#215 classification marking checking

[Daniel Black]

Some material may be marked with a security classification. DPR should do checks to document/image header/footer information to prevent material being archived that does not belong in the target repositories.

The sourceforge project spillguard may give some ideas/mechanisms as to how this checking could occur.

[Michael Carden]

Management of security classification metadata is explicitly out of scope for DPR.

[John]

I am not sure about this one. It does seem feasible, but...

Agencies are currently responsible for identifying (and handling appropriately) classified material in their possession. They are supposed to segregate any classified material prior to transfer to us.

In RecordSearch, the only way of telling if something is classified or not is the location - classified records are placed into secure areas, unclassified ones into general storage areas.

When records are requested by the public, things get 'access examined' to ensure that any material which is exempt from access is not released. This is the stage where security classification would alert the examiner that something may warrant exemption (but it certainly is not the only deciding factor).

Maybe this needs to be included into whatever access methodology is developed?

On the other hand, the Xena OCR process would also bring up stuff with a security/privacy marking (although it will pick up records which use the word 'confidential', but that does not necessarily mean that the document is classified). Picking up on the header/footer information would probably reduce the chance of misclassification

[Terry O'Neill]

Closed as rejected as too far out of scope for now and unlikely to ever be implemented in DPR.