Re: no names allowed, we serve types only

On Feb 25, 9:39 am, David BL <davi..._at_iinet.net.au> wrote:
> On Feb 24, 5:22 pm, Jan Hidders <hidd..._at_gmail.com> wrote:
> > On 23 feb, 18:08, David BL <davi..._at_iinet.net.au> wrote:

> > > C.Date presents this argument very well in section 20.9 of an
> > > Introduction to Database Systems where he claims that a coloured
> > > circle is not a subtype of circle (or vice versa).
>
> > The tuple that represents the circle is not the same thing as the
> > circle itself.
>
> Agreed, but I think the same (informal) reasoning applies to tuples.

I'll clarify that. Consider the following principle:

  Let T1, T2 be two given data types satisfying T1 <= T2.   Let there be a given encoding able to represent each   value of type T2 on a computer. Then that same   encoding is necessarily also able to represent each value   of type T1.

I believe this principle implies both:

      not coloured_circle <= circle

and not <a:t1,b:t2,c:t3> <= <a:t1,b:t2> Received on Thu Feb 25 2010 - 05:26:19 CET