Just in case I didn't express myself clearly... I had to brutally force it in there with SQL (administration interface didn't allow that by itself, so there's no bug for the time being ;)
BTW. I cannot register to the eZPedia, unknown account error during sign in validation... ;(
Fact remains your not the first to ask how can this goal be reached.
The SQL you used to reach this goal is valuable to others looking to do as you have done ... without having to study as hard as you must have to learn which SQL will work.
Which leads me to stepping up and asking for you to share the solution not just the question, 'Is this ok' when we can't tell what you have done really without the sql.
Their is always another reason hidden from view on the other side of the coin. Every time I turn it over while contemplating the question I think of another reason to share ....
Email kc at ez dot no perhaps he can help you outside the forum get yourself an ezpedia.org account ... if ezpedia is not functioning for you try a blog post ...
I had sharing in mind from the beginning: http://ryba.ez-publish.net/index.php/ez_publish/lab/custom_top_level_nodes_is_that_possible_round_i_exploration
I started with the question because I wasn't sure if it was a good direction and actually worth of spending time on. Now, that I dug a bit more, I suspect that no matter what the benefits, it will require too many core modifications... unless all served by own extension... But that's still a guess at this point.
You are not alone in searching for an answer to these questions. You can do <b>anything</b> using eZ ...
Though I strongly suspect my subconscious disagrees with your fears regarding core changes in kernel to reach your goal while retaining existing functionality and stability ...