On 9/16/10 9:49 AM, Robin Schwab wrote:
Those are very weak arguments: social impact...
Maybe. But an all-free-software policy can be defended, even on the
grounds of pure expediency.
I'm a developer and I've worked in situations that were a mix of
proprietary and free software. I much prefer it when it's all-free, or
at least as much as possible.
I have frequently had to tell people "sorry, that's impossible" when
working with a proprietary tool. When we are working with free tools,
the only limit is how much attention or skill we can bring to the
problem. If necessary, we can write the software ourselves.
Also, if we know that all of our software and data can be replicated
without limit on any server, backups, archives, even a developer's
machine, that's a huge win for many reasons. It's easier for ops, easier
to archive, easier for remote developers, easier for researchers. The
WMF doesn't have to have anybody doing useless tasks like tracking
software license compliance.
When we /a priori/ exclude one or the other we will
chance of using the really best software.
Yup. We are not using the best software today. We are trading that for
the assurance that we can do *whatever* we want, today, or in the future.
Given this situation the only alternative we have is
to actively let
somebody program the requested feature or to wait until somebody does it
spontaneously. Both ways it may take years to have a satisfactory result.
I agree that this sucks, and there ought to be some better way to fund
free software development. Right now we rely mostly on developer
caprice, or relatively unusual situations where a company finds it in
their interest to give away their code.
However, that is the angle we should be working on - how can we fund the
software we want to have - rather than looking for some compromise with
proprietary software, which we know will come back to haunt us.
Neil Kandalgaonkar |) <neilk(a)wikimedia.org>