Re: [regrep] [Urgent Issue] Should reference attributes in rim classesbe elements of type ObjectRef

[Farrukh Najmi <farrukh@wellfleetsoftware.com>]

carl mattocks wrote:
> Agreed - the proposed has a much cleaner O to O approach... 
>
>     *  offers more flexibility for having desirable Object properties.
>     *  offers greater clarity, less cross-reference, in query formulation
>

Actually it has more cross reference in query formulation as there will 
be several extra joins (1 more each reference attribute in the query).

My concern is that while the approach is cleaner it has potential for 
higher storage and query overhead. That is why I am not sure if we want 
to do this change.

> On Tue, Oct 7, 2008 at 12:00 PM, Farrukh Najmi 
> <farrukh@wellfleetsoftware.com <mailto:farrukh@wellfleetsoftware.com>> 
> wrote:
>
>     carl mattocks wrote:
>
>         Farrukh:
>          I understand the difference in declaration but do not fully
>         understand all the implications .. thus not sure of the value
>         in making the change.
>          For example, is the 'side note' the most significant outcome ?
>         /A side note on proposed change is that we would likely move
>         the home attribute from IdentifiableType to ObjectRefType and
>         also create a sub-class of ObjectRefType called
>         DYnamicObjectRefType which adds a QueryInvocation as selector
>         to select reference target for dynamic references./ 
>
>     No the side note is not the most significant. Its just some
>     related information to provide some additional context.
>     The most significant part is whether the cleaner (more OO)
>     approach of the proposed change warrants the potential loss of
>     efficiency (or increased implementation complexity to avoid loss
>     of efficiency).
>     -- 
>     Chair OASIS Business Centric Methodology TC
>     co-Chair (ISO/TS 15000) ebXMLRegistry Semantic Content SC
>     Ontolog ONION Cop Leader
>     VP Berkeley Town Underwater Search & Rescue Unit
>     CEO CHECKMi
>     vmail (usa) 908 322 8715
>     CarlMattocks@checkmi.com <mailto:CarlMattocks@checkmi.com>
>     www.CHECKMi.com <http://www.CHECKMi.com>
>     CHECKMi:Mate Semantics at your Service
>


-- 
Regards,
Farrukh Najmi

Web: http://www.wellfleetsoftware.com