On Thu, Feb 03, 2005 at 07:55:53AM +0000, Kees Jongenburger wrote: > On Tue, 1 Feb 2005 22:11:22 +0100, Rob van Maris <[EMAIL PROTECTED]> wrote: > > > > On Feb 1, 2005, at 9:42 PM, Homeijer, Michael wrote: > > > > > Using Apache BCEL (Byte Code Engineering Library), I create a > > > BeanFactory that wraps node with classes that can be evaluated using > > > jxpath or EL. > > > > Great stuff. But for EL both maps and javabeans are suitable, and I > > don't see any particular advantage of the latter. Because of this, I > > prefer the simplest option which is the first, i.e. a single wrapper > > class implementing Map. > The difference being that the BeanFactory allows a lot more then the > map approach. > -editing of a node > -getting the relations > > It it was only for the taglibs I would choose for the map approach > since the taglibs are power full enough to get relations etc. > I also see the bean aproach a a possible new way of extending mmbase.
I agree here, for what it's worth. I'd like to be able to specify a "high-level" interface to what is currently a bridge Node - possibly by extending it (but multiple implementations of the interface makes that a problem), or by extending an auto-generated bean. I wouldn't mind writing the whole bean class "by hand", provided I could use it seemlessly with whatever API would use the auto-generated beans. > For xjpath the map approach is not an option. I don't think either > needs to make it to the core., But I like the way Michael is thinking > about MMBase and standard apis. I have no clue what michael is thinking, but I _personally_ think that a nice extention mechanism on the Nodes is more important than just being able to use nodes easier in jstl or whatever (since I'm sort of doubting the added value of taglibs vs scriptlets/jsp methods in general, lately). Joost. _______________________________________________ Developers mailing list [email protected] http://lists.mmbase.org/mailman/listinfo/developers
