locking (taxonomy p.2)
Dave Long (dave@sb.aol.com)
Mon, 02 Sep 1996 05:28:46 -0700
In light of the taxonomy I proposed, it ought to be
clear why Christopher and Yaron would have conflicting
views over the relative importance of locking and
versioning.
Look at RCS: check-in and check-out deal with particular
revisions in the repository, but the fact that owner of
a locked work area file is the only one with write access
is a consequence of concurrency control between multiple
authors in the same work area.
One may also consider versioning without locking: the
work area is free game, and after the authors have
finished stomping all over one another, someone marks
a "stable" change set in the version system, and the
process begins again...
It seems pretty clear that although versioning systems
may implicitly acquire and release locks, locking itself
is a distributed authoring requirement.
-Dave