How to show directions on class diagram?

[Miguel Gonzalez]

Hello,

On a named association, how do I show the direction it has to be read?

Like on figure 6 where it says 'fulfilled by'...
http://www.agilemodeling.com/style/classDiagram.htm#RelationshipGuidelines

Thank you.

[Paolo F Cantoni]

Hi Miguel,

It IS tricky...

You need to select the label (with the Association Name) and then you can right-click to get the context menu.  Look for Direction>.

HTH,
Paolo

[sargasso]

Notwithstanding the above advice, let me just say "this".

"this"

re, your exemplary diagram, just HTF@! can an instance of an : : Order exist, if it hasn't already been delivered?

OMD how I hate half baked cr*p like that.


Enough, I'm going to get back in my casket.  Let s-o-s take care of it.

listen carefuly, my name is bruce




p.s. Scott's in the next crypt

[Paolo F Cantoni]

hi bruce (s-o-s)

Not withstanding you excellent point abut examples also being exemplars...

I'm not sure I understand your point.  Isn't the order the specification?  Therefore the order exists whether or not it's been delivered?

However, I note that in colloquial language we are often less rigorous with language than perhaps we might...

Which diagram are you complaining about?

Scott Ambler's usually pretty good at this stuff...  

Paolo

[sargasso]

Paolo,

In Mr Ambler's figure 6, there exists a class, or classifier (I don't really care which) "Order" with an association to another class or classifier) "Delivery".  On this association there are four adornments, a <<name>> (meaningless, as the OP said, unless it has a "reading direction" ) and two cardinalities.  Of these, I beg to ask in the southernmost instance, doesn't this mean that:

If I instantiate a :: Delivery  , then it (by virtue of it's cardinality) it MUST belong to a :: Order. Fair enough, but if I substantiate a  :: Order , then diagramatically, I can only do so if there are >= 0 Delivery's that are already extant, somehow (_undefined).  
 
Paolo, I'm not really sure where I'm going here, but.  There are some fundamental things, both in UML and in EA that need some serious and I mean serious (like M E N T H O I D S ) addressing.

as bl :-Xody always
bruce

[Miguel Gonzalez]

Thank's Paolo, that helped.

Miguel.

[Paolo F Cantoni]

Paolo,

In Mr Ambler's figure 6, there exists a class ...
bruce

Oops...  (I plead I was "rushing out the door"....)

You are, of course, correct bruce.

Maybe Scott was enacting that non-sense from a long time ago in data modelling the: "optional tending to mandatory"?  I think you're of an age to recall that...
 
You are perfectly correct... These kinds of things need fixing...

There are better ways of handling that.

"Rigor is your friend"

Paolo