[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [Fwd: [regrep-cc-review] Association Core Components - Represent with Associations]
Please see comments below marked with <JMC>. Chiusano Joseph wrote: > > Forwarding on behalf of Diego, who makes some great points below. > > I'm heading downtown now for a federal "E-Forms for E-Gov" meeting, and > will be back online in about 4.5 hours. In the interim I would encourage > anyone (observer or participant) to respond to Diego's points below - in > any event, I'll weigh in later today. > > Joe > > ------------------------------------------------------------------------ > > Subject: RE: [regrep-cc-review] Association Core Components - Represent with Associations > Date: Wed, 16 Jul 2003 15:51:37 +0300 > From: Diego Ballvé <diego.ballve@republica.fi> > To: "Chiusano Joseph" <chiusano_joseph@bah.com> > > Joe, > > This is getting interesting :) > Pleas forward to cc-review > > > <Quote1> > > Ok.. just keep in mind that CCProperties are not stored independently. > > Theiy are part of the referenced ASCC/BCC, meaning its attributes go > > to wherever the ASCC/BCC attributes go. If ASCC is a RIM Association, > > the Association will contain Slots with CCProperties attributes and > > ASCC attributes. > > </Quote1> > > > > That's an excellent point, Diego. Thinking back to the "Core Component > > Properties" entity (CCTS p.76) - it has the following attributes: > > > > - Property Term > > - Cardinality > > > > In terms of the Data Element Name, the only difference between an ACC > > and an ASCC is that the ASCC provides an Object Class Qualifier Term - > > e.g. "Residence" for the Object Class "Address". So I believe the > > Property Term would remain with the ACC, and not be placed with (or > > repeated on) the ASCC. > > Are you sure about the qualifiers for CCs? I was not and this is what > I've digged from the specs: > > [B14] The definition of a Basic Business Information Entity shall use a structure that is > based on the existence of the Object Class Term, the Property Term, and the > Representation Term, and enhanced by business related Qualifier Terms. <JMC> I see a discrepancy with the above requirement and Figure 7-1 on p.75. If we think in terms of the 11179 Data Element Terms and where they are placed (per the "11179 Terms/CCTS Entities document I sent recently - dated 7/7/03), we have: Object Class - resides with the ACC Property Term - resides with the BCC/ASCC Representation Term - resides with the CTT Since a BBIE is "derived" from a BCC, there would be no Object Class - an Object Class would not enter into the picture until aggregation is done. I recommend we consider this an inconsistency in Section 6 of the CCTS spec and defer to Section 7 as our "official" requirement. Does that sound good? </JMC> > [B15] The definition of an Association Business Information Entity shall use a structure > that is based on the existence of the Object Class Term, the Property Term and > the Object Class Term of the Aggregate Business Information Entity on which the > corresponding Association Business Information Entity Property is based, and > enhanced by business related Qualifier Terms. <JMC> Same response as my previous one. </JMC> > > If still we have qualifier for Object Class, then it's not shown on > those UML diagrams we've been studying. <JMC> Figure 7-3 on p.85 shows a Qualifier Term as an attribute of ABIE - which I believe is consistent with our discussions. </JMC> > > In terms of Cardinality, I believe this would be a different case. > > Suppose we had an ACC called "Address. Details", and 2 ASCCs > > (as in the > > CCTS example we have been citing) that ultimate "resolve" to > > "Residence > > Address" and "Official Address" (just using English terms here for > > simplicity). One may want to assemble a schema from Core > > Components that > > specifies a cardinality of (for example) "2" for Residence > > Address, and > > "3" for Official Address (that not make sense, but let's > > pretend it does > > for our purposes). If the Cardinality attribute were with the ACC, one > > would not be able to make that distinction. So it appears > > then that the > > Cardinality attribute should reside with the ASCC, so that this > > distinction can be made. > > > > How does that sound to you? > > 100% agree on cardinality. That is exactly the point!! > Now, extend it to Property Term: > If you have a Person. Details with 2 properties of type address (ASCC > to Address. Details) <JMC> Address is not a Property Term - it is an Object Class. </JMC> then you identify them using different Property > Terms (as you suggested, "Residence Address" and "Official Address"). <JMC> "Residence" and "Official" would be Object Class Qualifiers that qualify the Object Class "Address". </JMC> > If that would be stored in the ACC, then the Property Term would be > the same for both properties, what would not be correct. <JMC> The Object Class "Address" is the same for both entities (ASCC and ACC), which is correct. </JMC> > Same case for Business Term. You do need to redefine these > attributes since they are not necessarily the same in te ASCC as > for the ACC. <JMC> [S3] defines Business Term as "A synonym term under which the Core Component is commonly known and used in a business.". Not sure of the connection with ASCCs and ACCs, and where the significance is in your explanation above. Please clarify. </JMC> > See attached example Schema which illustrate possible BIEs for the > same example (BIEs, not CC!!!). It follows the UBL way.. I can soon > send a possible SubmitObjectsRequest for it, if required. Then you > can see the ExtrinsicObjects and Associations. <JMC> Thanks - I'll look it over (since it follows UBL, I will immediately recognize what is being done). ;) </JMC> > > <Quote2> > > I can then ask you why do we need BCC (which are similarly combined > > with corresponding CCProperty) as ExtrinsicObjects, if they require > > an Association to a DataType - and this RIM Association could contain > > BCC + BCC Property attributes??? (I'm stretching it to the limit here) > > </Quote2> > > > > Another excellent point. My response would be that a BCC is the basic > > stored block of information with which other "blocks" are combined > > (using term loosely) to form the entities that ultimately > > reside in (for > > our purposes) XML documents - BIEs and BBIEs. ASCCs are not > > fundamental > > building blocks - they build *upon* another fundamental > > building block, > > an ACC. So I view them as an extension of ACCs - they signify the role > > that an ACC plays within an another ACC, and act as the "glue" that > > joins 2 ACCs together. > > Agree with restrictions. IMO ACC are the blocks you can use and abuse. > Both BCC and ASCC are parts of these blocks, simple parts or composite > parts. They only exist inside the ACC, though. As the ASCC is the glue > that joins 2 ACCs together (through a property), the BCC is the glue > that joins ACC and DataType (through a property). And Yes, BCC is a > basic block, but you can still decompose it into ContentComponent and > SupplementaryComponent, plus restrictions. For the XML representation, > XML Elements and Attributes. <JMC> Let's keep in mind that we are disregarding Properties as entities for at least the time being. I personally don't view ACC as the "glue" that joins ACC and Data Type, but Figure 7-1 can certainly be interpreted that way (so I agree with your interpretation). I view ACCs simply as aggregates of BCCs. So: Where do you stand on the issue of using Associations to represent ASCCs? :) </JMC> > Regards, > Diego > > > > Diego Ballvé wrote: > > > > > > Chiusano Joseph wrote: > > > > > > > > <Quote1> > > > > We agreed on not to add Slots to RIM Association for clarity > > > > and I think > > > > it makes sense. > > > > </Quote1> > > > > > > > > Ah - I'm reconsidering here, in this context. :) I > > believe that this > > > > approach was not good for Properties, but is good for ASCCs. > > > > > > Ok.. just keep in mind that CCProperties are not stored > > independently. > > > Theiy are part of the referenced ASCC/BCC, meaning its attributes go > > > to wherever the ASCC/BCC attributes go. If ASCC is a RIM > > Association, > > > the Association will contain Slots with CCProperties attributes and > > > ASCC attributes. > > > > > > I can then ask you why do we need BCC (which are similarly combined > > > with corresponding CCProperty) as ExtrinsicObjects, if they require > > > an Association to a DataType - and this RIM Association > > could contain > > > BCC + BCC Property attributes??? (I'm stretching it to the > > limit here) > > > > > > Probably the reason is clarity. Desired? Needed? That I can't answer > > > alone. > > > > > > Regards, > > > Diego > > > > ------------------------------------------------------------------------ > Name: Page12Example.xsd > Page12Example.xsd Type: BizTalk Schema (text/xml) > Encoding: base64 > Description: Page12Example.xsd > > ------------------------------------------------------------------------ > > Joseph M. Chiusano <chiusano_joseph@bah.com> > Senior Consultant > Booz | Allen | Hamilton > IT Digital Strategies Team > > Joseph M. Chiusano > Senior Consultant <chiusano_joseph@bah.com> > Booz | Allen | Hamilton > IT Digital Strategies Team > 8283 Greensboro Drive Work: (703) 902-6923 > McLean > VA > 22012 > Additional Information: > Last Name Chiusano > First Name Joseph > Version 2.1 > > ------------------------------------------------------------------------ > You may leave a Technical Committee at any time by visiting http://www.oasis-open.org/apps/org/workgroup/regrep-cc-review/members/leave_workgroup.php
begin:vcard n:Chiusano;Joseph tel;work:(703) 902-6923 x-mozilla-html:FALSE url:www.bah.com org:Booz | Allen | Hamilton;IT Digital Strategies Team adr:;;8283 Greensboro Drive;McLean;VA;22012; version:2.1 email;internet:chiusano_joseph@bah.com title:Senior Consultant fn:Joseph M. Chiusano end:vcard
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]