home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.object
- Path: sparky!uunet!portal!ntmtv!irvine
- From: irvine@ntmtv.UUCP (Chuck Irvine)
- Subject: Re: Thoughts on the Progression from SA/SD Models to OO Models
- Message-ID: <1992Jul30.185737.5062@ntmtv>
- Sender: irvine@ntmtv (Chuck Irvine)
- Nntp-Posting-Host: nmtvs318
- Organization: Northern Telecom Inc, Mountain View, CA
- References: <1992Jul28.021455.3322@projtech.com> <1992Jul28.175326.10956@ntmtv> <1992Jul30.025233.19890@mole-end.matawan.nj.us>
- Date: Thu, 30 Jul 1992 18:57:37 GMT
- Lines: 34
-
- In article <1992Jul30.025233.19890@mole-end.matawan.nj.us>, mat@mole-end.matawan.nj.us writes:
- |> > I think a significant aspect of the Information Model as
- |> > described in Shlaer-Mellor is the fact that it (seems) heavily influenced
- |> > by the dictates and constraints of relational database modelling techniques.
- |> > In my opinion, the relational view of data is not as general or natural as
- |> > the object oriented view. While in some cases a relational model of data
- |> > will be most appropriate, in other cases it surely will not be. I'm a bit
- |> > concerned that the Shlaer-Mellor might force, or at least coerce,
- |> > implementations to adopt the relational modelling of data.
- |>
- |> > Any comments?
- |>
- |> I think that the E/R models of information and data transcend SA/SD,
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
- |> Codd's Normal Forms, OO, and other approaches. My view, held for only
- ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
- I agree. You seem to equate E/R modelling with relational database modelling.
- It is true that the first step in building a relational model is building
- an E/R model. However, the normalization of data for the relational model
- is not necessarily part of E/R, and it is this normalization which Shlaer-Mellor
- prescribes that I'm concerned with.
-
-
- |> about a year but not seriously challenged by anything, is that they
- |> these models are as fundamental to software engineering as circuit
- |> theory is to electrical engineering, or stochiometry to chemical
- |> engineering. (Which is not to denigrate SA/SD, nth Codd Normal Form,
- |> OO, or any of the other approaches ...)
- |> --
- |> (This man's opinions are his own.)
- |> From mole-end Mark Terribile
- |>
- |> mat@mole-end.matawan.nj.us, Somewhere in Matawan, NJ
-