Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » UML2 Tools » [ClassD] Is it possible to show...
[ClassD] Is it possible to show... [message #475931] Mon, 24 November 2008 21:50 Go to next message
Nicolas Rouquette is currently offline Nicolas Rouquette
Messages: 135
Registered: July 2009
Senior Member
....the following with the M3 build of UML2Tools:

a) name & language of a constraint.

I can show a the definition of a constraint with the menu action:
"Create Shortcut..." but I don't see:
- the language used for specifying the constraint (e.g., OCL, Java, ...)
- the name of the constraint. Sometimes, fully-qualified names are too
long. It would be nice to show:

- the fully-qualified name in reverse order, or
- 1,2,..N levels of name qualification, enough to tell which element on
the diagram the constraint belongs, or
- a line between the element that owns the constraint and the shortcut
to the constraint

b) generalizations/associations across packages

Suppose I have:
- package P1 with class A
- package P2 with class B
- P2::B has a generalization relationship to P1::A

If I initialize a diagram for P2, the diagram automatically shows *all*
of the contents of P2. For small packages, this is OK but it is
definitely too much clutter for large packages.

I can create a shortcut to P1::A in the P2 class diagram.
However, I can't create either a shortcut or show the existing
generalization relationship from P2::B to P1::A

The same comment applies to other kinds of relationships such as
associations.

-- Nicolas.
Re: [ClassD] Is it possible to show... [message #475938 is a reply to message #475931] Tue, 25 November 2008 00:25 Go to previous messageGo to next message
Tatiana Fesenko is currently offline Tatiana Fesenko
Messages: 530
Registered: July 2009
Senior Member
Hello Nicolas,

Thank you for the interesting questions.

> a) name & language of a constraint.
>
> I can show a the definition of a constraint with the menu action:
> "Create Shortcut..." but I don't see:
> - the language used for specifying the constraint (e.g., OCL, Java,
> ...)
In accordance with UML specification, it depends on UML Tool whether to show
language of an OpaqueExpression or not. Anyway, I submitted request #256356
'Optionally display language of the OpaqueExpression on the diagram'

> - the name of the constraint. Sometimes, fully-qualified names are too
> long. It would be nice to show:
> - the fully-qualified name in reverse order, or
> - 1,2,..N levels of name qualification, enough to tell which element
> on
> the diagram the constraint belongs,
Name of a constraint is displayed and editable in the properties view. The
name is not shown on the diagram.

> - a line between the element that owns the constraint and the shortcut
> to the constraint
Constraint#ConstrainedElement reference is a relationship between a Constraint
and the elements required to evaluate the constraint specification. The link
is diaplyed on the diagram and can be created using 'ConstrainedElement'
palette entry.

> b) generalizations/associations across packages
>
> Suppose I have:
> - package P1 with class A
> - package P2 with class B
> - P2::B has a generalization relationship to P1::A
> If I initialize a diagram for P2, the diagram automatically shows
> *all* of the contents of P2. For small packages, this is OK but it is
> definitely too much clutter for large packages.
Yes, I agree, that adding all elements to initialized diagrams can make the
diagrams unusable. That's the reason why request #199731 'Structure diagrams
synchronization' was created and fixed. Starting with Galileo M3 build, you
can show/hide particular elements from the diagram using 'Synchronize...'
action.
Besides, a new page 'Diagram Synchronization' was added to 'Initialize UMLClass
diagram' wizaed. The page allows you to select diagram contents and its synchronization
mode.

> I can create a shortcut to P1::A in the P2 class diagram. However, I
> can't create either a shortcut or show the existing generalization
> relationship from P2::B to P1::A
>
> The same comment applies to other kinds of relationships such as
> associations.
Hmm. I repeated the case you descried and it works completely fine.
I've created package P1 and class A in it, then created package P2 and class
B in it. Then I added a shortcut to class A to diagram P2, and created a
generalization from B to A.
Then I initialized diagram P2, added shortcut to class A, and the generalization
was displayed automatically.
And vice-versa, I initialized diagram P1, added shortcut to class B, and
again, the generalization was shown on the diagram.
Nikolas, how did you create a generalization? is there anything in your error
log?

Best wishes,
Tanya.
Re: [ClassD] Is it possible to show... [message #475939 is a reply to message #475938] Tue, 25 November 2008 05:01 Go to previous messageGo to next message
Nicolas Rouquette is currently offline Nicolas Rouquette
Messages: 135
Registered: July 2009
Senior Member
Hi Tatiana,

Tatiana Fesenko wrote:
> Hello Nicolas,
>
> Thank you for the interesting questions.

You're welcome.

>
>> a) name & language of a constraint.
>>
>> I can show a the definition of a constraint with the menu action:
>> "Create Shortcut..." but I don't see:
>> - the language used for specifying the constraint (e.g., OCL, Java,
>> ...)
> In accordance with UML specification, it depends on UML Tool whether to
> show language of an OpaqueExpression or not. Anyway, I submitted request
> #256356 'Optionally display language of the OpaqueExpression on the
> diagram'

Thank you.

>> - the name of the constraint. Sometimes, fully-qualified names are too
>> long. It would be nice to show:
>> - the fully-qualified name in reverse order, or
>> - 1,2,..N levels of name qualification, enough to tell which element
>> on
>> the diagram the constraint belongs,
> Name of a constraint is displayed and editable in the properties view.
> The name is not shown on the diagram.

I updated #256356 asking to show the last k members of the fully
qualified name of any NamedElement on a diagram.

>> - a line between the element that owns the constraint and the shortcut
>> to the constraint
> Constraint#ConstrainedElement reference is a relationship between a
> Constraint and the elements required to evaluate the constraint
> specification. The link is diaplyed on the diagram and can be created
> using 'ConstrainedElement' palette entry.

It looks to me that the behavior you describe depends on whether
Constraint#ConstrainedElement is an object that is shown as a Shape
(e.g., a Class) or as a simple label (e.g., an operation).

If Constraint#ConstrainedElement is a class, then the link is displayed
but only after I close the diagram and then re-open it.

If Constraint#ConstrainedElement is an operation, then there is no link
shown at all, even if I close/re-open/refresh the diagram.

Is this correct?

>> b) generalizations/associations across packages
>>
>> Suppose I have:
>> - package P1 with class A
>> - package P2 with class B
>> - P2::B has a generalization relationship to P1::A
>> If I initialize a diagram for P2, the diagram automatically shows
>> *all* of the contents of P2. For small packages, this is OK but it is
>> definitely too much clutter for large packages.
> Yes, I agree, that adding all elements to initialized diagrams can make
> the diagrams unusable. That's the reason why request #199731 'Structure
> diagrams synchronization' was created and fixed. Starting with Galileo
> M3 build, you can show/hide particular elements from the diagram using
> 'Synchronize...' action. Besides, a new page 'Diagram Synchronization'
> was added to 'Initialize UMLClass diagram' wizaed. The page allows you
> to select diagram contents and its synchronization mode.

Wow, it's nifty, pretty and handy!

Thank you *very* much for implementing this feature!

>> I can create a shortcut to P1::A in the P2 class diagram. However, I
>> can't create either a shortcut or show the existing generalization
>> relationship from P2::B to P1::A
>>
>> The same comment applies to other kinds of relationships such as
>> associations.
> Hmm. I repeated the case you descried and it works completely fine. I've
> created package P1 and class A in it, then created package P2 and class
> B in it. Then I added a shortcut to class A to diagram P2, and created
> a generalization from B to A. Then I initialized diagram P2, added
> shortcut to class A, and the generalization was displayed automatically.
> And vice-versa, I initialized diagram P1, added shortcut to class B, and
> again, the generalization was shown on the diagram. Nikolas, how did you
> create a generalization? is there anything in your error log?

There are no errors in the log; however, I start from an existing model
instead of creating one from scratch.

I hope that you can reproduce the problem with the following detailed
instructions:

a) Check out from dev.eclipse.org:/cvsroot/modeling the project:
org.eclipse.mdt/org.eclipse.uml2/plugins/org.eclipse.uml2.um l

b) Rename this project cvs.org.eclipse.uml2.uml to avoid creating
confusion with the installed org.eclipse.uml2.uml plugin.

c) navigate to: /cvs.org.eclipse.uml2.uml/model/

There, you'll find a bunch of *.uml resources.

If you had imported the installed org.eclipse.uml2.uml plugin, you
wouldn't have these *.uml resources unfortunately.

Now, you can explore two scenarios.

The first scenario is simple but it is long.
The second scenario is not so simple but it is short.

Scenario 1:

1) Select Infrastructure.uml

2) Initialize a class diagram for
Infrastructure::Core::Abstractions::Namespaces

You get a nice class diagram with two classes: Namespace and NamedElement.

3) In the diagram, create a shortcut for
Infrastructure::Core::Abstractions::Ownership::Element

THe diagram shows Element with its A_ownedElement_owner association.

The generalization relationship from ...::Core::Namespaces::NamedElement
to ...::Core::Ownerships::Element isn't shown.

4) Save and close the diagram.

5) Reopen the diagram.

Ta dah... now you can see the generalization relationship.

I've noticed that this behavior happens with relationships between
regular shape nodes and shortcut shape nodes.

Scenario 2:

1) Select Superstructure.uml

2) Initialize a class diagram for Superstructure::Classes::Kernel

3) Synchronize...

select: Element, Comment, DirectedRelationship, Relationship.

4) Arrange the elements similar to figure 7.3 of UML 2.2

That figure shows Element and Comment twice.
No problem...

5) We can create shortcuts!

Create a shortcut for Superstructure::Classes::Kernel::Element
Create a shortcut for Superstructure::Classes::Kernel::Comment

6) Ok, but now what?

Currently, the associations between Element and Comment
(A_annotatedElement_comment and A_ownedComment_owningElement) are shown
between the same Element and Comment node shapes.

To replicate fig 7.3 in UML2Tools' class diagram, we have to visualize
the associations A_annotatedElement_comment and
A_ownedComment_owningElement with different pairs of shape nodes for
Element and Comment.

Since these associations already exist, we are not creating new
associations; instead, we need to do somehting like this:

- delete one of the two associations from the diagram
- select the two shape nodes between which we want to display a specific
association (similar to synchronize but for a pair of nodes instead of
the whole diagram)

-- Nicolas.


> Best wishes,
> Tanya.
>
>
Re: [ClassD] Is it possible to show... [message #475964 is a reply to message #475939] Thu, 04 December 2008 20:55 Go to previous messageGo to next message
Tatiana Fesenko is currently offline Tatiana Fesenko
Messages: 530
Registered: July 2009
Senior Member
Hi Nicolas,
Sorry for the delay with the answer, I was on vacation.

> I updated #256356 asking to show the last k members of the fully
> qualified name of any NamedElement on a diagram.
Great!

>>> - a line between the element that owns the constraint and the
>>> shortcut to the constraint
> It looks to me that the behavior you describe depends on whether
> Constraint#ConstrainedElement is an object that is shown as a Shape
> (e.g., a Class) or as a simple label (e.g., an operation).
> If Constraint#ConstrainedElement is a class, then the link is
> displayed but only after I close the diagram and then re-open it.
I think you set a reference in the properties view, right? If so, it's known
GMF problem - https://bugs.eclipse.org/bugs/show_bug.cgi?id=254871. Refresh
with F5 should help.
Alternatively you can use 'Constrained Element' palette entry, it doesn't
require refresh.
>
> If Constraint#ConstrainedElement is an operation, then there is no
> link shown at all, even if I close/re-open/refresh the diagram.
>
> Is this correct?
You are correct, contraints should be connectable to operations. There is
a similar request in bugzilla - https://bugs.eclipse.org/bugs/show_bug.cgi?id=235158
..

>> ... request #199731
>> 'Structure diagrams synchronization' was created and fixed.
> Wow, it's nifty, pretty and handy!
>
> Thank you *very* much for implementing this feature!
Yes, it is. Thanks to Michael Golubev who implemented it. I personally also
like this feature a lot. The good news, that many improvements have been
done after M3, so in M4 it will be even better.

> I hope that you can reproduce the problem with the following detailed
> instructions:
> Scenario 1:
Hm, the generalization was shown immediately in my workspace. Could you please
submit a request into bugzilla?

> Scenario 2:
> ...
> 4) Arrange the elements similar to figure 7.3 of UML 2.2
> That figure shows Element and Comment twice.
Yes, that would be great if the user could manage elements and links between
such elements this way.

> 5) We can create shortcuts!
> Create a shortcut for Superstructure::Classes::Kernel::Element Create
> a shortcut for Superstructure::Classes::Kernel::Comment
Success!!!
>
> 6) Ok, but now what?
> To replicate fig 7.3 in UML2Tools' class diagram, we have to visualize
> the associations A_annotatedElement_comment and
> A_ownedComment_owningElement with different pairs of shape nodes for
> Element and Comment.
>
> Since these associations already exist, we are not creating new
> associations; instead, we need to do somehting like this:
>
> - delete one of the two associations from the diagram
> - select the two shape nodes between which we want to display a
> specific
> association (similar to synchronize but for a pair of nodes instead of
> the whole diagram)
Alternatively you can reroute source or target of the element.

Tanya.
Re: [ClassD] Is it possible to show... [message #475965 is a reply to message #475964] Thu, 04 December 2008 22:24 Go to previous message
Tatiana Fesenko is currently offline Tatiana Fesenko
Messages: 530
Registered: July 2009
Senior Member
Nicolas,

>> To replicate fig 7.3 in UML2Tools' class diagram, we have to
>> visualize
>> the associations A_annotatedElement_comment and
>> A_ownedComment_owningElement with different pairs of shape nodes for
>> Element and Comment.
>> Since these associations already exist, we are not creating new
>> associations; instead, we need to do somehting like this:
>>
>> - delete one of the two associations from the diagram
>> - select the two shape nodes between which we want to display a
>> specific
>> association (similar to synchronize but for a pair of nodes instead
>> of
>> the whole diagram)

I created a sample diagram for the given figure and put it to CVS and wiki
: http://wiki.eclipse.org/MDT-UML2Tools_FAQ#Can_I_display_a_si ngle_element_several_times.2C_like_it_is_shown_on_figure_7.3 _of_UML_2.2_specification.3F

Tanya.
Re: [ClassD] Is it possible to show... [message #623194 is a reply to message #475931] Tue, 25 November 2008 00:25 Go to previous message
Tatiana Fesenko is currently offline Tatiana Fesenko
Messages: 530
Registered: July 2009
Senior Member
Hello Nicolas,

Thank you for the interesting questions.

> a) name & language of a constraint.
>
> I can show a the definition of a constraint with the menu action:
> "Create Shortcut..." but I don't see:
> - the language used for specifying the constraint (e.g., OCL, Java,
> ...)
In accordance with UML specification, it depends on UML Tool whether to show
language of an OpaqueExpression or not. Anyway, I submitted request #256356
'Optionally display language of the OpaqueExpression on the diagram'

> - the name of the constraint. Sometimes, fully-qualified names are too
> long. It would be nice to show:
> - the fully-qualified name in reverse order, or
> - 1,2,..N levels of name qualification, enough to tell which element
> on
> the diagram the constraint belongs,
Name of a constraint is displayed and editable in the properties view. The
name is not shown on the diagram.

> - a line between the element that owns the constraint and the shortcut
> to the constraint
Constraint#ConstrainedElement reference is a relationship between a Constraint
and the elements required to evaluate the constraint specification. The link
is diaplyed on the diagram and can be created using 'ConstrainedElement'
palette entry.

> b) generalizations/associations across packages
>
> Suppose I have:
> - package P1 with class A
> - package P2 with class B
> - P2::B has a generalization relationship to P1::A
> If I initialize a diagram for P2, the diagram automatically shows
> *all* of the contents of P2. For small packages, this is OK but it is
> definitely too much clutter for large packages.
Yes, I agree, that adding all elements to initialized diagrams can make the
diagrams unusable. That's the reason why request #199731 'Structure diagrams
synchronization' was created and fixed. Starting with Galileo M3 build, you
can show/hide particular elements from the diagram using 'Synchronize...'
action.
Besides, a new page 'Diagram Synchronization' was added to 'Initialize UMLClass
diagram' wizaed. The page allows you to select diagram contents and its synchronization
mode.

> I can create a shortcut to P1::A in the P2 class diagram. However, I
> can't create either a shortcut or show the existing generalization
> relationship from P2::B to P1::A
>
> The same comment applies to other kinds of relationships such as
> associations.
Hmm. I repeated the case you descried and it works completely fine.
I've created package P1 and class A in it, then created package P2 and class
B in it. Then I added a shortcut to class A to diagram P2, and created a
generalization from B to A.
Then I initialized diagram P2, added shortcut to class A, and the generalization
was displayed automatically.
And vice-versa, I initialized diagram P1, added shortcut to class B, and
again, the generalization was shown on the diagram.
Nikolas, how did you create a generalization? is there anything in your error
log?

Best wishes,
Tanya.
Re: [ClassD] Is it possible to show... [message #623196 is a reply to message #475938] Tue, 25 November 2008 05:01 Go to previous message
Nicolas Rouquette is currently offline Nicolas Rouquette
Messages: 135
Registered: July 2009
Senior Member
Hi Tatiana,

Tatiana Fesenko wrote:
> Hello Nicolas,
>
> Thank you for the interesting questions.

You're welcome.

>
>> a) name & language of a constraint.
>>
>> I can show a the definition of a constraint with the menu action:
>> "Create Shortcut..." but I don't see:
>> - the language used for specifying the constraint (e.g., OCL, Java,
>> ...)
> In accordance with UML specification, it depends on UML Tool whether to
> show language of an OpaqueExpression or not. Anyway, I submitted request
> #256356 'Optionally display language of the OpaqueExpression on the
> diagram'

Thank you.

>> - the name of the constraint. Sometimes, fully-qualified names are too
>> long. It would be nice to show:
>> - the fully-qualified name in reverse order, or
>> - 1,2,..N levels of name qualification, enough to tell which element
>> on
>> the diagram the constraint belongs,
> Name of a constraint is displayed and editable in the properties view.
> The name is not shown on the diagram.

I updated #256356 asking to show the last k members of the fully
qualified name of any NamedElement on a diagram.

>> - a line between the element that owns the constraint and the shortcut
>> to the constraint
> Constraint#ConstrainedElement reference is a relationship between a
> Constraint and the elements required to evaluate the constraint
> specification. The link is diaplyed on the diagram and can be created
> using 'ConstrainedElement' palette entry.

It looks to me that the behavior you describe depends on whether
Constraint#ConstrainedElement is an object that is shown as a Shape
(e.g., a Class) or as a simple label (e.g., an operation).

If Constraint#ConstrainedElement is a class, then the link is displayed
but only after I close the diagram and then re-open it.

If Constraint#ConstrainedElement is an operation, then there is no link
shown at all, even if I close/re-open/refresh the diagram.

Is this correct?

>> b) generalizations/associations across packages
>>
>> Suppose I have:
>> - package P1 with class A
>> - package P2 with class B
>> - P2::B has a generalization relationship to P1::A
>> If I initialize a diagram for P2, the diagram automatically shows
>> *all* of the contents of P2. For small packages, this is OK but it is
>> definitely too much clutter for large packages.
> Yes, I agree, that adding all elements to initialized diagrams can make
> the diagrams unusable. That's the reason why request #199731 'Structure
> diagrams synchronization' was created and fixed. Starting with Galileo
> M3 build, you can show/hide particular elements from the diagram using
> 'Synchronize...' action. Besides, a new page 'Diagram Synchronization'
> was added to 'Initialize UMLClass diagram' wizaed. The page allows you
> to select diagram contents and its synchronization mode.

Wow, it's nifty, pretty and handy!

Thank you *very* much for implementing this feature!

>> I can create a shortcut to P1::A in the P2 class diagram. However, I
>> can't create either a shortcut or show the existing generalization
>> relationship from P2::B to P1::A
>>
>> The same comment applies to other kinds of relationships such as
>> associations.
> Hmm. I repeated the case you descried and it works completely fine. I've
> created package P1 and class A in it, then created package P2 and class
> B in it. Then I added a shortcut to class A to diagram P2, and created
> a generalization from B to A. Then I initialized diagram P2, added
> shortcut to class A, and the generalization was displayed automatically.
> And vice-versa, I initialized diagram P1, added shortcut to class B, and
> again, the generalization was shown on the diagram. Nikolas, how did you
> create a generalization? is there anything in your error log?

There are no errors in the log; however, I start from an existing model
instead of creating one from scratch.

I hope that you can reproduce the problem with the following detailed
instructions:

a) Check out from dev.eclipse.org:/cvsroot/modeling the project:
org.eclipse.mdt/org.eclipse.uml2/plugins/org.eclipse.uml2.um l

b) Rename this project cvs.org.eclipse.uml2.uml to avoid creating
confusion with the installed org.eclipse.uml2.uml plugin.

c) navigate to: /cvs.org.eclipse.uml2.uml/model/

There, you'll find a bunch of *.uml resources.

If you had imported the installed org.eclipse.uml2.uml plugin, you
wouldn't have these *.uml resources unfortunately.

Now, you can explore two scenarios.

The first scenario is simple but it is long.
The second scenario is not so simple but it is short.

Scenario 1:

1) Select Infrastructure.uml

2) Initialize a class diagram for
Infrastructure::Core::Abstractions::Namespaces

You get a nice class diagram with two classes: Namespace and NamedElement.

3) In the diagram, create a shortcut for
Infrastructure::Core::Abstractions::Ownership::Element

THe diagram shows Element with its A_ownedElement_owner association.

The generalization relationship from ...::Core::Namespaces::NamedElement
to ...::Core::Ownerships::Element isn't shown.

4) Save and close the diagram.

5) Reopen the diagram.

Ta dah... now you can see the generalization relationship.

I've noticed that this behavior happens with relationships between
regular shape nodes and shortcut shape nodes.

Scenario 2:

1) Select Superstructure.uml

2) Initialize a class diagram for Superstructure::Classes::Kernel

3) Synchronize...

select: Element, Comment, DirectedRelationship, Relationship.

4) Arrange the elements similar to figure 7.3 of UML 2.2

That figure shows Element and Comment twice.
No problem...

5) We can create shortcuts!

Create a shortcut for Superstructure::Classes::Kernel::Element
Create a shortcut for Superstructure::Classes::Kernel::Comment

6) Ok, but now what?

Currently, the associations between Element and Comment
(A_annotatedElement_comment and A_ownedComment_owningElement) are shown
between the same Element and Comment node shapes.

To replicate fig 7.3 in UML2Tools' class diagram, we have to visualize
the associations A_annotatedElement_comment and
A_ownedComment_owningElement with different pairs of shape nodes for
Element and Comment.

Since these associations already exist, we are not creating new
associations; instead, we need to do somehting like this:

- delete one of the two associations from the diagram
- select the two shape nodes between which we want to display a specific
association (similar to synchronize but for a pair of nodes instead of
the whole diagram)

-- Nicolas.


> Best wishes,
> Tanya.
>
>
Re: [ClassD] Is it possible to show... [message #623360 is a reply to message #475939] Thu, 04 December 2008 20:55 Go to previous message
Tatiana Fesenko is currently offline Tatiana Fesenko
Messages: 530
Registered: July 2009
Senior Member
Hi Nicolas,
Sorry for the delay with the answer, I was on vacation.

> I updated #256356 asking to show the last k members of the fully
> qualified name of any NamedElement on a diagram.
Great!

>>> - a line between the element that owns the constraint and the
>>> shortcut to the constraint
> It looks to me that the behavior you describe depends on whether
> Constraint#ConstrainedElement is an object that is shown as a Shape
> (e.g., a Class) or as a simple label (e.g., an operation).
> If Constraint#ConstrainedElement is a class, then the link is
> displayed but only after I close the diagram and then re-open it.
I think you set a reference in the properties view, right? If so, it's known
GMF problem - https://bugs.eclipse.org/bugs/show_bug.cgi?id=254871 Refresh
with F5 should help.
Alternatively you can use 'Constrained Element' palette entry, it doesn't
require refresh.
>
> If Constraint#ConstrainedElement is an operation, then there is no
> link shown at all, even if I close/re-open/refresh the diagram.
>
> Is this correct?
You are correct, contraints should be connectable to operations. There is
a similar request in bugzilla - https://bugs.eclipse.org/bugs/show_bug.cgi?id=235158
..

>> ... request #199731
>> 'Structure diagrams synchronization' was created and fixed.
> Wow, it's nifty, pretty and handy!
>
> Thank you *very* much for implementing this feature!
Yes, it is. Thanks to Michael Golubev who implemented it. I personally also
like this feature a lot. The good news, that many improvements have been
done after M3, so in M4 it will be even better.

> I hope that you can reproduce the problem with the following detailed
> instructions:
> Scenario 1:
Hm, the generalization was shown immediately in my workspace. Could you please
submit a request into bugzilla?

> Scenario 2:
> ...
> 4) Arrange the elements similar to figure 7.3 of UML 2.2
> That figure shows Element and Comment twice.
Yes, that would be great if the user could manage elements and links between
such elements this way.

> 5) We can create shortcuts!
> Create a shortcut for Superstructure::Classes::Kernel::Element Create
> a shortcut for Superstructure::Classes::Kernel::Comment
Success!!!
>
> 6) Ok, but now what?
> To replicate fig 7.3 in UML2Tools' class diagram, we have to visualize
> the associations A_annotatedElement_comment and
> A_ownedComment_owningElement with different pairs of shape nodes for
> Element and Comment.
>
> Since these associations already exist, we are not creating new
> associations; instead, we need to do somehting like this:
>
> - delete one of the two associations from the diagram
> - select the two shape nodes between which we want to display a
> specific
> association (similar to synchronize but for a pair of nodes instead of
> the whole diagram)
Alternatively you can reroute source or target of the element.

Tanya.
Re: [ClassD] Is it possible to show... [message #623361 is a reply to message #475964] Thu, 04 December 2008 22:24 Go to previous message
Tatiana Fesenko is currently offline Tatiana Fesenko
Messages: 530
Registered: July 2009
Senior Member
Nicolas,

>> To replicate fig 7.3 in UML2Tools' class diagram, we have to
>> visualize
>> the associations A_annotatedElement_comment and
>> A_ownedComment_owningElement with different pairs of shape nodes for
>> Element and Comment.
>> Since these associations already exist, we are not creating new
>> associations; instead, we need to do somehting like this:
>>
>> - delete one of the two associations from the diagram
>> - select the two shape nodes between which we want to display a
>> specific
>> association (similar to synchronize but for a pair of nodes instead
>> of
>> the whole diagram)

I created a sample diagram for the given figure and put it to CVS and wiki
: http://wiki.eclipse.org/MDT-UML2Tools_FAQ#Can_I_display_a_si ngle_element_several_times.2C_like_it_is_shown_on_figure_7.3 _of_UML_2.2_specification.3F

Tanya.
Previous Topic:UML 2.1 Diagram - Association Class?
Next Topic:Show entry/exit actions in state machine diagrams
Goto Forum:
  


Current Time: Fri Oct 31 10:53:24 GMT 2014

Powered by FUDForum. Page generated in 0.01983 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software