Re: builtins operating on rif:iri

> The only currently defined "built-in" that accepts rif:iris is
> pred:iri-string...

Well, it's the only one that works with iris as iris.  All of them
syntactically accept rif:iris, and will work with the literal they are
equated to, if at some point they are equated to a literal.
Until/unless that equation happens, the reasoner can't do anything with
them...

There was this test that Michael confirmed [1]:

         ex:foo = 1
     entials
         external(numeric-equal(ex:foo, 1))

And here's another:

PREMISE:
     <http://example.com/foo> = 1
     p :- iri-string(?x, 'http://example.com/foo') and
          numeric-equal(?x, 1)
CONCLUSION:
     p.

Right?

But.....  yeah, I guess they don't mess up safety.  (I was confusing
myself by thinking about them as unbound variables, but they're
(roughly) existentialy quantified variables, and it's the universally
quantified ones that are a safety issue.)

      -- Sandro

[1] http://lists.w3.org/Archives/Public/public-rif-wg/2009Apr/0121.html