CLARITY re: challenge: end of life for 6.2 is premature with
Bruce M. Simpson
bms at FreeBSD.org
Thu Jun 5 12:03:07 UTC 2008
Jo Rhett wrote:
> I am suggesting that given that the current bug list for 6.3-RELEASE
> is both (a) too large and (b) breaks things that work fine in 6.2 ...
> that I think pushing 6.2 (the real stable release) into EoL is a bit
> rushed. I sympathize with the development costs of maintaining old
> versions. Again, I will help in any way I can.
I'm sorry to hear about the problems you've been having.
It is worth remembering that FreeBSD is an open source project, and
it's maintained on a best-effort basis -- it is offered for free and
without any warranty. Like any other open source project, risk
management and change management becomes a two-way street, because
that's the trade-off struck with the open source model.
The risks, as well as the benefits, have to be factored in carefully
to your company's technology strategy, as I'm sure you're aware.
I'm very surprised that the 6.3 train has been a big issue for you,
although speaking from the development side of the fence, there are a
lot of moving targets, and vendor support of the OS does play a part.
It is difficult to offer any more specific advice without knowing in
more detail exactly what's causing such problems for you, although I see
you've offered general pointers, the folk directly involved need to be
pointed at direct information.
The FreeBSD Project just doesn't have the resources to do
compatibility testing on the scale of e.g. Windows Hardware Quality
Labs, as I'm sure you are also aware.
I take on board what you say about your organisation holding back on
an upgrade because there are PRs filed for the hardware you use, and
having worked in an investment banking environment, I understand this
level of conservatism is warranted.
However, I point out again: it's the open source model, and where
hardware compatibility is concerned, it really is a case of "suck it and
Always has been, no different anywhere else. Open source requires
user participation. Microsoft run the WHQL because their status as a
going concern depends on it.
I'm pleased to hear about your offer of hardware resources for
developers. However, this is only part of the problem.
To my mind, you need to find the right people, with the right
skills, to deal with the issues, and quite often, those guys are already
in demand, and thus their time can attract a high value. Open source
succeeds because money is not the only motivation.
The alternative is DIY, and that is "the point".
If you need firm guarantees about support, consider contracting with
someone to do that. Many companies using FreeBSD already outsource this
kind of support requirement to 3rd parties. There are also FreeBSD
hardware vendors who support FreeBSD as a platform.
If you want someone to take responsibility, make 'em an offer.
More information about the freebsd-stable