Re: Mini code-sprint report

On Thu, 2005-10-27 at 01:44 +0200, Henrik Nordstrom wrote:

> For me the situation is somewhat different.
>
> I have customers demanding certain Squid-2 extensions (or to be exact
> extensions to the current STABLE release),

That's my point! We will be stuck with Squid2 until we make Squid3
stable. Customers do not care about the version number; they care about
features and stability. The better you make Squid2, the less likely we
will ever find resources to make Squid3 stable. We are "optimizing"
short-term but losing a lot more long-term (assuming Squid3 is something
worth moving towards, and I think it is).

> As a
> result I am most often working with something remebling the proposed 2.6
> in one way or another.

I do feel your pain and am quite familiar with the situation, but I
think that opening up Squid2 cures symptoms and not the disease. The
primary incentive for Squid sponsors to look at Squid3 is the problems
with Squid2. Squid2 frozen state and associated large number of Squid2
patches is one of the biggest Squid2 problems.

> Getting someone to sponsor development in Squid-3 is as you note extremely
> difficult. But getting them to sponsor development to the current STABLE
> release AND squid-3 tends to work out much better in my experience.

... only if you promise to do more work for about the same amount of
money (in my limited experience), which effectively means that Squid3
code gets half-tested patches and becomes worse.

> Getting someone to sponsor the efforts required to make Squid-3 STABLE is
> not something I see likely. Even is hard get people motivated to
> sponsoring the maintenance of the current STABLE release which comparably
> is a much smaller job.

I am quite certain that if we keep Squid2 frozen and pull our efforts
together, we can find customers willing to CO-sponsor serious Squid3
stability work. We may already have a couple sitting on the fence. If
their perception of Squid2 frozen state changes, they will simply
[continue to] demand Squid2 work instead, and Squid3 will continue to
deteriorate...

$0.02,

Alex.
P.S. There may be a personal angle/bias to this story, I guess. When
discussing Squid-related projects with customers I have been telling
them that Squid2 is frozen and, hence, they have a primary choice of
getting a forever-isolated Squid2 patch or a core Squid3 feature. Now,
if we open up Squid2, it looks like I was essentially lying. Sure, their
patch may wait a year to be accepted into Squid2, but that is a lot
different than "never"! :-(
Received on Thu Oct 27 2005 - 09:26:00 MDT