Re: [Arx-users] Further thoughts on ArX and simplicity

[Walter Landry]

Kevin Smith <address@hidden> wrote:
> Walter Landry wrote:
> > Kevin Smith <address@hidden> wrote:
> > 
> >>Ah. Would it be accurate to say that *conceptually* a branch would be 
> >>partially ordered, but a branch *as stored in a particular archive* 
> >>would be fully ordered?
> > 
> > Not even within a particular archive.
> 
> That confuses me. A branch has a sequence of revisions, from ,0 through 
> ,n. Some patches could be swapped without changing the end result, but 
> many couldn't. Regardless of that, they *were* applied to this 
> particular branch within this particular repo in some specific sequence, 
> and that specific sequence is recorded via the parentage scheme you 
> describe below. So wouldn't the sequence represented by this branch 
> serve to "fully order" the patches within this archive?

First, let me clarify what I mean when I talk about a "branch".  Given
the picture


  ,0-AAA
     ^
     |
     |
  ,1-BBB
   ^   ^
   |   |
  /     \
,2-CCC  ,2-DDD


I call the whole thing a branch.  The individual offshoots ,2-CCC and
,2-DDD are microbranches.  Monotone calls microbranches "forks".

With that said, it is true that for an individual archive, there is
some order in which revisions were added.  However, that order is
different for different archives.  For example, I write revision
,2-CCC, and you write ,2-DDD.  In my archive, ,2-CCC came first, while
in yours ,2-DDD came first.  When we sync, then we will have different
orderings.  So there really is no natural ordering for the two
revisions.  Relying on the accidental ordering would be error-prone.

However, we both started from ,1-BBB, so it comes before both ,2-CCC
and ,2-DDD.

Cheers,
Walter