Re: Identity modelling (was: dbdebunk 'Quote of Week' comment)

From: Marshall Spight <marshall.spight_at_gmail.com>
Date: 30 Aug 2005 10:46:37 -0700
Message-ID: <1125423997.908155.125310_at_z14g2000cwz.googlegroups.com>


Gene Wirchenko wrote:
> On Tue, 30 Aug 2005 13:08:28 +0300, "x" <x_at_not-exists.org> wrote:
>
> >"David Cressey" <david.cressey_at_earthlink.net> wrote in message
> >news:wIVQe.3985$_84.3105_at_newsread1.news.atl.earthlink.net...
> >
> >> One interesting case is that of the URL. Is a URL a key or a pointer?
> >Does
> >> it name the resource, or does it locate the resource?
> >
> >It is a pointer, not a key.
> >The URL and the resource are not tied together.
>
> It is a key since it does not specify the location. In a way,
> DNS is concerned with translating this key into a pointer.

To me, one of the most important characteristics of a pointer is that it has an operation, dereference, on it, that gets the pointed-to information. There also exists a reverse operation, reference, that gets the pointer for a datum. In C, these are * and & respectively. Note that SQL has no such operations.

It strikes me that GET http://example.com/index.html is a dereference, so I'd call URLs pointers. DNS likewise has pointer operations.

To a large extent it's a matter of perspective, though.

Marshall Received on Tue Aug 30 2005 - 19:46:37 CEST

Original text of this message