Responding to vincent.hemery...
Mine is the v2.0 '03 spec.
Hmmm. I think OMG has shot themselves in the foot again with that "at
least one..." It seems to me that to have a decision one needs at least
two choices. And if it is an AD, what else would it be deciding than
which flow to take?

Having a look at the March 2003 v1.5 specification makes me better
understand your point of view.
Mine dates from February 2009 (v2.2) and I had a small look one these
points at the September 2009 2.3 XML version:
Merge node and decision node are now two distinct objects described in
separate chapters.

Responding to vincent.hemery...
My version of the v2 spec dates from '03 so I can't speak to any changes
OMG has made since. But in my spec it is pretty clear that the
difference is definitional. If the diamond has one output, it is a merge
node. If there are two or more outputs it is a decision node and a

VDC is conducting its annual survey of mobile and embedded engineers
so if you are involved in the engineering of mobile or embedded
systems/software, this is your chance to influence key solution
suppliers.
The research covers embedded software, hardware, tools, and
development practices. Your thoughts will improve our insights into

This is what I was refering to. But that comment was coming from a
misunderstood due to the differences in the terminologies we use. But
now I see what you meant.
The point that mostly interests me in your answer is :
Maybe I should precise that I am currently developing a model-based
Activity diagram editor.

Responding to vincent.hemery...
aka Merge Node.
aka Merge Node. The nature of the flows does not change the way flow
paths are followed.
Decision nodes and merge nodes are basically the same thing so the same
UML symbol is used. They different from Fork and Join nodes in that they
do involve a split into multiple flows in parallel (Fork) or an implied

Responding to DesperateDan...
Lloyd made an assumption about your problem that wasn't as obvious to
me, so I would like to push back on whether that was really what you
were doing.
A use case captures a logical context for the actor interacting with the
system. That high-level interaction may be broken up within the use case

Sorry, I was meaning the two situations (though control node is a
generalization of decision node):
- a decision node has two incoming control flow and one outgoing
control flow (and no decision input flow)
- a decision node has two incoming object flow and one outgoing object
flow (and no decision input flow)

Dan,
A Use Case references a category of Activities, and Activities reference a
category of Actions.
You can put the Activities for one Use Case under its own Activity diagram
and each Activity can have its own Activity Diagram full of individual
actions, etc.
Ken

Dear all,
I'd like to know your opinion on the this topic "Sequence Diagram:
What You See Is NOT What You Get". I'm referring to the combined use
of CombinedFragments, Operand and Messages.
It is possible to DRAW Message within Operands in Combined Fragments
but this is only "graphical sugar" for documentation purposes: we are

I have a use case diagram with an actor associating to three use
cases.
I then have a textual description which expalins the actor's steps of
interactions with all three use cases.
I've described all the steps under ONE use case heading.
My question is: Should I have layed out my text under 3 separate use

Responding to vincent.hemery...
Control nodes and decision nodes are different animals. OTOH, the flow
is simply a data flow. Thus a control node consumes (processes) flow
messages or generates them. A decision node simple routes the flow to
the right control node once it has been generated. IOW, control nodes

Responding to S...
This is probably a tool issue so a better place to query would be a
Borland user forum.
There is insufficient information. What is the context of drawing the
line? An Interaction Diagram? Object Diagram? what do you mean by
multiple messsages? The callee responds with multiple returns at

Jap Boon Churn wrote, On 2/26/2010 9:44 PM:
UML 2 provides a connector for this purpose. It can span pages, although
I don't think it can span diagrams. In other words, you can use it on
one diagram and print that diagram on two pages.
-Jim

Hi,
If my activity diagram is long, is there a connector notation that I
can use in activity diagram to connect from one to another page?
Thanks.

Hi,
I am using Borland Together UML tool. In this tool, If I select return
option for a method call, a single dotted lined will be drawn to show
return message.
But, my method has multiple return messages for different conditions.
Is there any notation using which I can represent return message
directly.

Please disseminate the DANCE workshop CFP to your colleagues and PhD
students that could be interested.
============================== ============================== ============
DANCE: International workshop on Distributed Architecture
modeling for Novel Component Embedded systems
In conjunction with NOTERE 2010

According to the UML specification UML Superstructure Specification,
v2.2 (12.3.22, p360):
[1] A decision node has one or two incoming edges and at least one
outgoing edge.
Hence, there is nothing in the specification that prevents the
situations :
- a control node has two incoming control flow and one outgoing

Thank you for you honesty haim, however I would disagree with you on that.In
fact I have found 3 to 4 more sites that will do just that. Generalization
in the dark isn't such a good idea, regardless of how much experience one
has, not to mention that part of the "grinding" process involves seeking out

Responding to Firewolf2210...
Actually it isn't a good place for class projects and home work.
Grinding through that stuff on your own is part of the learning process.
If there is something fundamental you do not understand, forums can
help. But forums aren't here as a shortcut to free up more time for beer