# Re: TRUE and FALSE values in the relational lattice

Date: Thu, 21 Jun 2007 19:57:57 -0000

Message-ID: <1182455877.045340.10100_at_c77g2000hse.googlegroups.com>

On 21 jun, 20:39, "David Cressey" <cresse..._at_verizon.net> wrote:

*> "Vadim Tropashko" <vadimtro_inva..._at_yahoo.com> wrote in message
**>
*

> news:1182446779.883211.130730_at_e9g2000prf.googlegroups.com...

*>
**>
**>
**> > On Jun 21, 3:14 am, Jan Hidders <hidd..._at_gmail.com> wrote:
**> > > I would assume the following operations: (in yet another notation)
**> > > - E * E : natural join
**> > > - E + E : generalized union
**>
**> > Eh, 3 different people, 3 different notations:-)
**>
**> > > - {()} : the relation with empty header and a single empty tuple
**>
**> > OK
**>
**> > > - [A,B,..D] : the empty relation with header {A, B, ..., D} (possibly
**> > > the empty set)
**>
**> > This is nice proposal. How abotut small letters for attributes and
**> > capital letters for relations?
**>
**> Why not the following?
**>
**> {A,B,..D ()}
*

Because that seems to imply it contains the empty tuple, but it contains no tuples. Also if A, B, .., D is empty this becomes {()} but that is already something else, and yes, I do need a symbol for [] (i.e., the empty relation with the empty header).

- Jan Hidders