Hi Alexandru, Well, if I am the only one making efforts to understand others view
points, then this discussion is useless :-(.
I'm feeling the same way, glad you brought it up.
As a matter of fact, I've been designing APIs for quite a few years, and I can guarantee you that setValue() should always have different semantics than remove() (they are different at language level; and if you check different API you'll see that almost everywhere this difference is kept).
Um, yes. I'm in violent agreement ;) Quite obviously, the java language supports any object variables being set to null, and the some of the apis also supports setting of null values, for example: http://java.sun.com/j2se/1.5.0/docs/api/java/util/HashMap.html. The key is not removed, just the value. I mean, its fine that there are other ways of doing things, but I'm surprised that some people don't seem familiar with the Java approach. I'm NOT surprised that some java developers find it odd that the JCR standard veers from the Java standard, especially when use cases have been provided that describe mis-match problems (based on the fact that JCR and Java work differently wrt nulls). I hope our input can result in some change. Best, Mark I will probably stop by now, considering that I brought enough details
to the discussion and presented my point. It is now up to the others to see if they are valid/helpful or not -- I only hope that people looking at it are trying to keep their mind opened as I am. bests, ./alex -- .w( the_mindstorm )p.
-- Best, Mark Waschkowski
