Re: Database-valued attributes?
Date: Thu, 13 Nov 2003 21:35:03 -0500
Message-ID: <582dnSCLLv7goimiRVn-jg_at_golden.net>
"Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message
news:bovs9n$1dou$1_at_gazette.almaden.ibm.com...
> "Bob Badour" <bbadour_at_golden.net> wrote in message
> news:NqydnbqEFO6qei-iRVn-tw_at_golden.net...
> > "Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message
> > news:botpmf$ooe$1_at_gazette.almaden.ibm.com...
> > > "Bob Badour" <bbadour_at_golden.net> wrote in message
> > > news:GLadnaU5LYVh0i-iRVn-tg_at_golden.net...
> > > > "Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message
> > > > news:bot3ev$oqa$1_at_gazette.almaden.ibm.com...
> > > > > "Bob Badour" <bbadour_at_golden.net> wrote in message
> > > > > news:EZKdnUmgA62gjiyiRVn-tw_at_golden.net...
> > > > > > > poss rep with more than 1 component a tuple of those
components,
> > or
> > > a
> > > > > > list, a set, an array or just 'a thing with components'?
> > > > > >
> > > > > > It is a possible representation. The remainder is internal.
> > > > >
> > > > > Thanks.
> > > > >
> > > > > Can't say I'm happy with such dereliction of definition though.
I'd
> > like
> > > > to
> > > > > see the logical model cover such matters rather than leaving it to
> be
> > an
> > > > > 'internal' matter. Oh well.
> > > >
> > > > Physical independence is not a dereliction.
> > >
> > > One of us in the group said
> > >
> > > "I doubt Dijkstra would consider a division between applications and
> data
> > > management appropriate. I know he expressed skepticism regarding the
> > > division into conceptual, logical and physical."
> > >
> > > Has anyone got the quote? I think I share some of his skepticism...
> >
> > As I recall, it was in one of his trip reports, but I could not find
> > anything by searching so I assume that one has not been transcribed yet.
>
> Maybe the above was inspired by his comments in 1976 on "Database
> Technology" in this trip report
>
> http://www.cs.utexas.edu/users/EWD/ewd05xx/EWD577.PDF
>
> His list of "crazy expressions" does include
> "logical abstractions"
> "mapping of one level of abstraction into the layer below"
> etc
>
> So I guess one could read that as skepticism regarding the division into
> conceptual, logical and physical.
>
> However I think I saw another quote somewhere in which he questioned the
> idea that the 'physical' was not also 'logical'. I.e is 'physical'
therefore
> 'illogical' ?
>
> At the least, I'm not a big fan of the terminology we use. What does the
> word logical, in the context of "logical model" really mean?
The ISO/IEC Standard Vocabularies for IT (ISO/IEC 2382) make that rather clear. The logical level of discourse involves an abstract representation of data suitable for communication or processing without regard to actual media or physical devices. The physical level of discourse involves the actual encodings of data on physical media or devices.
You get an idea--that's conceptual.
You imagine a picture to represent the idea--that's logical.
You draw the picture on acetate--that's physical.
The ink on the acetate is physical as is an overhead projector and a
projection screen.
The picture is logical and is the same picture whether viewed on the acetate
or on a projection screen.
The idea is conceptual.
Also note that (according to the standard vocabularies) information is conceptual and data is logical or physical.
Why these terms? I don't know. We have to call them something. Just as Dijkstra points out that physical is not illogical, neither physical nor logical are inconceivable. Aristotle ran into the same problem when he named everything metaphysical after a book's position on his shelf.
Is a Turing Machine logical or physical? I don't know, and I can think of it as either. Received on Fri Nov 14 2003 - 03:35:03 CET