Re: Has E/R had a negative impact on db?

SD wrote:
> On 4/20/2006 9:12 AM, Jon Heggland wrote:
> > JOG wrote:
> >> I don't like entities. In fact I despise entities, as the enemy of good
> >> information philosophy. [...]
> >>
> >> Okay, so for those in the know this isn't an issue and E/R is a useful
> >> tool. But for those not in the know (which appears to be a lot of the
> >> industry) it promotes the fallacy of the Entity/Relationship
> >> distinction, of impenetrable wrappers, and encourages the mindset that
> >> has lead to OODBMS, XML databases, etc.
> >
> > Agreed. ORM is much better in this regard. It was a major revelation to
> > me to think in terms of *facts* instead of entities.
>
> E/R is too close to the logical model (in RDBMS) to richly describe and
> discover the problem domain. Does a business analyst understand E/R?
> ORM was a revelation to me, but sadly it appears to be dying because
> tech heads like to think they are god and keep "knowledge" to themselves.
>
> SD

I'm not sure what you mean when you say 'E/R is too close to the logical model (in RDBMS)'. In fact I'd say the exact opposite, I feel that it is too far away.

In RM everything is stored as a relationship (mathematical relation plus a header), and as such it pretty much denies the concept of entities, period. If more people thought like this, whether they use RM or not, we'd have far better multi-user persistence of data. In contrast E/R encourages its user's to think in terms of objects and links, and these artifices seem incredibly detrimental to me. Received on Thu Apr 20 2006 - 19:26:51 CEST