On Wed, 22 May 2002 21:39:54 +0100, Matt Patterson wrote: >Is there an easy way around this problem? Is there a problem? With UTF-8, you need no stinkin' entities! But if you insist on burning cpu cycles: >>> ''.join([c.encode('ascii', 'ignore') or "&#%d;" % ord(c) for c in u'= \u8314=E4=F6=FC=DF?abc']) '茔„”á?abc' Ciao, J=FCrgen