On lack of support for convergence (was: Some thoughts on loom)

[John Yates]

Rob, replying to Forest Bond, writes:

> In this case bzr can't tell that B is actually meant to be successive
> work from trunk-2 aka A.

IIUC this is a consequence of bzr's current lack of support for convergence.

As Forest's experience shows, bzr's semantics are difficult to predict for
those who have not internalized an accurate system model.  That an accurate
model diverges from the natural, intuitive one is unfortunate.  It is not
dissimilar to saying:

  "Yes our system looks a lot like integer arithmetic.  We even reuse the
   same notation and vocabulary. But, unlike other practitioners, we opted
   not to implement associativity.  Do not be mislead by false similarities."

I fully understand the design choice.  Nonetheless, this is another (perhaps
unwitting) instance of bzr placing barriers in the way of acquiring and
internalizing a useful, confidence inspiring mental model.

The fact that systems against whom bzr competes for mindshare do implement
convergence probably means that bzr will find itself, most likely, forever
on the defensive on this point.

/john


-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.ubuntu.com/archives/bazaar/attachments/20080415/00913ab2/attachment.htm