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

Re: ports as real types...

> > Well, my take on this one is very simple: ports should not be REQUIRED
> > to be disjoint from other types.  The report allows implementors the
> > freedom to implement ports in any way that seems reasonable.  What
> > reason can you advance for their being required to be disjoint?
> Humm.  I'm a bit puzzled by this.  What's the reason for lalowing
> implementors this freedom?

So far as I can tell, there is no technical reason that would not
also apply to, say strings, characters and booleans, yet the latter
three are opaque, as they should be. [Or maybe, they shouldn't be!
If there is enough opposition to making the port type opeque in the
name of freedom, I'll be glad to propose the logical consequence of
this, namely to remove the type disjointness requirement from these
other types.] Note that port type is the *only* type that is left
dangling in the proverbial implementation-dependent wind, an unsafe
and unnecessary situation in my opinion.

Code in haste, repent at runtime. | internet: oz@sis.yorku.ca
	- notebooks of a heretic  | phone:416 736 2100 x33976