Re: VOTE - Apple communit

[richard.b@genie.com (Richard Bennett)]

In article <3692E8D2.7A50F346@cyberhighway.net>, Frank Carney
<frankcNOSPAM@cyberhighway.net> wrote:

> This closed door policy of development is crap.

If I may post a "no flame" response to this, considering I'm propably one
of the people you're referring to.

> If the Apple II
> programmers really want to create high quality, fast release programs
> that "benefit" the Apple community ( what is left of it) then they
> should look at how Linux is developed.

The bazaar in my opinion is not appropriate with the size of the Apple II
developer community.

> Release often (sometimes several
> times a day), allow people access to the code, use internet beta
> testers, delegate tasks, and allow code submissions.

You can count the number of Apple II developers on one hand, and they're
all busy working on their own projects. You wouldn't get any number of
people to delegate programming tasks to, or receive code submissions from.
Remember, most work is in assembly.

I use a small group of tried and true beta testers (ranging from 4 to 15
people generally, depending on the project) who are pretty experienced in
beta testing, or are very familiar with the way I work. Extending past
those people would cause a number of problems:

1. New folks would have to get used to how I work, which means they're
effectively useless in the short term, and I spend time hand holding which
wastes my time.

2. Word gets out about what I'm working on, which slows development. There
IS pressure on you once people know you're working on something. And the
last thing you want to do is have the project die, as there have been too
many over the years that were never completed. I never announce a project
until it is pretty much complete.

3. Its hard to find testers dedicated to the particular project you're
working on. For example, Marinetti requires TCP/IP aware people, with a
current PPP connection, and lots of knowledge on setting up connections
and protocols.

It is also my experience that most (not all, but most) people who want to
beta test, only want to do it so they can get a free copy of the software,
and don't really care about looking for bugs or beating on the program.
(My internal beta testers are not like this by the way)

> The result is that
> commercially developed software has a 15 to 50 percent bug rate while
> Open Source Softwate has a 7 percent bug rate. ( Linux itself has a 9
> percent bug rate.)

By what standard are these percentages? Marinetti 2.0 was released on 27th
September 1998, and not a single bug has been found. Two instances of
strange happennings have been reported, but at this stage its uncertain if
Marinetti is to blame. Not bad for an in-house project, with few testers
and no applications to test with. I put that down mainly to my beta
testers.

> My source of information on stats and the development cycle is here:
> http://www.tuxedo.org/~esr/writings/cathedral-bazaar/

I am familiar with it. In my opinion, Apple II development these days is
best left to a mini version of the cathedral. And even then, its not an
accurate description.

> I hope this gets people thinking about what they want from this
> "community".  To me all I see is mud flinging back and forth about
> really stupid things.  Anyway, I have probably said to much.  Thanks for
> your time.

I already have what I want. The satisfaction that my friends can now give
their Apple IIGS an IP address.

If you'd like to discuss this further, I'd be more than happy to continue
over in A2PRO on Delphi.

Richard Bennett