[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Squeeze Frozen, NOT



On Mon, 1 Mar 2010 13:14:48 -0500 (EST), Andrei Popescu wrote:
> On Mon,01.Mar.10, 11:28:36, Stephen Powell wrote:
>> Timed release cycles for Debian is a bad idea.  
>> ...
>> Trying to go to timed release cycles
>> is either futile or fatal.
>
> There is no plan for timed release cycles, only timed freezes.

Hmm.  Well, I suppose that timed freezes is better than timed
release cycles, but it still breaks the traditional concept of
"release goals", does it not?  Historically, a new release had
a set of release goals associated with it.  With timed release freezes, that
pretty much throws release goals out the window.

This reminds me of the
"extended Heisenberg uncertainty principle" for programming.  In the
traditional Heisenberg uncertainty principle of atomic particle physics,
there are two quantities, position and momentum, that cannot be known with
certainty.  The more certain you are of a particle's position, the less
certain you are of its momentum, and vice versa.

The extended Heisenberg uncertainty principle of programming, one of the
corollaries of Murphy's law, involves three quantities: time,
resources ($), and project.  (1) If there is a precisely-known
problem that must be solved and a precisely-known time frame within
which this task must be completed, then one cannot guess how much it will
cost.  (2) If there is a precisely-known problem that must be solved and
a fixed amount of resource is available, then one cannot guess how long
it will take.  (3) If there is a fixed amount of time allotted for the
project and a fixed amount of resource is available, one cannot guess how
much of the project will be completed.  If one is lucky enough to
accurately predict all three variables in advance, then what one deals
with is not in the realm of programming!

A manager of a programming staff at a company used to have a sign on
his wall that read something like this: "Everyone who comes to me wants
his project done (a) fast, (b) cheap, and (c) high quality.  I tell
them to think it over, pick two of the three, and call me back."
That pretty well sums it up.  You want it bad?  You get it bad.
You want it really bad?  You get it really bad.  The worse you want
it, the worse you get it.

Since the resources are fixed at $0 (donated free labor) and the time
interval is fixed at one year, the release goals go out the window.
You just can't eat your cake and have it too.


Reply to: