Re: More on lists and sets

From: vc <boston103_at_hotmail.com>
Date: 22 Mar 2006 10:51:58 -0800
Message-ID: <1143053518.460841.305470@g10g2000cwb.googlegroups.com>

Mikito Harakiri wrote:
> vc wrote:
> > Mikito Harakiri wrote:
> > > vc wrote:
> > > > Mikito Harakiri wrote:
> > > > > It is remarkable that list intersection laws mimic the ones for the
> > > > > union:
> > > > > A /\ A != A
> > > >
> > > > This is not correct.
> > >
> > > Let A = 1->1. Then A /\ A = 1->1->1->1.
> >
> > According to your unknown private definition, maybe, but that's not
> > how list intersection is usually defined.

>

> Let's focus on bags, first. Do you agree that
>
> {1,1} /\ {1,1,1} = {1,1,1,1,1,1}

If '/\' means 'intersection', then, no. An accepted bag intersection definition would result in {1,1} (Ullman, minimum number of occurences).

>

> ? (Intersection is join. Join is cartesian product followed by
> selection).

Intersection aint no join, it's a much more primitive set-theoretical operation.

>

> As we descend down the chain lists-bags-sets, the list intersection
> operation has to be compatible with the bags one.

Why 'has it to' ? They are different structures serving different purposes. If it does not suit your purpose, you can invent a different operation called 'my_intersection', why mis/abuse the simple and intuitively clear 'intersection' ?

>The standard list

> intersection does not. Besides, the second argument in the standard
> list intersection behaves like a set, not list.
Received on Wed Mar 22 2006 - 12:51:58 CST