>You were right, the HTMLParser of htmllib is more permissive. He just ignores the bad tags ! The HTMLParser on my distribution is a she. But then again, I am using ActivePython on Windows...