Re: Questions regarding the WBEM/CIM XSDs
Kirk Augustin <[email protected]>
| Newsgroups | gmane.network.open-pegasus.general |
|---|---|
| Message-ID | <[email protected]> |
I don't use XSDs that much, but I do not think the XSD model can let an association be known. I believe that not only are associations only defined by MOF files, but that even that is not possible for a human to follow, and that the simpler "profiles", as limited and focused subsets, are necessary in order for understanding in human terms. http://en.wikipedia.org/wiki/CIM_Profile And XSD can describe the hierarchical components of an object, but can not describe relationships with independent objects like associations, that can contain references to several objects. If you tried, then I think you would find the cyclic nature of the MOF description would make it impossible. I believe RDF is to show the relationship of properties to a node, but not a relationship between nodes, as associations do. But perhaps I am not following. Kirk & Trish Augustin 11821 NW McNamee Rd Portland, OR 97231 ________________________________ From: J B <[email protected]> To: Kirk Augustin <[email protected]> Cc: [email protected] Sent: Wed, October 6, 2010 10:59:13 AM Subject: Re: Questions regarding the WBEM/CIM XSDs I can see that argument, and agree with it. However, my original question is still valid, I think. :) Without the MOF to guide me, using only the XSDs as a model, how would that association even be known? At least the MOFs are typed and have the same UMLPackagePath hint. And the definition of a cim:cimReference is wide open: <xs:complexType name="cimReference"> <xs:sequence> <xs:any namespace="##other" maxOccurs="unbounded" processContents="lax"/> </xs:sequence> <xs:anyAttribute namespace="##any" processContents="lax"/> </xs:complexType> So I am not even sure what it would point to in an actual instance document unless it was meant to be the entire XML instance document of an OperatingSystem and a ComputerSystem, which does not make sense to me. I would hope that I could create an instance of a ComputerSystem and an OperatingSystem independent of the relationship type. The ultimate goal I have is to generate an RDF model of this and use it to capture my data, but I will settle on just better understanding and capturing the information in straight XML as defined by the XSDs. Is there any example of actual instance documents using the XSDs? Thanks! Birch On Wed, Oct 6, 2010 at 1:50 PM, Kirk Augustin <[email protected]> wrote: > I am not aware of any way that a XSD hierarchy can describe an association. > An association is a relation, as in a relational database, between 2 other > objects. > A relation contains instances of 2 other objects, but does not have to exist > and is not part of either one. > For an XSD to describe a relation, it would have to not have a hierarchical > description, but maybe I just don't use XSDs that way. > > Kirk & Trish Augustin > 11821 NW McNamee Rd > Portland, OR 97231 > > ________________________________ > From: J B <[email protected]> > To: [email protected] > Sent: Wed, October 6, 2010 10:07:39 AM > Subject: Questions regarding the WBEM/CIM XSDs > > In looking over the the WBEM/CIM XSDs, I am puzzled by the use of > associations. As an example, let's take CIM_ComputerSystem and > CIM_OperatingSystem. According to the MOFs, there is an explicit > association of InstalledOS(0..*) and RunningOS(1..1). I would have > expected some element in one of the above XSDs to explicitly call out > this association, but instead, I see two other XSDs, CIM_InstalledOS > and CIM_RunningOS. I am assuming that the GroupComponent and > PartComponent refer to CIM_ComputerSystem and CIM_OperatingSystem, > respectively since they are of type cim:cimReference. But without the > benefit of having the MOF and UML diagrams available to me, how in the > world would I know this association exists? I would have expected > that the XSDs and the MOFs would ultimately describe the same model, > but it does not appear that way. What exactly would be the value of > the cim:cimReference element? > > So if the above is actually correct, can anyone provide me with a > sample XML instance document that shows how to actually tie a > CIM_ComputerSystem instance document to a CIM_OperatingSystem instance > document together? > > Thanks! > James Birchfield >