At 11:55 PM 7/31/2000 +1000, David J N Begley wrote:

>On Sun, 30 Jul 2000, Doug Chamberlin wrote:
>
> > David, you comments suggest that its "all or nothing" when it comes to open
> > source. That any commercial benefit of an open source effort violates some
> > rule about open source.
>
>Not at all - I mentioned previously that an open source project, in order to
>succeed, does not rely solely on financial investment; this can help, sure,
>but it is the community which makes the project succeed - both financial
>investment and opportunities will arise from this. One only needs to look at
>some other successful OSS projects (ignore Linux) to see that this is true.

I agree completely about the importance of the community and have spent
much of the past seven months encouraging the growth of our community.
In most successful open source projects, including apache, there has been
a core of knowledgeable developers who helped others refine their patches
and develop an understanding of the bones beneath the flesh of the code.
So it should be for InterBase. Since InterBase is being delivered to the
open source community with 15 years of commercial history, there is benefit
to including people from that group in the initial "inner circle."

The inner circle should not be closed - as others develop a reputation
for creating patches that improve InterBase, they will join the circle.

I'm looking for a way that people from the original group can serve
as mentors - not just on an "as available" basis, but consistently,
especially in the beginning. The InterBase "engine" is a complex
little beastie and doesn't take kindly to abuse. Being an old hand,
I consider some of what has been done in the Borland years to be
abuse and would like to see it undone. Perhaps someone without my
background (or Charlie's or Deej's or Dave's or Jim's) could sort
the wheat from the chaff ... but the help of some of the people above
would increase the chance of success. And, like all of you, they
(we) must pay our bills somehow.

>My point isn't that IB/NewBase/whatever *must* work as other open source
>projects, but:
>
>(1) it is wrong to argue that without financially-driven development, an
> open source project cannot succeed and will merely result in chaos
> (history proves that this just isn't true); and,

Again, InterBase starts as a complex product with 15 years of history.
The need for financially-driven development (meaning compensated
senior mentors) will diminish over time, but now, I think the need
exists. Some of the important projects are large - handling
multi-processor, shared memory systems is one - and require co-
ordinated effort.

>(2) if the NewBase project is going to operate differently to other OSS
> projects then this should be made clear to any potential new
> users/developers - if you just say, "This is an open source project" with
> no further explanation then obviously users/developers from other OSS
> projects will come with certain expectations.

When I thought that the company formerly known as ISC would be
responsible for the original release of the open source code,
it had been my expectation that we would define and publish the
protocol we were following. I also expected that we would start
with the funding necessary to pay our experts as we built the
peripheral businesses. Again not so. We start with debts, debts
we incurred anticipating the need to take off quickly and to
make available such items as production quality documentation.

Such is life.

>A project can perfectly meet the "open source definition", thereby correctly
>be called "open source" - but if it's goals/vision/operation are sufficiently
>different to so many other OSS projects, it's only fair to publish them
>somewhere so everyone is clear what to expect. Remember, my focus here is
>*new* users/developers - it's all well and great for existing IB users to have
>decided these things for themselves, but if they ignore newcomers the IB
>community will find it hard to grow beyond its existing userbase.

InterBase should grow beyond its original user base in two
dimensions - first, among the community - developers with the
skill, experience, and generosity to improve the product,
and second among those who are simply looking for a database
that doesn't get in the way. Apache and Linux would be of
much less interest without the latter group - people who want
reliable tools but don't want to get in the business of forging
hammer heads.

Best regards,

Ann