Datatypes R2 Issue 80

Introduction

Taken from ISO Datatypes draft #1:

There is a problem with the CE definition. CE exists to allow a specification to define that the concept must not be post-coordinated, since more infrastructure and work is required to handle post-coordinated concepts. Translation allows for post-coordination on the basis that it may be required to perform a proper translation. So in CE only the root concept (the first time it was encoded) is constrained to not allow for post-coordination. But if the concept that is being represented was first encoded in a different code system than that the one in which it is targetted to be represented (without qualifiers), then it is not possible to represent the original system. Possible solutions to this:

• restrict CE so there is no qualifiers on the translations
• specify that the original code cannot be sent in this case
• redefine the rules for CE so that the translation with the purpose Mandatory is constrained to not include qualifiers

Discussion

This was discussed in committee (joint INM/Vocab meeting 2 May 2007)

example from discussion:

 <translation id="qw2" code="a" codingRationale="orig">
   <qualifier... "z">
   <originalText>Broken Logic</originalText>
 <translation code="c" source="qw" codingRationale="fun"/>

Disposition

Motion in INM 2-May 2007:

Change CD & CE so that the root code is the code that meets conformance requirements - if there is such, or any code if there is not.

Use CodingRationale to indicate which is the original code. Add an implementation note explaining the relevance of use of codingRationale

Dump nested translations. We accept the use case that it is sometimes useful to know which code was translation from which code. We will solve this by adding a reference property to the translation which is an optional reference to the source code from which this code was translated which will be implemented using xml id in the XML ITS.

Status

Approved