Re: more closed-world chatter
Date: 7 May 2007 00:34:16 -0700
Message-ID: <1178523256.613959.176320_at_n59g2000hsh.googlegroups.com>
On May 7, 1:56 pm, Jon Heggland <jon.heggl..._at_idi.ntnu.no> wrote:
> David BL wrote:
> > Consider these two types
>
> > T1 = int
> > T2 = { 1,5,17,22,105 }
>
> > I have the feeling it's not such a good idea to treat these in the
> > same way. Otherwise what does it mean to distinguish between type and
> > value? I would rather call T2 a value than a type, and keep the type
> > system simple.
>
> What criterion do you use to decide between "type" and "value"? Isn't T1
> as much a set of values (and thus a value) as T2? Consider T1 = byte and
> T2 = { -128, -127, ..., -1, 0, 1, ... , 126, 127 } ...
The distinction is based on whether a proposed type will in fact tend to be treated as a value at run time, and manipulated as such. I agree this is rather vague.
My concern is that if types can be manipulated as values at run time, then we have the means to solve complex problems using just the type system. In that case why make a distinction between types and values?