[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: transitivity does not imply type-checking

Sebastian Egner <sebastian.egner@xxxxxxxxxxx> writes:

> For clarification, I was seeking advice from one or more of the
> authors of R5RS to find out what they meant when they required
> <= etc. to be 'transitive.'

I found a discussion about transitivity on the rrrs mailing list (google
for "rrrs transitive"). The issue there was the behaviour of the
predicates when applied to a mixture of exact and non-exact args. I
suspect that some Lisps do not guarantee transitivity in that case, and
that is what motivated the comment in R5RS re Lisp.

> Frankly, I am quite happy with this choice because it contributes to
> robustness, more than it hurts performance---unless you use chain<=? 
> on very long lists, in which case I would recommend not using APPLY

I agree. A programmer cannot rely on the short-circuiting behaviour
anyway, since that is not required by R5RS.