Conversation

Notices

1. Brandon Hall (bthall@bikeshed.party)'s status on Saturday, 20-Jul-2019 12:22:47 EDT Brandon Hall
   Am I correct in thinking that #Git addresses the issue of "who has what versions of X and what are the differences between those versions?" The flow graphs illustrate changes to versions, over time ... While always looking forward in time.
   
   This feels like a solution to problems that emerge from working with strongly linear systems, such as email. ("Linear" is variously thought of and defined, and I'm not familiar enough with it to give a good definition, but note that even though the branching nature of Git would suggest that it isn't linear in terms of always being directed in the same absolute direction (angles emerge), it is never directed backwards.)
   1. Ashwin (ashwinvis@mastodon.acc.sunet.se)'s status on Saturday, 20-Jul-2019 14:46:40 EDT Ashwin

      in reply to

      @bthall
      You mean:

      https://en.m.wikipedia.org/wiki/Directed_acyclic_graph

      1. Andrew Miloradovsky (amiloradovsky@functional.cafe)'s status on Saturday, 20-Jul-2019 15:23:06 EDT Andrew Miloradovsky

         in reply to

         @ashwinvis @bthall "Linear" means total order, that any two elements are in a relation (precedes, follows, or both). Partial order doesn't require that (elements may neither precede nor follow one another); path-connectedness of vertices in a directed graph is an example.
         To be quite precise, one distinguishes (partial) orders and pre-orders, in the sense that the former also require that precedes-and-follows is iff they're equal w.r.t. some deeper equality…

         1. Andrew Miloradovsky (amiloradovsky@functional.cafe)'s status on Saturday, 20-Jul-2019 15:31:42 EDT Andrew Miloradovsky

            in reply to

            @ashwinvis @bthall That is to say, path-connectedness relation in a directed graph (not necessarily acyclic) is only a pre-order (i.e. just reflexive and transitive), since all the elements in a non-trivial cycle would follow and precede each other, yet be different vertices.
            While in a DAG, there are only trivial cycles, so if one vertex follows and precedes another, they're necessarily the same. So it is a "proper" (without "pre-") partial order.