Re: relations aren't types?

From: Marshall Spight <mspight_at_dnai.com>
Date: Mon, 29 Dec 2003 06:07:22 GMT
Message-ID: <uwPHb.66027$VB2.129057_at_attbi_s51>


"Bob Badour" <bbadour_at_golden.net> wrote in message news:J8ednZnlZMCYJ3KiRVn-tA_at_golden.net...
> "Marshall Spight" <mspight_at_dnai.com> wrote in message
> news:cNKHb.64146$VB2.125134_at_attbi_s51...
> > "Bob Badour" <bbadour_at_golden.net> wrote in message
> news:MPOdnepoTvzeP3OiRVn-hA_at_golden.net...
> > >
> > > Sorry. Time is just as atomic as any other type including relation
> types.
> >
> > I'm not clear on why types have to be atomic. I have not seen a
> > reason for it.

>

> It is not a question of whether they have to be. They just are. The value 5
> does not change when we represent it as 5+0j where j is the imaginary unit.

I agree that type integer is atomic. I also agree that byte, boolean, char, and float/double are necessarily atomic. However, every user defined type will necessarily be composed of some combination of other user defined types or system defined atomic types, and could easily be decomposed back into those types. So they are not "atomic" by any definition of atomic I'm aware of.

There is no way to decompose the value 5. (Generally I prefer my examples to use either 3 or 7, but I will just go with the whole "5" thing for now.) But the value 12:34:56 PM *can* be decomposed. For example, one can decompose it into (hour(12), minute(34), second(56)).

> Alpha as an attribute type means the user cannot perform any operations on
> the values, which distinctly limits its utility too; although, not as much
> as omega.

One presumably would have a way to tell what the actual type of the value for the attribute is at any given row.

Marshall Received on Mon Dec 29 2003 - 07:07:22 CET

Original text of this message