double post please delete
from test images in article posted by bgm it seems like iterative parallax mapping have better visual quality
here is paper if someone is interested 
http://www.cescg.org/CESCG-2006/papers/TUBudapest-Premecz-Matyas.pdf
here are papers about relief mapping
http://fabio.policarpo.nom.br/Publications.html
and here is paper about cone mapping
anyway - I like more first demo too cause it looks more 3d but at low angles it’s not looking so good so ability to automaticaly flaten texture at these low angles is great thanks
I don’t think it works on my computer (mac). The plane on the right looks good, but it has been made with the old method. The plane on the left seems to simple be using normal mapping, I can’t really see the difference. I just tested martinsh’s file on the other topic, who is using this method (correct me if I’m wrong), and here again, it just looks like if it was only using normal mapping, there isn’t any real relief impression.
Woah! Those papers are awesome psychotron! Thanks for the links! 