Protect of the original XML when importing from a spider

[benoitg]

When in a network of repositories, it is important that metadata in never altered while traveling trough the network. Only the authoritative repository (the one that has the master copy of a learning resource's metadata) should be allowed to propagate changes.

But for various reasons, it is possible that intermediate repositories store the metadata in a different intermediate representation, or even need to modify it. An example in Eurêka is that it tries to store only one vCard per person, and group them by organisation.

For these reasons a copy of the original data in it's original format must be stored. Currently this is done by making the LOM XML cache read-only for read-only data. This is problematic for two reason:

• There are other reasons to make a resource read-only than it being imported from elsewhere.
• The system is dependent on the original resource being in LOM-XML format, which may not always be the case.
• Keeping important data in the cache makes it vulnerable to programming or human errors.

Separate fields (or a separate table) should be used to store the original data (one field for data, another for format).