DataSource Registry: Use a new "unique_id" for identification instead of Name

[Siegfried Gevatter]

Add a "unique_id" parameter to the data-source registry, which takes an arbitrary string used to identify the data-source. This will make it possible for the Name to be mutable (description and event_templates already are), giving us the following advantages:
 - Data-sources can internationalize their name and send us the translated version (there won't be a problem if the user changes his system's language, the translation changes, or the string was untranslated and gets translated in a later version).
 - Data-sources can change their name if they want to, without leaving a stale data-source information under their old name.