Re: Opinions please : You can't (shouldn't) fix Stupid, remove the Stupid first and then see if you still have a problem.

> On Fri, Oct 29, 2010 at 8:47 AM, James Ashley
> <[email protected]>wrote:
>
>> On Thu, Oct 28, 2010 at 2:25 PM, Johnathan Meehan
>> > The best thing about code that doesn't exist is that it's guaranteed
>> > 100% bug-free. :)
>> >
>> > To my own opinion, I tend to be a little more cautious even when I'm
>> > sure I'm sure. I stick to one thing at a time, make a note of problems
>> I
>> > see whilst fixing the current and get back to them later. My real
>> > concern, especially against a large code base, is missing a connection
>> > and actually making my life harder with multiple bugs on the go (read,
>> > burnt very subtly but very badly before).
>>
>> Yeah. A lot of the time those nasty tangled-up messes are there for a
>> variety of reasons. Cumulative bug fixes and work-arounds for nasty
>> issues.
>>
>
> Sometimes people do things that are amazingly Stupid and make you go
> "WTF!?".
>
> But by far people do the sanest and best thing they know... but being
> merely
> human fail to see where ultimately it all ends.
>
And often we know where it will ultimately lead but don't or can't care
about it.
Either it's (supposed to be...) a short term thing ("in the next release
we'll replace it with a more permanent solution") or you're pushed by time
and budget constraints to make the dirty quick fix because there's no way
you can do a proper job in the time available.


> You may in retrospect and humility examine each step in turn and conclude
> each step was sound, perhaps not perfect, but basically sound.
>
> But the quickest glance at the end result reveals that the result is
> obviously blindingly wrong.
>
Even the most deliberate and careful analysis can lead to something that
in 20-20 hindsight is fundamentally flawed.
I've worked on systems like that, systems that were theoretically perfect
(flawless design and construction according to all the books and
professors, in fact designed by those professors who'd written those books
in at least one case) but just didn't work (performance disasters,
maintenance nightmares, etc.).

> A close reading of human history will reveal many such cases where the
> result was deaths of thousands and tens of thousands.
>
the plague epidemic in medieval Europe comes to mind...
population of a continent drops by over a third over a few years, but it
ultimately leads to the Renaiscance so while bad the end result was
positive (even if the people of the time would rather it never happened no
doubt).