Re: Server at Paris VIII

[David Kastrup]

"address@hidden" <address@hidden> writes:

> For me, it is faster to write a lot of kludgy code, see what it
> breaks, and then write good code [1] rather than trying to write good
> code from the get-go.  So cheap regtests definitely help.

That sounds uncomfortably like: "It's ok to design stuff above one's
head since we have a procedure for finding most of the easy to diagnose
bugs."

There is a rule of thumb: you can debug code at about half the
complexity that you can design it.  If you design at the limits of your
capacities, you will not be able to debug your code.

And this sounds like you think we have mechanisms helping with designing
code beyond the limits of your capacities.

Maintaining such code usually eventually reaches a point of equilibrium
where each fix breaks as many things as it fixes, partly because the
underlying assumptions are incompatible.

We see this kind of effect in the recent bout of circular dependencies.

-- 
David Kastrup