UBL Context Methodology Subcommittee

Context Methodology Session Agenda

• Review "eight questions"
• What do context rules apply to?
• Gather and review use cases

Review of "Eight Questions"

• No controversy on most questions
• Additional discussion on several questions
  • Use of XSD derivation
  • Use of subtractive refinement
  • Hierarchical value spaces for context drivers
  • Assembly rules

Use of XSD Derivation

• Tendency to agree that use of XSD derivation will have advantages for implementers
• Some issues since certain aspects (restriction of code lists) are not supported
• Further issue relating to the removal of required fields (see later)
• Stuck to decision to remain agnostic during the design phase and tackle this issue later

Use of Subtractive Refinement

• General agreement that subtractive refinement will be needed for effective modeling
• One controversial point raised: removing of required fields
  • Breaks interoperability
  • BUT: Might be "demanded" by data modelers
  • OR: "Moral hazard" of making everything optional
• Possible solutions:
  • Have intermediate types with everything optional
  • Have a "gray shade" of required
  • Require default values
  • Rebuild incompatible types from scratch

Hierarchical Value Spaces for Context Drivers

• Some values may match exactly only, others will match all subcategories
• Decided that this is probably a specific case of a more general matching issue
• One suggestion was to use a mechanism similar to the axes of XSLT
• Postponed as it will be one of many issues tackled as part of the iterative context rule language design process

Assembly Rules

• Should we build schemas directly or using assembly rules
• Assembly rules would have explanatory and possibly design-methodological value
• Could possibly be unified with context rules by having an "any" context
• Will be addressed eventually but is not currently a priority

What do Context Rules Apply To?

• Could apply to abstract analysis artifacts or concrete schemas or schema fragments
• Decided to apply them directly to schemas as implied by UBL goals
• Arofan to write position paper

Gather and Review Use Cases

• Great list of initial use cases compiled in joint session with Library Content SC
• List reviewed and approved by Context Methodology SC
• Will be made available for further review and extension
• Authoring of Context Methodology spec can now begin!

Resolutions

Resolution: The derivation relationship of new types created using the context methodology to the original types shall be left open at present. This may be harmonized with XSD derivation mechanisms in the future, and aspects of the methodology that would prevent this or make it more difficult should be avoided if possible.
RESOLVED

Resolution: Subtractive refinement shall be one of the mechanisms for creating interoperable types from core types.
RESOLVED

Resolution: The context rule language shall be defined ad hoc according to requirements established by the Context Methodology Subcommittee, in collaboration with other UBL subcommittees, rather than using a standard transformation language such as XSLT.
RESOLVED

Resolution: Assembly rules will be an important part of the overall methodology for building UBL schemas but will not be addressed at present as there are other issues with higher priority.
RESOLVED

Resolution: To avoid needless dependencies, we shall not attempt to harmonize the names of the tags used in the planned context rule language with related names in XSD.
RESOLVED