I have to make a remark on this extension that has probably a braoder scope for all location referencing extensions. The observation was actually made when reviewing an extension produced in a German Cooperative Systems field trial, where a new type of linear has been created consisting of a polygone with points given as pairs of ETRS89 coordinates.
The approach taken was the same as in this OpenLR extension, i.e. the top level class of the linear LR was created as a specialisation of "Linear". I think this approach is wrong, since Linear is a container and itself meant to be empty - without attribute - where the class only consists of a choice of plugged-in LRs.
Hence, the better way to plug in a new linear LR would be to create a specialsed class "ExtendedLinear" - also empty - which has the top level LR class - "OpenlrLineLocationReference" in this case - as a single component plugged into it.
If the extension should finally be adopted for the level A standard, this aggregation/composition would be deleted an replaced by one to the "Linear" class.
This is a very messed up thread.. This is a response to Josefs comment and suggestion about the ExtendedLinear. I agree with this. It would add an extra layer but logically I agree with this.
A new version will be uploaded soon.
Just a note.
OpenLR shall soon be able to express point locations (I do not what is actual status, may e it is already there).
Where to fit it into datex2 structures then?
* keep it as special case of linear
* in case of point use another container
* using other approach
Jan Vlčinský
Yes, this extension alreday support OpenLR Point references which I think you are referencing to. For Points it extends the Point DATEX II class.