Re: An object-oriented network DBMS from relational DBMS point of view

On Mar 19, 10:07 am, Alexandr Savinov <s..._at_host.com> wrote:
> Marshall schrieb:
>
>
>
> > On Mar 18, 6:04 am, Bob Badour <bbad..._at_pei.sympatico.ca> wrote:
> >> Drago Ganic wrote:
>
> >>> P.S.: I personally differentiate between code and data. To me data is like
> >>> mass in physics. It is static and has structure. I move it around, destroy
> >>> it etc.
> >> I suggest you find a way to check out the standard vocabularies. One can
> >> move code, and in fact "code motion" is an important concept. One can
> >> destroy it, create it etc. While code and data have very different
> >> meanings, the above nonsense is just nonsense.
>
> > Amplifying:
>
> > All code is data. An unparsed Fortran source file is data. Code
> > compiled into Java bytecodes is data. An x86 object file,
> > a PowerPC executable, an abstract syntax tree: all are data.
>
> No, code and data are fundamentally different. This difference is
> analogous to the difference between references and objects (identities
> and entities).

Why would you even put your arguments that way? You know most on cdt would wholly disagree that identity exists outside of attributes. You appear stuck in thinking of wrapped up 'entities' with 'links' between them, as though those two things are somehow different constructs. The simple "is married to/marriage" example surely shows that distinction is blurry /at best/.

> Storing code and data in ordinary files is only an
> implementation issue (rather bad solution by the way).
>
> RM does not use such a thing as code or reference or identity or any
> other representation/access means and hence this difference between code
> and data is not important (it does not exist actually). Yet it makes our
> life very difficult and is the cause of serious problems in data modelling.
>
> > Marshall
>
> --http://conceptoriented.com
Received on Mon Mar 19 2007 - 12:16:15 CET