[PEAK] Rules clarification

[Phillip J. Eby]

At 09:02 AM 2/28/2008 +0200, Sergey Schetinin wrote:
>It also seems that there's some bug triggered by declarations that
>mean nothing, for example:

I wouldn't say it's a bug - your code has an error.  Granted, it 
would be better to catch it sooner.  :)

The problem here is that the default tuple-based engine doesn't have 
any code that notices when it has invalid data, which then causes 
problems later when you switch to the predicate engine further 
below.  (The predicate engine will fail immediately if it sees invalid data.)

Adding more sophisticated validation is possible, but I'm a bit 
concerned about redundancy, since the additional machinery will exist 
only to create errors where none currently occur.  :)