Search



Page tree

  

Terminology servers at Level 1 will accept any expressions represented in valid classifiable form, as in Level 0. Additionally, some expressions that are not already in classifiable form will also be accepted. These will mainly be syntactically compliant expressions that are not MRCM compliant, but which the server can unambiguously transform to derived expressions that are MRCM compliant, using predefined automated transforms.

Loose Attributes

A loose attribute is any attribute within a CTU expression that is not in an attribute group AND the MRCM specifies:

  • that it should be grouped, or
  • that it is in the wrong domain

Scope of Transformations

Level 1 transformations of CTU expressions may apply to expressions meeting the following criteria.

  • Transformations can only apply where there is no definition status
  • There must be only one focus concept
  • Level 1 transformations can only apply to loose attributes (as defined above)
  • When the expression contains a mixture of loose attributes and attribute groups, the attribute groups are effectively treated as for Level 1, so no transformation of these are attempted

Transformation Process

Expression syntax, concepts being active and MRCM attribute-range validation should be applied before the transformation step. Implementations should fail fast if these are not met and not attempt transformation. This simplifies the preconditions for specific transformations.

The process of applying the transformations is iterative. This includes: collecting the set of remaining loose attributes, then applying the next transformation, and repeating until each transformation has been applied once. When each specific transformation is applied it should be greedy and transform all the loose attributes that meet the conditions of that transformation.

The list of transformations must be applied in a specific order; to ensure that the same classifiable expression is reached every time by all implementations.

Transformations

The allowed patterns can be summarized as follows:

The following pages elaborate on the preconditions and behaviour of each pattern.


Feedback
  • No labels