Oleg,
Personally I have an equally lax opinion on releases as you. It is even
laxer when it comes to API changes (I support the Linux kernel way of
big API changes at any time). I was getting so conservative because our
release practice has been so during development of 3.0. Please
understand that I am totally fine with your decision. I live in
Switzerland so I know what democracy is about :-)
Odi
Oleg Kalnichevski wrote:
Odi,
Let us face the simple hard fact that all our quality assurance efforts
are exclusively reactive. If there's a bug report or a post at the
httpclient-user list, we'll react. Otherwise, we just go on with our
daily routine. I do not think this is a major revelation to anyone
vaguely familiar with OSS what all those release versions are purely
fictional. We could have released 3.0rc3 as 3.0-final and 3.0rc4 as
3.0.1. The first RCs that we release are all in fact of solid BETA
quality but certainly nowhere near release quality. There are certain
type of users out there who just never use BETA software as a matter of
principle. That's why we always experience a spike in bug reports as
soon as we release the first RC because that greatly expands our user
base to include all those 'will not touch no stinkin' BETA' type of
folks. We can sit here and see bug reports trickle in at a rate of 2
trivial bug reports per month for the rest of our professional carrier,
or we can declare 3.0 good enough and get our user base massively
increased because all those managers will no longer feel that can get
blamed for using BETA quality software in their VERY important business
applications.
To me there's only one true milestone in OSS, that is, the API freeze.
All that goes beyond that is just an illusion of stability/instability
Evil Comrade Oleg
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
