Devkit schemas #659
Devkit schemas #659
Conversation
a formal representation of Rights Requests so systems and components cna share them @Clementinev please have a look and add the current work you're doing here.
Architecture proposal for v2 of Privateform (#596)
Update request list, treatments list and data categories list
…ope-gdpr init Prohibited scope
Co-authored-by: Marko <[email protected]>
| - **Configuration Maps**, defined at configuration time: | ||
| - *Retention Policies*: For each **Known Selector** (or Data Category implying every `selector` used within that Data Category) a data [Retention Policy](./RFC-PRIV.md#retention-policy), | ||
|
|
||
| - *Provenance*: For every **Known Selector** keep track of one or more [Provenance](./RFC-PRIV.md#provenance) objects. |
There was a problem hiding this comment.
In the RFC document, concept Provenance has a property system which is defined as
the ID of the System having generated the Data Capture Fragment ... or ID of the System having initiated a transfer
I concluded Provenance is related to a Data Capture Fragment, not a selector so
For every **Known Selector** keep track of one or more Provenance
is confusing.
What is provenance in the context of selectors?
There was a problem hiding this comment.
Provenance has System
System generates Data Capture Fragments
Data Capture Fragments have selectors
So you may be able in a situation to answer for CONTACT.ADDRESS.SHIPPING that it is the kind of data thay you obtain from other Systems (System A, System B,,...), while BEHAVIOR.CONNECTION is something you (System ID of your System) generate DERIVED.
There was a problem hiding this comment.
So: if
CONTACT.ADDRESS.SHIPPINGselector has a retention policyRT_aand scopeS_a, and if we, for a particular fragment, manually define the retention policyRT_fandS_f, the fragment will have retention policyRF_fand scopeS_f(no union)?
If we don't define those 2 properties, the fragment will inherit them from the selector?
@m4rk055 : my understanding is that in this case, if S_a includes CONTACT.ADDRESS.SHIPPING, then the the Fragment both inherits the global policy RT_a and in addition abides to RT_f.
You can have two policies at the same time:
- Delete after 1 year after collection
- Delete after contract end
We must have a way to correctly resolve them.
There was a problem hiding this comment.
my understanding is that in this case, if S_a includes CONTACT.ADDRESS.SHIPPING, then the the Fragment both inherits the global policy RT_a and in addition abides to RT_f.
Can you add the information to this document that the final scope of a fragment in this case is union of both scopes?
You can have two policies at the same time:
What happens if two policies are conflicting?
- Delete after 1 month
- Keep for at least 1 year
Can you add information on how to resolve them?
There was a problem hiding this comment.
I added clarity about this
Co-authored-by: Marko <[email protected]>
Co-authored-by: Marko <[email protected]>
|
@m4rk055 @filipblnt , please:
Please follow DR-0002. Filip is marked as the Assignee, and it is written that It is the responsibility of the Assignee to act and not let PR remain at a stale point |
| - **Configuration Maps**, defined at configuration time: | ||
| - *Retention Policies*: For each **Known Selector** (or Data Category implying every `selector` used within that Data Category) a data [Retention Policy](./RFC-PRIV.md#retention-policy), | ||
|
|
||
| - *Provenance*: For every **Known Selector** keep track of one or more [Provenance](./RFC-PRIV.md#provenance) objects. |
There was a problem hiding this comment.
my understanding is that in this case, if S_a includes CONTACT.ADDRESS.SHIPPING, then the the Fragment both inherits the global policy RT_a and in addition abides to RT_f.
Can you add the information to this document that the final scope of a fragment in this case is union of both scopes?
You can have two policies at the same time:
What happens if two policies are conflicting?
- Delete after 1 month
- Keep for at least 1 year
Can you add information on how to resolve them?
It can be modeled using Privacy Scope + Data Range. No need for it now, and it would have complicated things for the relational databases.
|
LGTM after a quick review. |
|
Thank you @noelmace for this dram-come-true merge I have been waiting for quite some time. I celebrate today. 🍾 |
Making of the Schemas component of the HLA