Version, Subset and Profile IDs - suggestions for NDRs.

[Tim McGrath <tmcgrath@portcomm.com.au>]

The current draft of the NDRs discusses the use of three new elements 
for Version, Subset and Profile identification  The PSC has been trying 
to  understand how we should recognize these (if at all) in the models.  
In doing so we think there may be more clarification and refininements 
needed to the current draft rules.

ELD 13 states:

> The 'UBLProfileID' element MUST be declared immediately following the 
> 'UBLExtension' element and MUST be based on xsd:string, with 
> xsd:maxOccurs="1".

ELD 14 states:

> The 'UBLSubset' element MUST be declared immediately following the 
> 'UBLProfileID' element and MUST be based on xsd:string, with 
> xsd:maxOccurs="1".

Both these come under the optionality explained on lines 651-652 of the 
NDR document ...

> The following elements are not necessarily required. However, when 
> they are used the following rules (ELD12, ELD13, ELD14)apply in the 
> order below.

Firstly, we would propose that these elements are of type IdentifierType 
like all other identifiers in UBL. And to be compliant with CCTS naming 
rules, these elements names must end with "ID".  As they may not be 
defined by UBL, more likely by implementors, we would also suggest 
leaving off the "UBL" prefix.  The names then become "ProfileID" and 
"SubsetID".

Secondly, we are not convinced that Profile and Subset identifications 
are "not necessarily required".  They should be in all schemas to 
support compatibility with those implementations chosing to use them.  
Their use and values can be optional but the elements should be declared 
in all document schemas.  This also avoids the problem of having to 
state where they must appear.

This leads to the possibility of separating Profile and Subset from the 
discussion about Extension Scheme (section 3.4) and allows that to talk 
only about xsd:any.  ProfileID and SubsetID are not extensions.

It may be simpler to consider ProfileID and SubsetID in the same way as 
the Version identifier.  That is, metadata about the document 
implementation.

Rule VER15 currently states:

> Every UBL document schema MUST include a required element named 
> "UBLVersion" as the first child of its root element. This element MUST 
> have a default value that matches the value of the xsd:version 
> attribute of its containing schema.

Again, we recommend this be of type IdentiferType and hence its name be 
UBLVersionID.  In this case the value is only defined by UBL and so the 
prefix is legitimate.

Our suggestion is that we will build UBLVersionID, ProfileID and 
SubsetID into each document model (spreadsheet). This means that we will 
always have three elements in every document schema - UBLVersionID, 
ProfileID and SubsetID. 

So the NDR rules really only needs to say something like...
"Every UBL document schema MUST include elements named "UBLVersionID", 
"ProfileID and "SubsetID" as the first three children of its root element."
and
"The element named "UBLVersionID" MUST have a default value that matches 
the value of the xsd:version attribute of its containing schema."

Does this seem reasonable?  Does it still satisfy the original 
requirements for these elements?

Perhaps this can be added to the agenda for the Atlantic call this week.

-- 
regards
tim mcgrath
phone: +618 93352228  
postal: po box 1289   fremantle    western australia 6160
web: http://www.portcomm.com.au/tmcgrath