Fwd: Re: [Uml-devel] Patch implementing classifier scope

[Andrew Sutton]

----------  Forwarded Message  ----------

Subject: Re: [Uml-devel] Patch implementing classifier scope
Date: Wed, 16 Oct 2002 22:17:12 -0400
From: Andrew Sutton <ansutton at kent.edu>
To: Jonathan Riddell <jr at jriddell.org>

On Wednesday 16 October 2002 03:46 pm, Jonathan Riddell wrote:
> > - I did not found any about the specification of "static"s in the
> > specification. Is there anybody who knows this.
>
> It's certainly there.  I can't remember the exact word used (most likely
> static) but operations and attributes which are owned by the class rather
> than an object from that class are underlined.  Constrictor methods are an
> example of this (but not destructor).

yep. it most definitely is. there's two different attributes that appear in a
couple different objects that determin scope. variable names are ownerScope
and targetScope, both of type ScopeKind (values sk_classifier (static) and
sk_instanct (local)).

lets see... a Feature has an ownerScope attribute. Feature is the base class
of both structural (attributes) and behavioral (method and operation)
features. the ownerScope attribute can is available for all three - and yes.
UML differentiates between operation and method.

an assocation end has a targetScope (also scopeKind) that specifies whether
the target of the association (the class to which the end is connected) has
classifier scope or instance scope.

on a side note, i've been doing some brainwork on implementing a template
library to generically implement attributes and relationships, making the
implementation of the actual UML model much, much faster. i've been a little
stuck on the concepts of bidirectional relationship management. not easy. i
won't be able to do much for a while though.

andy

-------------------------------------------------------

-- 
Andrew Sutton
ansutton at kent.edu