I first read about the idea of the Semantic Web today, and it totally struck a nerve in me. I've got to admit that I'm not sure that I totally understand what Berners-Lee is going for, but I really like one idea that came up while I was reading about it, and that is the standardization of cross-platform (cross-discipline?) data.
From the perspective of an amateur historian like myself, I know that there's tons and tons (or, more appropriately, gigs and gigs) of data out there about Portland and the people that have lived in Portland, but there is no easy way of searching or aggregating that data. For example, there are genealogists that are contributing great information about people from Portland's past, but they have no regard for historians of the city (like me). I'm doing research about the city and people that have lived here, but my research is not likely to cross paths with the genealogists. We're referred to as two different "silos" of data accumulation. How awesome would it be if those two "silos" were to be somehow integrated? My posted research could benefit the genealogists, and their research could benefit mine. It'd allow people to do research without being redundant in regards to previous research.
Admittedly, I haven't read much on the idea of the technology, and wouldn't even understand how this might be implemented, but the idea is really exciting to me. I think that the idea of standardizing data across all possible spectrums of collectible data is nigh impossible, but I've got to admit that it's one of the fantasies I imagine in my head while I'm drifting off to sleep.