Re: More on identifiers

From: David BL <davidbl_at_iinet.net.au>
Date: Sun, 7 Jun 2009 23:36:44 -0700 (PDT)
Message-ID: <c8db9431-eeef-491d-b6b7-2a4ab3d10895_at_r3g2000vbp.googlegroups.com>


On Jun 7, 10:04 pm, JOG <j..._at_cs.nott.ac.uk> wrote:
>
> My first comment would be that the scenario you have described doesn't
> (strictly) necessitate abstract identifiers. Instead you'd have N
> relations, for the N different object types, with each of those
> relations also have an "in_box" attribute. You don't necessarily need
> a "boxes" relation to fully describe the information in full. This
> would mean however that you would need to query N relations to
> determine what was contained in any given box.

I was assuming some of the properties of the items are parameterised on numerical quantities. Even though these quantities would typically be quantised into a finite set of possible values, your suggestion would require many billions of relations making it impractical. Received on Mon Jun 08 2009 - 08:36:44 CEST

Original text of this message