Posted by hyperpat on August 1, 2008
Over at Tor.com, the new website tor has established for interactive discussion of all things speculative, Jo Walton posted a commentary on just what the difference is between sf and fantasy, and just how some books really can’t be pigeonholed into one category or another. Which leads to a problem for brick-and-mortar stores that try to shelve these two categories separately, as wherever they put such books, it will be missed by those expecting it to be in the other category.
The problem is really much more general than just the divide between sf and fantasy, and gets into an entire field of library science dealing with indexing and cross-referencing massive amounts of data, at least some of which is subject to highly subjective evaluation by those irrational and sometimes contradictory beings called humans. Coming up with at least partial solutions to this problem is important. Many, many times in the world of science today, a fact discovered in one field, say entomology, has great relevance to another field, say a search for cancer-curing drugs in the field of medicine. But this fact won’t be noticed by the medical researchers unless they have some tool that properly indexes the discovered fact as being relevant to their field.
The Google model for searches is great for topics of wide interest. What it won’t do is find data that is obscure and of importance to only a few specialists. It’s also completely mechanistic – it can’t get that “aha!” moment that humans do when seeing one data item that suddenly provides an answer to a problem in something apparently totally unrelated. Artificial intelligence probably won’t really be useful (or really ‘intelligent’) until it can make such connections. We’ve still got a long way to go before we get to “Computer! Tell me why these Acturians are blowing up all our grain silos on Deneb IV.” (certain Star Trek episodes to the contrary).