General - Replace LocalisedCharacterString by PT_Freetext
|Submitting Organisation:||Discussion (link):||https://themes.jrc.ec.europa.eu/discussion/view/38527/localtype-as-pt-freetext-rather-than-as-localisedcharacterstring|
In IGN France, we have decided for attribute localType to use the harmonised and multilingual EGN classification and to provide this localType both in English and in French.
Some of our source localType match with 2 types in the EGN classification (e.g. our source pumping stations are both « Economic activity facilities » and« Hydrographic point features » )
As localType is defined as LocalisedCharacterString in INSPIRE data specifications, this results into 4 independant values whereas we have in fact 2 values, each of them in 2 languages.
Options and Justification:
PT_Freetext versus LocalisedCharacterString was an issue in TWG GN.
Initially the TWG applied PT_Freetext. Then Clemens Portele advocated LocalisedCharacterString for reasons I do not remember in detail, but probably related with the Generic Conceptual Model and/or International Standardization.
Citations from an Email by Clemens, 02 Oct 2008:
Change the type of LocationInstanceName.name to LocalisedCharacterString. This is based on the fact that all names will have a language associated with them, but there will be only one language represented in any name. Therefore, LocalisedCharacterString is the appropriate type. PT_FreeText does not add any value as there would be also a single text group only.
I will propose a change in the gazetteer model in the Generic Conceptual Model to change the type of geographicIdentifier back to CharacterString to better support gazetteers that do not use PT_FreeText for identifiers/names. In this context it may be worth to point out that both PT_FreeText and LocalisedCharacterString are subtypes of CharacterString. As a result, an instance of PT_FreeText or LocalisedCharacterString may in principle also be the value of a property that is modelled with a value type of CharacterString in the UML model.
ISO 19139 promotes PT_Freetext . PT_freetext being subtype of char string the GCM is still consistent with the ISO
The name experts were open to both solutions. If we now use LocalisedCharacterString it was the consistency within INSPIRE (and eventually with International Standardisation) that lead to this decision.
In case the LocalisedCharacterString is changed back to PT_Freetext, one has to check the INSPIRE model carefully to make sure that the GN model still fits with the overall model.