Chris Howe wrote:
I haven't looked at the MRP stuff at all.  But I can't fathom why it
would have a problem with determining its scope without relying on the
current convention of a Purchase Order v. Sales Order.  There is no
problem using a denormalized field for convenience or performance sake,
but problems arise when you use a denormalized field where a
fundamental convention exists.  When we override fundamental
conventions, we lose the benefit of the convention.  In addition, when
we write applications that aren't responsible for their scope, every
other application is restricted to the assumptions of the non-scoped
application.  That by definition keeps us from producing a "universal
and generic" ERP.

Sorry, I still think this is dead wrong and if you think it is an important 
change you'll really have to look at the details and see just how much work 
such a change would create (and that aside from the rest of the model changes, 
like eliminating the return, request, quote, etc models). The distinction 
between purchase and sales orders is used all over the place for different user 
interfaces and services and high level business processes. If we didn't have 
that type field we would have a heck of a time figuring it out in things like 
ECA rules and such.

We don't live in an abstract world where we model for theory. Things are 
modeled the way they are to pragmatically represent reality in as generic and 
customizable way as possible. That is the goal anyway, and I don't see how 
these changes help that goal.

Chris: unless you have something that addresses how this would help in a 
pragmatic, rubber-on-the-pavement way then discussing this is a waste of time. 
To be honest the main reason I'm answering is to make sure other people are 
clear to so that no one will start doing stuff like this, especially without 
reviewing the impact and problems it would cause first!

-David


Reply via email to