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

Re: [Scheme-reports] [r6rs-discuss] redefining eqv?

On Thu, 2010-12-23 at 20:39 -0500, Eli Barzilay wrote:

> Earlier today, Peter Kourzanov wrote:
> > 
> > Still returning to Scheme, I would like to be also in control
> > concerning eqv?, [...] when I don't care about exact semantics and
>                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

What I meant, exact semantics of eqv? Which are already loose enough.

BTW...Can someone with enough grey hairs remember why we have the horde
of predicates like =, eq?, equal? and eqv? What I understood is that
eqv? is sort-of one-size-fits-all idea gone astray

> > using something like _syscall when semantics are at stake.
> Yeah, I'm most definitely in a very different "we" -- I always care
> about "exact semantics" of any code I write, and I certainly hope that
> you do too, if you write any code that runs anything I interact with.

Alright, fair enough. Propose a version of case that could use any
equivalence predicate without having to supply the predicate to every
instance of case. What would that be: Monadic style? 
Something like R6RS hashtable library?

In presence of pattern-matching, equivalence becomes an interesting
dimension. Most matchers use equal? but I wonder if that is always
sufficient. Use-case - matching to detect sharing...

Scheme-reports mailing list