Igor, I would like MOWS to extend MUWS and not change
it. Therefore, you wouldn't change the inheritance hierarchy, but add to
it. Your words are right, but I interpret them differently in UML.
When you say the resource is a WS endpoint, you are really saying that the
endpoint takes on the properties / associations / characteristics of a
resource - that makes endpoint the subclass. This leaves resource as a
general abstraction, but in the case of MOWS says that the endpoint IS-A kind
of a resource. I really can't see it as the other way
round.
Andrea
Thank you, Andrea. I will be incorporating your changes into the
document since Word has turned a bad joke on us again :) -- the template
formatting went astray...
With regards to the figure 5. In MOWS we need to express that a resource
here IS-A Web service endpoint and that is the one we're managing. How
else would you indicate that in UML? I guess we could remove the resource from
the diagram altogether, but still manageable resource here IS-A Web service
endpoint not the vice versa. In MUWS a manageable resource IS-A arbitrary
resource not the reverse.
-----Original Message----- From: Andrea
Westerinen [mailto:andreaw@cisco.com] Sent: Tue 11/30/2004 10:23
PM To: wsdm@lists.oasis-open.org; Sedukhin, Igor S Cc:
Subject: Here is the MUWS document with my editorial
markup
Also, I had a
comment on Figure 5 that I highlighted in the attached document. My
problem with the figure is that it restricts an arbitrary resource to be a
Web service endpoint. This is not correct. It seems that the
inheritance line should run from manageable resource (the superclass) to
the Web service endpoint (the subclass). This would indicate that
the endpoint IS-A manageable resource in MOWS (which is what we want, I hope
:-).
Andrea
|