Re: What databases have taught me

Tony D wrote:
> topmind wrote:
> > After thousands of debates and failed attempts to find objective
> > metrics beyond execution speed and matching stated external
> > requirements, I am leaning toward the "art" viewpoint. One man's
> > spehgetti code is another man's masterpeice. Just because I find it a
> > flaming tangled mess does not mean the next guy will. Related:
> >
> > http://www.geocities.com/tablizer/science.htm
> >
>
> Ye gods. I'll limit my response to that to a comment on the metrics.
> How about "correctness" ?

Yes, but as pointed out in the link, there are multiple paths to the same "answer". Nobody is defending something that produces the wrong answer here.

> Or, "provability" ?

Very few are attempting to use formal proofs to prove their software correct. It has generally been easier to have 3 teams write different code based on the same specs and then have the system use majority rules if one is off.

> Or, "ability to reason
> about what this software actually does,

This is rather subjective, though. I've encountered people who could reason about Goto's far better than I could. Somebody will invariably say, "If you were smart like me, you could reason about my code." I guarentee it.

> without resorting to stuffing
> some more-or-less random test cases through it as some kind of
> demonstration that it kind-of, maybe, perhaps does what we want it to,
> for these semi-random test cases at least ?"

I have kicked around approaches to measure the code impact of various change scenarios. The problem is that people also perceive change differently, I've found out, such that they would assign different frequency estimates, which were required to get a total score.

-T- Received on Sun Jul 09 2006 - 02:38:55 CEST