It seems that in the new sharing model, the DAG doesn't really mean anything since transitions are approved on a per-share status. However, the DAG can actually be interpreted as an Auto-Ack of the dependent functions, as mentioned in issue #53.
The Interceptor was originally for "optimistic"/"enclave"/"discoverable" data elements that can be automatically accessed by a function during runtime. However, with the new implementation, the "discoverable" data elements are instead added to a share, before computation. With the Shares, I think it might be safe to say that all accessible data elements to a function will be used.
Derived Data Products, in our model, are now treated as just another Data Element. I think this clears some things up.
It seems that in the new sharing model, the DAG doesn't really mean anything since transitions are approved on a per-share status. However, the DAG can actually be interpreted as an Auto-Ack of the dependent functions, as mentioned in issue #53.
The Interceptor was originally for "optimistic"/"enclave"/"discoverable" data elements that can be automatically accessed by a function during runtime. However, with the new implementation, the "discoverable" data elements are instead added to a share, before computation. With the Shares, I think it might be safe to say that all accessible data elements to a function will be used.
Derived Data Products, in our model, are now treated as just another Data Element. I think this clears some things up.