Jens Axel Søgaard <xxxxxx@soegaard.net> writes:

> I'd prefer as many errors to be signaled as possible, but I guess that's
> a matter of taste.
>
> What's your view?

I agree with the general principle. In SISC we *signal* errors in almost
all of the situations that R5RS just states "it is an error". The n-ary
predicates are one of the exceptions, because forcing type checking of
all args changes the average O(...) complexity of the predicates for the
worse. We are still debating the issue though, which is why I have taken
such an interest in that particular aspect of srfi 67.

Matthias.