search

DATEX-II-EU / DatexII

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

IMPORTED (234) - Issues with PositionConfidenceElipse (Bugzilla Bug 19) #19

Open datexii opened 5 years ago

datexii commented 5 years ago

This issue was created automatically with bugzilla2github.py

Bugzilla Bug 19

Date: 2019-02-28T16:29:57+01:00 From: @ingeniumfaber To: Jean-Philippe Mechine <Jean-Philippe.Mechin@cerema.fr>

Last updated: 2019-08-23T14:30:06+02:00

datexii commented 24 years ago

Comment 56

Date: 2000-01-01 01:01:01 +0100 From: @ingeniumfaber

The v3.0 Level A model contains a new class PositionConfidenceEllipse. The definition property of this class states that it describes an ellipsoidal shape that describes the positional accuracy of the point corresponding to "a predefined confidence level (e.g. 95 %)". This confidence value is not stated anywhere in the model! Knowing an ellipsoidal confidence area without knowing the corresponding confidence level seems useless. Proposal: 1) Add an attribute to the class to specify the corresponding confidence level in %. The attributes semiMajorAxisLength and semiMinorAxisLength are typed MetresAsFloat. This type allows for extremely large values (2^24-1 * 2^104). Nevertheless, the corresponding xxxCodedError have a value "outOfRange". Actually, it is the only value besides the value "unavailable". "unavailable" seems useless for an optional attribute. What would be the semantic difference between the attribute not being present in a data instance or the value being present with any possible value and the corresponding error being set too "unavailable"? "outOfRange" is confusing, since it states that the value is exceeding 40.93m. What would a data set with value 50 and "outOfRange" mean? It doesn't seem wise to allow large values and hide that most of them are invalid in the definition of a corresponding errors flag's value! If the limit of 40.93m is semantically intended, a new datatype should be specialised from MetresAsFloat that has corresponding facets limiting the value space. Proposals: 2) Remove literal "unavailable" from enumeration "PositionConfidenceCodedErrorEnum". 3) Change datatype of attributes semiMajorAxisLength and semiMinorAxisLength to a new typed specialised from MeteresAsFloat with facets limiting the permissible value space to the range stated in the definition of the corresponding xxxCodedError attributes.

datexii commented 5 years ago

Comment 57

Date: 2019-02-28 16:29:57 +0100 From: bugzilla admin <webmaster@datex2.eu>

--- Bug imported by webmaster@datex2.eu 2019-02-28 16:29 CET ---

This bug was previously known as bug 234 at https://bugzilla.datex2.eu/show_bug.cgi?id=234 This bug blocked bug(s) 2.

Unknown platform other. Setting to default platform "All".

LBlaive commented 2 weeks ago

The issue is relevant. It actually contains several items. The source of these definitions and values is the ETSI TS 108 894-2:2014 (version 1.2.1). Since this publication two new editions have been published, the last one referenced as 2.1.1 of 2022. This latter has introduced number of changes regarding some of the points detailed in this issue. It is to note some data elements and data frames have been replaced by new ones.

Especially the problematic statement aforementioned about the confidence level is now defined by a fixed value of 95% (“…the estimated accuracy with a confidence level of 95 %”).

If this method to define a position confidence is kept it is wise to update this model part by adopting these items and definitions.

Regarding the issue of error codes (i.e. “out of range” and “unavailable”) it is also based on the ETSI definitions. The proposal to limit the acceptable values needs to be assessed regarding the objectives (i.e. is there a need to stick to the ETSI definitions or is it preferable to accept higher values for an optional attribute?).