Re: Index management

Joel Garry wrote:

> Beyond necessary. References are a prerequisite for academic
> veracity.
>
> Still not convinced that a few inaccuracies cast doubt on an entire
> body of work. But agree that reproducible test cases are much better
> than bald assertions. But am also concerned that not everything in a
> complex system is reproducible or testable.

The old 'ghost in the machine' idea. True enough, which is why one always allows for exceptions (or should do), and tries not to sound like Moses (difficult, perhaps, to achieve the required tone of moderation when first advancing an idea that flies in the face of what passes for accepted practice, but there eventually, we hope).

What it isn't, however, is two completely contradictory things at the same time, nor a matter of 'high emotion'.

Not being able to pin down every last nuance doesn't mean abandoning any effort to be scientific about gathering test-cases and evidence; or proposing a hypothesis and trying hard to blow it to bits. As, indeed, you say. And as, indeed, I thought Richard's paper was.

Regards
HJR Received on Thu May 06 2004 - 18:52:18 CDT