Re: Using the core location vocabulary to query national address data

On 2014-06-13 19:28, Stasinos Konstantopoulos wrote:
> Frans, all,
>
> On 13 June 2014 16:54, Frans Knibbe | Geodan <frans.knibbe@geodan.nl> wrote:
>> This got me thinking:
>>
>> 1) If one of the goals of Location Core is to harmonize worldwide addresses,
>> perhaps it is a good idea to consider having something like an address
>> template. The various bits that can define an address are already there, but
>> in order to build an address that could be used in mail traffic for
>> instance, one needs to know about the order in which address components
>> should be placed. For the Netherlands that would be something like
>>
>> <thoroughfare> <locator designator (first level)>  {locator designator
>> (second level)} {locator designator (third level)}
>> <postCode> <postName>
>> <adminUnitL1>
>>
>> Could it be that something like this exists somewhere? I don't believe such
>> a thing is part of the INSPIRE specifications.
> Natural Language Generation (NLG) systems have such mechanisms
> already, where templates specify language-specific syntactic and
> grammar details and have slots that are filled in by
> language-independent data. Some of these systems, such as NaturalOWL,
> are also specifically geared towards generating text from Semantic Web
> data.
>
> In NaturalOWL, these templates are represented as annotation
> properties over the RDF data. Different templates can be defined for
> each language, but this can extended to different (country x language)
> pairs to, for example, capture:
>   - the French rendering of an address in Belgium
>   - the Dutch rendering of an address in Belgium
>   - the Dutch rendering of an address in the Netherlands
> etc.
>
> If you could send me sample RDF and the wished-for textual result, I
> can try to check if the current mechanism and representation is
> adequate with any modification or extension.
>
> s

Hi Stasinos,

It is good to know that such  template mechanisms already exists. I do 
not have any strong requirements for the assembly of full addresses from 
address components, I just thought a reference to a template would be 
useful for consumers of address data that are based on LOCN, so it got 
me thinking. And I wondered if it makes sense to think about some kind 
of extension to the vocabulary to allow specification of how an address 
should be built from its parts.

For the dataset that I have published I have decided that it would be 
best to add extra triples with locn:fullAddress properties. After all, I 
think the rendering of an address primarily depends on national 
conventions. For example, should a French address be used in other 
countries it makes sense to keep the French format.

The nicest things about just using locn:FullAdress is that it is already 
there and that it is simple.

Regards,
Frans

------------------------------------------------------------------------
Frans Knibbe
Geodan
President Kennedylaan 1
1079 MB Amsterdam (NL)

T +31 (0)20 - 5711 347
E frans.knibbe@geodan.nl
www.geodan.nl <http://www.geodan.nl> | disclaimer 
<http://www.geodan.nl/disclaimer>
------------------------------------------------------------------------