Why not have the "Places" feature set up in the same way that "People" are currently set up in the software?
Since genealogy programs already have the capacity to record all of the core information for individuals, creating a parallel approach to places should be merely a mirroring and adaptation to fit locations. The program already creates the relationship between events and locations, it just needs to create the parallel system that allows relationships between locations.
A "place" evolves over time and many sites already show the "parent" county or location when referencing historical locations. So a parallel family tree for places seems like a appropriate way to deal with locations.
Place type (town, county, state, country, eclesiatical district, etc)
Place key dates ("begin" and "end" similar to "birth" and "death")
"Current" or "Historic" (similar to "living" or "deceased"
Place events (similar to other events for individuals)
Place relationships (between town, county, country, historic and current).
GPS locations would work similarly to the DNA for people to create the link between historic and current locations.
"Historic" places could be mapped because they are linked to the "current" location while staying true to the source documentation and period being recorded. Just like the "also known as" option with people, the user could choose the "historic" or "current" location as "preferred".
Places and their history, media, and relationships to other places play as valuable a role in genealogy research as the research we do on individuals. Genealogy is more than just names of people and the dates of birth and death. It is the stories of the places our ancestors lived and the events that touched their lives that make genealogy so compelling.
In Ancestry.com and Ancestry Wiki:
Currently, "location" information is scattered and difficult to locate requiring hours of additional research time and no "shared" repository for genealogists. The Ancestry Wiki pages are attempting to do this, but participation is still in its infancy and the lack of templates for the "non Wiki programmer" restricts participation. Embedding a database template similar to the family tree structure would increase participation and provide a repository for all genealogists. Just imagine the potential if it was linked to the Ancestry.com collections with "location" research links to source documents similar to finding source documents for people.