ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <>
Subject Re: Ant Coding guidelines
Date Thu, 17 Jan 2002 06:57:53 GMT

----- Original Message -----
From: "Tim Vernum" <>
To: "'Ant Developers List'" <>
Sent: Wednesday, January 16, 2002 17:21
Subject: RE: Ant Coding guidelines

> > It's easy to focus people on a simple "screw up and you get
> > woken up" rule,
> > but that doesnt address productivity, only code quality.
> Lucky you.
> We have the reverse.
> "Screw up too many times and we stop calling you."
> It's the ones who are competent that keep getting called
> to fix things.

That's an expectation management issue. They expect you to be reachable; you
need to manage not to be :-)

I could not get over the v-mail I had collected while climbing one of the
local mountains; I had my mobile with me in case I got into trouble, not in
case they did, and there was no way it was going to be switched on. It wasnt
that they had much of an issue anyway, they had just been getting

I did make a tactical error on the week I took off by letting slip I was
going away on the sunday am; I should have turned up at work on friday with
my bike on the car and announced I was driving off at 4pm, which would have
stopped my saturday being wasted in the Cold Room, diagnosing installation
problems like "file server clock in completely different time zone".

My new expectation management theory is "never fix a bug in less that 24
hours"; even if it is to a running system, even if it is trivial. This is
not some rigorous 'you must get past soak tests first' theory, though that
is what I can claim; it is a "stops people harassing me about whether the
bug is fixed or not" tactic...the ease of web service deployment has led
mgt. to believe that any bug can be fixed in fifteen minutes, see.


[I will be talking about this at the WebServices DevCon in March, for the
curious, "When Web Services Go Bad"...The slides will be up on the web some
time later]

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message