Re: Strategy

From: Alex Rousskov <rousskov_at_measurement-factory.com>
Date: Mon, 22 Sep 2008 08:35:31 -0600

On Mon, 2008-09-22 at 09:13 +0200, Kinkie wrote:

> I also take your comment as an encouragement to keep on working with
> SBuf. If that's not the case, please stop me now, as while I have some
> time to devote to squid, time is still too precious a resource to
> waste it fruitlessly.

I encourage you to select Squid projects that match your interests and
expertise the best. As for the String project, which you have already
selected and invested in, I hope you will find a way to lovingly
implement valid recommendations, even when you feel that your design
choices are superior.

Again, this only applies to the String project: If a reviewer makes a
mistake, it should, of course, be pointed out and discussed. On the
other hand, if it is a choice between A and B, you did A, the reviewer
strongly recommends B, you know that both are doable, please follow the
reviewer's recommendation without going into "better than the old code",
"just like the old API", "maximum efficiency", "low-level performance
optimization", "code compactness", "private API", "undefined C++ coding
guideline", and similar arguments where it is virtually impossible to
prove anything in a reasonable amount of time.

> > Teamwork requires a balance of responsibilities and trust. The String
> > project has not been a teamwork so far. It was Kinkie doing a lot of
> > hard coding work and then me spending obscene amounts of time fighting
> > with Kinkie over basic C++ principles. No balance, no trust.
>
> There can be no balance, for the reason that I write the code and you
> have the power to veto it. That makes the relationship inherently
> asymmetric.

You are confusing balance with equality or symmetry. We indeed have
different rights and responsibilities, there is no symmetry, but the
system can still be balanced. This is very natural, I think.

> I _do_ trust you though, although it doesn't mean that I'll just take
> your word at face value if I believe have arguments to bring to the
> table.

That's exactly the problem in a project like String. See above for my
suggestions on how to bring the cost of this project down to an
acceptable level.

HTH,

Alex.
Received on Mon Sep 22 2008 - 14:35:46 MDT

This archive was generated by hypermail 2.2.0 : Mon Sep 22 2008 - 12:00:04 MDT