Re: theoretical question on the RDBMS

This is a good point. Is the concept of "file" part of the theory? If having one file with the data and programs violates some principle, then what's the definition of "file"? Does it violate the same principle if the data and programs are on the same disk partition, or the same "file system," or the same physical storage device? Does it matter if they are on the same virtual something but not the same physical something? I can't imagine the theorists worrying about hardware like this (but point me to where they do, if they do). For the record, the word "file" doesn't appear anywhere in the SQL-92 specification.

I don't see where files and devices say anything about independence, and I suspect Paul is right that insofar as the theory talks about independence, it almost has to mean something logical, not physical.

Steve Kass
Drew University

Paul wrote:

> "Alan" <alanshein_at_erols.com> wrote in message news:<ajb18p$19emf7$1@ID-114862.news.dfncis.de>...
> > Access is a desktop application that supplies tools and a data structure
> > with which one can place and retrieve data. Because the tools and data are
> > both in the same layer (the .mdb file), it violates one of the most basic
> > principles of an RDBMS- independence of data and programs. It also supports
> > ODBC connectivity to RDBMSes.
>
> I thought relational theory just required that data and programs be
> independent at the logical level? So that physical details such as
> whether it is all stored in a single .mdb file are irrelevant. Access
> has a database engine (JET?) which can be used independently of the
> rest of the Access tools I think.
>
> Of course, relational purists will argue that SQL isn't even
> relational and thus any DBMS that uses SQL can't be relational...
>
> Paul.
Received on Wed Aug 14 2002 - 14:18:40 CEST