Re: Entity and Identity
Date: Mon, 28 Sep 2009 14:28:50 +1000
Message-ID: <8YWvm.74955$u76.42712_at_newsfe10.iad>
Bob Badour wrote:
> I have no use for pointless extensions.
Neither do I, and I'm not indulging any.
> Given that Halpin formalized and extended Nijssen's text based methods
> with graphical methods,
NIAM is Nijssen's graphical precursor to ORM. Terry didn't create the first graphical representation. And there was other work prior to those, too.
> I fail to see a need to extend them back to text.
Most join constraints (which occur frequently) cannot be expressed in NIAM, FCO-IM, CogNIAM, ORM or ORM2, but are trivial to express in CQL. As are queries, unit conversions, and various other things that either have no precursor in the above technologies, or no graphical representation. None of the tools generate code for the kind of fact-oriented programming API that my project has.
Apart from that, I argue for a number of good reasons why availability of text is an advantage. I'm not going to repeat those arguments here.
> That's nice. I suppose that's why you cite your friend Halpin and ignore
> the real source of his methods.
I didn't come here to give you a history lesson. The metamodel of ORM2 is different *in many ways* from that of CogNIAM, and the two are working hard to come to an agreement at present, more than a decade after one would have thought it complete. In the absence of a published metamodel, I metamodelled ORM2 using NORMA and generate much of my code from that metamodel (which is also published).
> If your work has substance, you will be able to describe at least one
> advance it makes succinctly and without any hand-wavy bullshit or ad
> hominem sales pitches. You already know what I concluded from what you
> have written so far.
-- Clifford Heath, Data Constellation, http://dataconstellation.com Agile Information Management and DesignReceived on Mon Sep 28 2009 - 06:28:50 CEST