Someone asked, I think last week - what the moral responsibilities of an IA were. Here's an article that really brings up some TRUE instances where end-user impact of IA-type decisions can be less than what the designer intended: http://online.wsj.com/article_email/0,,SB1038261936872356908,00.html