search

DATEX-II-EU / DatexII

Main repository for issues and bugs for the DATEXII standard
0 stars 0 forks source link

Definition of Percentage (Bugzilla Bug 463) #463

Closed datexii closed 7 months ago

datexii commented 12 months ago

This issue was created automatically with bugzilla2github.py

Bugzilla Bug 463

Date: 2023-10-05T15:01:25+02:00 From: @JoergFst To: @JoergFst CC: @iancornwellmottmac

Last updated: 2024-02-16T19:04:33+01:00

datexii commented 12 months ago

Comment 1865

Date: 2023-10-05 15:01:25 +0200 From: @JoergFst

The definition of "Percentage" data type is poor: "A measure of percentage.". It has got no facets, the data type is Float.

It does not become clear, which is the proper value range: 0..1 (in the sense of a ratio) or 0..100 (in the sense of percentage).

It is noted, that for example in Czech Republic the 0..1 interpretation is used widely, whereas the 0..100 interpretation is common in other countries, too.

A decision should be made for one of the variants, including a proper definition and facets.

Further note, that in CEN/TS 16157-2 part 2 (i.e. LocationReferencing) the following contradictionary expressions can be found:

7.4.1.2 Semantics "A percentage as distance expression i.e. the ratio between the absolute distance [...]"

and

C.3.3.2 "Percentage [...] A value of 0 (zero) represents a location at the start of the linear element. A value of 100 represents a location at the end of the linear element."

datexii commented 12 months ago

Comment 1867

Date: 2023-10-06 14:10:28 +0200 From: @iancornwellmottmac

An email thread discussed the issue described here. The consensus so far is that "Percentage" should entail an amount with respect to 100, not a ratio where 1 is parity, but also that given reported misinterpretations we should improve the definition to avoid doubt.

Ian suggested a few possible definitions, from concise to verbose, then there was a request that we do include some kind of illustration, so this verbose one seems to meet that requirement:

"A ratio expressed as an amount per hundred, such that for example the value 100 represents the case where a quantity is equal to the reference value it is compared with, while the value 50 represents a ratio of one half."

The consensus so far was that no facets should be specified, since there are valid reasons that in some contexts the number could be <0% or >100%.

(The base type is currently Float, and although this was noted to be a debatable topic, there was no strong argument for change.)

datexii commented 7 months ago

Comment 1916

Date: 2024-02-16 19:04:33 +0100 From: @JoergFst

Fixed for version 3.5. with new definition proposed above.