Re: OK, it's the matter of database structure rather

In article <hoete393k8ijv13q7iflp4ogei14e093vo@xxxxxxx>,
John Kelly <jak@xxxxxxxxxxxx> wrote:
On Mon, 17 Sep 2007 10:28:16 -0700, Roy Terry <roy@xxxxxxxxxxxxx>
wrote:

John Kelly wrote:

There is no "real" cost of building a stronger algorithm which can
handle any possible, albeit unlikely, load of data, aside from the
extra mental effort required to think things all the way through.

Some costs include:

4. Increased risk of defects based on poor understanding and time
constraints in maintenance.

To "think things all the way through" means better understanding. And
the result is better code with less risk of defect.

There is no defense for lazy programming, claims of cost effective
"engineering" notwithstanding.
.
.
.
Hmmmm.

I'm wondering what defense there is of emotion-laden arguments
that rest on disparagements like "lazy", as opposed to sound
engineering analysis of the system and trade-offs involved in
code quality.

"Better code with less risk of defect" comes at a cost. There's
abundant, even overwhelming, anecdotal evidence that organizations
are unwilling to pay many or most of those costs, in many real-
world situations.

The comp.lang.tcl ethos, I assure you, esteems rigorous analysis
and deep understanding far more than many other public programming
hang-outs you're likely to encounter. They're good enough
engineers, though, also to think carefully about the trade-offs.
.