It's a great pity that the first users of every computer technology are the very people who are least likely to benefit from that technology. They have already been through the ritual scarification of the present technology and so have a lot invested in it. This is the main reason our software appears so very bad to real users. As I've circulated drafts of this proposal, I've received various negative reactions to the idea of localizing information. It's interesting that the most negative criticisms come from those programmers who are best adapted to the current milieu of raw search. This reminds me of the arguments that raged on and on during the seventies about the malevolence of gotos and the advantages of structured programming. That argument lasted a full decade before most programmers grudgingly agreed to give up the ``power'' of being able to arbitrarily jump to any point in their programs without leaving a trace. The programmers who fought the hardest were those who through great effort and native cunning had already mastered the arcane art of making code work even when its flow of control looked like a bowl of spaghetti. They didn't need to abandon gotos, they thought, things were wonderful just the way they were. Well, they were wrong, and I think the defenders of raw search on the web are wrong. But that's a judgment only the future can make.