RE: [wsdm] [Omod] prettied up UML sample

["Sedukhin, Igor S" <Igor.Sedukhin@ca.com>]

Actually, in general I agree with you, Fred. I think eventness is more important conceptually than being ro/rw. I tried to work out the constraints, but they are such an ugly thing in UML that it is really hard to get it setup right. There are so many variations and possibilities there. Also nomatter what I did I could not get property constraints to showup in a Visio diagram.

It is a legitimate approach to model this as stereotypes (although probably not as nice as I would like it to be). Stereotypes are just a semantical tags of a metamodel of one's choice. So in other words one could associate any kind of semantics with stereotypes of certain elements and treat them as anything.

It is limiting in UML since there is only one stereotype per one instance of an element of the model (e.g. a property in a class). There is a way to have a meta-model diagram that explains stereotypes and aggregates them as necessary. That can provide a solution for modeling metadata. For example we may have a stereotype for a <<metric>> or <<mutable>> and  such. The problem is that we'd need <<roMutable>>, <<rwMutable>>, etc. defined as well. That is a pain.

-- Igor Sedukhin .. (igor.sedukhin@ca.com)
-- (631) 342-4325 .. 1 CA Plaza, Islandia, NY 11788

-----Original Message-----
From: Fred Carter [mailto:fred.carter@amberpoint.com] 
Sent: Tuesday, October 28, 2003 5:47 PM
To: wsdm@lists.oasis-open.org
Subject: Re: [wsdm] [Omod] prettied up UML sample

Thus quoth Sedukhin, Igor S (~ 28-Oct-03 1:50 PM ~)...

> In response to
> ACTION: Igor to research how to pretty-up the presented UML diagram to 
> be UML 2.0 compliant with respect to readonly modifier and research 
> how to hide +/- if they are not used in the diagram.

Not being one to let my ignorance get in the way ;-), since we're going so far as to use "rw", "ro",  etc., might it not be better to use {} instead of << and >> ?  I suppose they could be stereotypes uber everything, but it seemed like a constraint on the function would be equally appropriate.  Not sure I have a strong opinion here, especially since I'm not a UML expert, but it's worth considering.

Also, if it's not a stereotype, should it move to the end?

Thus,
     <<rw>> SampleOptionalProperty[0..1]: xsd:dateTime becomes
     SampleOptionalProperty[0..1]: xsd:dateTime {rw}

 From an "information priority" point of view, I tend to think that the "event"-ness is much more important than the mutability of a property. 
That is, first I'd sort things [mentally] into events vs. properties, then, within properties, by type and/or optionality, and finally by mutability... until I know what it is, I don't much care if I can change it...

(If you send me the .vsd file, I'd be happy to tweak it to "get the full effect".  Once I see it, I might even disagree with myself :-/ )

Just one opinion...


> 
> <<MOWS UML Model Sample1.png>>
> 
> --* **Igor Sedukhin* .. (igor.sedukhin@ca.com)
> --* (631) 342-4325* .. 1 CA Plaza, Islandia, NY 11788
> 
> 
> ----------------------------------------------------------------------
> --
> 
> 
> ----------------------------------------------------------------------
> --
> 
> To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsdm/members/leave_workgroup.php.


--
Fred Carter / AmberPoint, Inc.

mailto:fred.carter@amberpoint.com
tel:+1.510.433.6525 fax:+1.510.663.6301


To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsdm/members/leave_workgroup.php.