I think the first step is to create (or contribute to) an identity validation system on the Cardano blockchain (or on a blockchain which can be referenced by it) that allows for various levels of identity verification and trust but which doesn’t expose any aspect of privacy without explicit, per-instance consent. There would also be a need for users to establish consensus between themselves and possibly to also form trust networks (ie such as between members within a neighborhood or village), and also without exposing the privacy of the identities in that network. I think there are lots of identity validation projects on various block chains, so there are probably plenty of subject matter experts and white papers on this topic.
Second, I think there would be a need to define assets to be owned. For land and the buildings contained on them, perhaps a GIS system combining GPS, photos, and scans of any existing documentation, legal documents, descriptions, landmark references, or agreements could establish definitions of property. There might be some variations between countries or villages as far as what are legally allowable definitions of property which may need conforming to based on location.
Then I think there would also need to be an arbitration process where either through consensus or decisioning offered by government entities that disputes over property could be settled and recorded. There may be a need for this system to also register liens related to financing, but if some alternative could be used that would be preferable. Avoiding repossessions and housing crises would be ideal in developing areas, especially as a way to avoid predatory lending practices.
Another aspect to solve would be how to import various forms of pre-existing ownership registrations. One difficult problem to solve would be when there’s disagreements around the political entity that is in charge of the containing land area, or when people are displaced by war or changes in government.
Lastly for this entity defining component, there would be a need to include for conditions like co-ownership, tenancy, inheritance, etc.
Besides this component there would need to be the financing aspect which would ideally operate as independently of the entity validating process as possible.
The mechanisms involved with lending could be really long conversation and there are many facets to it. There are a lot of projects and thought which have explored this area in depth, though most involve staking currency assets as collateral (which would be easier to deal with than trying to stake physical property as collateral). But the first step I think is establishing the definitions of individuals and property.
One last thought is that ideally the system could be flexible enough to handle lending for other types of purchases as well such as livestock or equipment to facilitate running a business, etc. If the lending was simply done as currency borrowed against staked collateral as most defi lending currently works, then what funds are used for wouldn’t matter. But I think a lot of detailed discussion around the lending aspects starts to get off topic. Like you said, the main hurdle is defining ownership, so treating that as the main focus and keeping that as a separate endeavor from any financing aspects is probably the best approach.