> params[:item].each {|id, hash| Item.find(id).update_attributes(hash)} > > Which should convert the serialized data into the acts_as_nested_set. Make > sense?
It does. I'm wondering what happens if you have a very large tree (+10.000 nodes) and you load part of it on demand. In this situation, the whole tree is never loaded at once for scalability reasons, so it is not possible to compute the left and right values client side. Maybe it's a specific use case. If so, another serializer could be written that only sends parent_id information for the loaded tree ? I never used act_as_nested_set, but my implementation allows you to add nodes in an arbitrary order, then you can "rebuild" the tree. It is also possible to add a node anywhere without rebuilding the nested set. There is an article on the difference between "adjacency list" (parent id) and what they call "modified preorder tree traversal" (left + right) : http://www.sitepoint.com/article/hierarchical-data-database/2 Something else that came to mind : is it possible to allow other parameters to be sent? Like a 'modified' flag, so we can only update the DB with modified nodes. All this is very cool to have in jquery, thanks a lot for making it happen! -- Philippe Jadin Thinkedit, a flexible data and content management system : http://www.thinkedit.org _______________________________________________ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/