Hi, On Wed, Jun 20, 2012 at 2:34 PM, Marcel Reutegger <[email protected]> wrote: > how about going one step further and define some conditions > where *no* observation of changes is possible? > > that would allow for lots of concurrent writes without worrying > how to diff them or get them into a sequence.
The diffing part isn't a problem, as we already have that functionality. The sequence part is trickier, but it shouldn't be an issue either if we use the solution described in Michael's latest post. BR, Jukka Zitting
