Hi all, You know, I don't really care what we do (here). But, we should not start discussing this over and over. Once we decided it would be a good idea to do the mapping the JavaScript way, that is getters are turned into (read-only) properties. To prevent the collisions, that Bertrand mentioned, we agreed to insert synthetic objects (properties, children).
Now, I am perfectly fine going back to not having the methods be properties and access properties and child nodes as properties of the node - which is equally natural, of course. But then, we should finally keep it one way or the other, ok ? BTW: If a node has a child node and property with the same name, the node is returned and not the property. And of course, this mechanism will most probably not easily support SNS, which should not be used anyway ;-) What do others thinkg ? Regards Felix Am Freitag, den 01.02.2008, 15:27 +0100 schrieb David Nuescheler: > Hi Bertrand, > > > The simplest way might be to require the use of node.getPath() instead > > of node.path, and use javascript property names for content access > > only, not for JCR getters. > I completely agree to that solution. > > I think I have a strong bias for using node.propname for accessing the > properties > and the node.childnodename for child nodes. > > One of the reasons to resolve the issue this way is that this would allow > compatibility between ujax (jst) and .esp when it comes to content access. > > (especially if we would introduce a currentNode on the server side aswell) > > I think accessing the original Node methods with node.getPath, > node.getSession() > is perfectly acceptable. > > regards, > david
