTED]>
ZMS Ltd


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to