On 03-08-11 11:52, Bob Tarling wrote:
> In argouml we manage well-formedness rules in 3 ways
>
> Enforce/Restrict - The system refuses to let us modify the model if
> it'll break a WFR - this restriction is what the user considers a
> defect as reported in issue 6332
>
> Enforce/Mutate - The systems modifies the model to compensate for some
> other change to keep the model well-formed. An example would be when
> trying to make both ends of an association aggregate. When making the
> second end aggregate argouml will change the first end to
> non-aggregate.
>
> Critics - The user can break the WFR but will be informed by critic
> and will have a wizard to fix. An example is duplicate names of
> classes in a namespace.
>
> Regarding Issue 6332 I think we should just go for critics at least in
> the short term until we have some way to mutate the model developed.
>
> The user is right in what he says in that issues anyway - we can
> already break that WFR by just moving a class to another package after
> the association is drawn.
>
> Are we in agreement? Can anyone provide the critic/wizard for this?
>
> Any thought?
>
> Bob

See also my reply in issue 6332.

A remark though: The 3 ways of handling WFRs do not exclude each other:

- we need a Critic even if we completely enforce - since an XMI import 
can cause the WFR to be broken anyhow.

- in the case of issue 6332 we enforce the WFR in the combo for 
usability and performance reasons. Let's not throw that away, but 
instead offer an additional method (for advanced users) that allows to 
break the WFR.

------------------------------------------------------
http://argouml.tigris.org/ds/viewMessage.do?dsForumId=450&dsMessageId=2808762

To unsubscribe from this discussion, e-mail: 
[[email protected]].
To be allowed to post to the list contact the mailing list moderator, email: 
[[email protected]]

Reply via email to