Communicating with the public (was Re: Possibility for FreeBSD 4.11 Extended Support)

Bill Moran wmoran at collaborativefusion.com
Fri Dec 22 06:05:06 PST 2006


In response to "Adrian Chadd" <adrian at freebsd.org>:

> On 22/12/06, Bill Moran <wmoran at collaborativefusion.com> wrote:
> 
> > I could be wrong, but I get the impression that this whole EOL issue with
> > 4.x is partly a result of not reminding people when the EOL date for 4.x
> > is every 5 minutes.  The result is that it's just hitting home for a lot
> > of people now that it's the 11th hour.
> 
> The trouble Squid had was its push to a new codebase (2.5 -> 3.0)
> without adequately considering what users wanted. After all, if users
> don't get any of what they want then there's probably no chance of any
> paid work out of it.. Users cried for new features but with the
> stability of the existing codebase. In the end the developers caved
> and provided Squid-2.6 which seems to have begun reinvigorating the
> project somewhat.
> 
> I'm not saying thats the case here, but all the people I've seen
> complain about 4.11 isn't because the upgrade path isn't -there-, its
> because the upgrade path doesn't give them stability. People then
> answer "but its stable for mee!"; both sides don't end up agreeing.
> tsk .:)

Agreed.  The problem is that I'm _not_ seeing any problems.  The result
of this is:
1) I'm not motivated to do anything about it.
2) I don't even know what to do if I was motivated.  Until this week,
   I didn't even know any stability problems existed in post 4.x
   systems until today, so I _couldn't_ do anything about it.

I'm guessing you could say #1 and #2 for any number of developers.

There are rumblings about stability issues.  The problem is there's
very little helpful information.  My prediction is that these problems
will persist until one of the following conditions is met:
1) Someone knowledgeable just gets interested and starts working on
   the problem.
2) Someone who needs these features puts some effort in to gathering
   some truly useful information.
3) Someone who needs these features decides to pay someone knowledgeable
   to work on it.

It's interesting that another party who posted to the list earlier was
complaining about how his stability issues went unfixed, yet he had
_zero_ useful information on where the problem was originating from.
After 5 minutes of searching the PR database, I found an open issue
regarding lockups with quotas.  This other guy never connected the dots?
Never did any diagnosis?  Never added his $.02 to the open PR?

_That_ is why these things aren't getting fixed.

Again, the thing that _absolutely_ boggles my mind is that these folks
want to divert developer support _away_ from fixing these issues and
to supporting legacy software.

Quit bitching and go use Dragonfly.

-- 
Bill Moran
Collaborative Fusion Inc.


More information about the freebsd-stable mailing list