Non-place feature types

[docuracy]

The LP standard limits types to places, but the format has been found usefully to serve datasets in which the features are actually geolocated objects, for example archaeological finds.

Visualisation software (e.g. Peripleo) can recognise and perform faceting based on these types without the need for any complicated configuration.

PROPOSAL: Allow any type properly referenced to a published vocabulary.

[VincentDucatteeuw]

I think this proposal should be discussed in-depth with the working group as it touches on the possible distinction between space (geolocation) and place (relational space embodied with meaning). I'm wondering why it wouldn't be possible to still model these geolocations as places within Linked Places and model the objects within Linked Traces and provide a reference to the place.

But I understand that there are certain objects where the distinction is difficult to make. Would it be possible to post some of these examples so that other people can give their thoughts as well?

[docuracy]

The primary problem we faced with implementing Linked Traces was in providing resolvable references to the many thousands of geo-points in our datasets: it is both impractical and of extremely limited use to create referenceable gazetteer entries for these points, which in most cases have no associated place-names. Archaeological finds are an obvious example, but I've experienced the same problem when trying to geo-annotate historical documents too. This was why I experimented with using linked traces within LPF, and that's what led to some of the suggestions I've made for extending LPF.