Re: MV Keys

x wrote:
> "Marshall Spight" <marshall.spight_at_gmail.com> wrote in message
> news:1141267829.869557.170030_at_t39g2000cwt.googlegroups.com...
> > David Cressey wrote:
> > > "Marshall Spight" <marshall.spight_at_gmail.com> wrote
> > >
> > > > To my current way of thinking, there are only the two compound types:
> > > > list and set.
> > >
> > > A tuple is a compound type. And I don't think a tuple is either a list
> or a
> > > set.
>
> > An n-ary tuple is an n-ary relation whose cardinality is constrained to
> > be exactly 1. Lots of people hate this idea, but it seems to work
> > very well.
>
> An n-ary tuple is an 2-ary relation whose cardinality is constrained to
> be exactly n. Lots of people hate this idea, but it seems to work
> very well.

I *did* say there were many valid ways to think about it. :-)

This certainly works, but it's not very convenient, and probably not even practical unless you algebra has pivot as an operator.

OTOH, the viewpoint I suggested is both practical and convenient. But these are design issues, not theoretical ones.

Marshall Received on Thu Mar 02 2006 - 11:41:46 CST