Re: PROV-ISSUE-124: Constraints on Used Relation (PROV-DM and PROV-OM) [Conceptual Model]

Tim

it does exist. Indeed there are numerous constraints that I call "non-actionable", such as "traceability assertion" for example, 
which describe semantics but cannot be used to make new assertions, or even to check consistency.

There is a proposal to push all constraints into a separate section, and in that setting it will be easier to make this distinction.

-Paolo


On 11/22/11 7:58 PM, Timothy Lebo wrote:
> Luc and Paolo,
>
> Does this distinction among constraints still exist?
>
> If so, could/is it described in the latest DM?
>
> Thanks,
> Tim
>
>
>
>>>> We are proposing to make a distinction between
>>> - inferences
>>> - so-called constraints that are there for the purpose of interpretation
>>> - constraints that need to be enforced in the data model to be "well formed".

Received on Tuesday, 29 November 2011 09:22:14 UTC