Datatypes R2 Issue 14

Introduction

It's clear that a physical quantity is a tri-value:

• the number
• the units
• the thing being measured

In a PQ, we have only modelled the first 2 parts, on the basis that the third comes from the context. There is a few cases where the third element, what is being measured, doesn't come from the context. In this case, people want to encode this information it the unit, and cannot really do that, it is wrong to try. This always causes confusion with dimensionless quantities, where the unit is 1 (the "unity"), but confusion is not limited to dimensionless quantities, for example, 1 m might be a length of a tape or a wave length, which are vastly different measurement concepts.

We need to resolve whether

EITHER

• the context should provide the third piece of information and needs to be fixed if it doesn't

OR

• we need to extend PQ to handle the third part.

Discussion

It is extremely important that it be clear that Observation.code (or Observation.definition) set the context for any PQ in an Observation.value and that whatever attribute is of type PQ define the measurement concept specifically. Gschadow

The context (such as the definition of the attribute) in which the PQ is used must define what thing that the PQ is a quantity of.

• Add examples
• add warnings against being an idiot.

Disposition

Grahame added clarification text explaining why UCUM is required.

Status

Proposed

Links

Back to Data Types R2 issues