User:John Cummings/Archive/sandboxt

Importing Wiki Loves Monuments lists into Wikidata
Dear all

I’m posting this to start a discussion on migrating the information in the Wiki Loves Monuments (WLM) database to Wikidata as part of the Connected Open Heritage project. There have been discussions about the work previously here and here.

Below we have outlined what we can see as the potential approaches and issues with each approach. We would very much appreciate your feedback including the following questions:
 * Which approach do you think would work and why?
 * Are there ways to mitigate or overcome the issues highlighted?
 * Are there any other approaches or issues that have been missed?
 * If the combination of the two approaches outlined is a good approach, what combination would work best? Would this depend on the country or could universal guidelines be used?

Approaches
Wiki Loves Monuments lists for different different countries are constructed in different ways: In addition some countries have lists with unclear documentations as to the source of the data and methodology used to select objects.
 * Some countries use only official immovable heritage lists
 * Some countries add sites onto official national and regional registries e.g South Africa.
 * Some countries do not have official lists.

There appear to be three approaches to importing the WLM database into Wikidata taking the construction of existing databases into consideration:
 * 1) Import directly from the WLM database into Wikidata.
 * 2) Import from the reference sources used to created the WLM databases where known.
 * 3) A combination of the two approaches.

Issues
There seem to be a number of issues with each approach:

Importing directly from the WLM database into Wikidata

 * By importing straight from the WLM database rather than going back to the sources there might be no referenceable sources connected to the items created, potentially we will be creating up to 1,384,000 unreferenced items. However this could be solved (in the short term) by an "imported from Monument database", wlm-id:, date: " reference. This is not ideal but is essentially how importing from Wikipedia works today.


 * Even if the data is based on official lists there is no simple way of determining which changes have been made to the data since the time it was imported.


 * Whilst facts are not copyrightable there might be elements of the Wikipedia lists (and hence the monuments database) which are copyrightable. If so there is an issue in that the Wikidata license is more restrictive than that of Wikipedia.

Import from the reference sources used to created the WLM databases

 * Some WLM countries add sites onto official national and regional registries e.g South Africa and some countries do not have national registers.


 * Some of the original official documents/lists might be copyrighted. It is currently unclear what copyright restrictions might apply if we are looking at importing the WLM lists/Monument database vs. the original lists that the WLM lists are derived/ copied from?

Again we would very much like your feedback on this
 * This approach would not make use of the work that went into standardising the info from the various sources for the Monuments database.

Thanks