[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: Time-Based Versioning - Conceptual Overview [Initial Draft]
What I like about this is that it recognizes that things and the “facts” about them change over time where as it is the “same thing”. In a world of complex attacks this basic principle would seem to be absolutely necessary.
What is not clear to me is the recognition of the timeframe some statement is true and valid (e.g. John was in Paris from 2/1/2016 to 2/5/2016) and the “timestamps” of who asserted this fact and when (Joe said this on 2/16/2016 at 4pm EST).
Given this fidelity in the relationships we can then say that Joe, John and even the location of John have a consistent identity, even as their state and information about them change.
Are you handling both of these scenarios?
Also – what are the alternative proposals for versioning?
This paper seeks to provide an overview of the Time-based Versioning concept in the context of supporting OASIS CTI TC discourse and exploration of options to meet community requirements.
Time-based Versioning provides a universal scheme that satisfies the following stated community objectives/requirements for versioning:
I only need/want the latest version of the model.
I need to establish the state of the model (1) at a given point or (2) over a range in time.
I need to request a copy of the model and all of its states (1) at a given point or (2) over a range in time.
Time-based versioning separates Structure from State which allows versioning of Structure independently of its State.
[Note: This version of the paper is being published in it''s initial draft form given comments to the list today indicating a possible cut-off of discourse on Versioning]