I'll have to move release date, planned for September, in October.

As I mentioned in the comment of previous post, the main reason is how Menu Specification took longer to implement than I expected. Now I understaind why LXDE guys decided to not implement it: it is simply too complicated for this simple task.

Yes, it is powerfull, but the things could be simplified a little bit. Hell, you even have a pattern matcher in it!

For example, why do I have to care about legacy menus; they are not used any more and the spec should mention it in initial draft, not carry it through all versions. You implement it once, do the conversion and there are no more needs for legacy menus.

Or merging: couldn't one perl/python script solve this? Even better, both things could be put in one script called when desktop starts.

Current implementation in EDE is not full; menu merging and layout customizing is not yet supported and I will leave them for the future version(s). Those things that are supported are passing official menu regression tests.

Of course, partial implementation will not cripple the final menu presentation if sane menu configuration is used (e.g. like the one shipped with Xfce). Or you are definitely trying to complicate the things :P