Hi, sorry for being one day late with my comment.<br><br>I approve of the EEP and feel that it should be implemented.<br><br>However I am a little sceptical of the implementation provided Björn. If I read it correctly then in an "unknown" attribute all atom/integer expressions at any level will be converted to {atom,integer}. This may be a felt to be very limiting and it would require a careful explanantion of what an attribute is.<br>
<br>Richards solution would probably be worse as it affects normalise/1 used anywhere.<br><br>Robert<br><br>