UBL Naming and Design Rules SC

 View Only

[ubl-ndrsc] Arofan's RT/CCT Draft Definitions

  • 1.  [ubl-ndrsc] Arofan's RT/CCT Draft Definitions

    Posted 04-10-2002 11:48
    I agree with just about everything Arofan says here right down to the very end where he makes the point about the terms being backward (that RT and CCT should properly be reversed).  It was informative to finally hear such a lucid explanation of the history behind these two constructs.  If we could get this sorted out to the satisfaction of UBL and CC folks this would be a huge step forward. The only thing I'm left wondering about is this: in light of Arofan's proposal do we really need what is now called RT's at all? What's the alternative (to dropping RT's) -- do we leave RT's in the UBL model and do they get reified as XSD (complex) types?  And then do we go around extending those types with the CCT (complex) XSD types?  If that's the case then how will I reuse one of the CCT's between two RT's -- for instance the Text CCT is gonna be a really widely used structure -- across many RT's.  There isn't any multiple inheritance in XSD -- that's what you'd need in order to implement the many-to-many relationship between base and derived types suggested in Arofan's proposal. An alternative is for the UBL _model_ to treat RT as an adornment or stereotype for documentation purposes but to elide it when it comes to the XSD model -- or to stick it in as a post-validation default attribute the way we're doing with uuid's.  I don't want to get deep into exactly how we do that.  I think we can work that out in a separate thread. I propose we add a little clarification to the first leg of Arofan's proposal: (1) Have a set of Representation terms in the UBL working model which function as semantic primitives, as originally intended by the CC group in ebXML. (The list may need to be altered slightly to include some missing types, but will not undergo wholesale expansion). This indicates what the business purpose of the data is, in an abstract sense, wholly separate from how it will be represented when syntax bound.  When the UBL working model is syntax bound into XSD the RT information will be represented in a manner similar to uuid's.  There will be no XSD types for RT's.  There will be XSD types for CCT's. -Bill