Teresa Nielsen Hayden has posted a great blog-entry on the kinds of things that we lose when we burn libraries:
Recently I was deeply vexed by the news that a professional author, who of all people should know better, has dismissed the burning of the National Library in Baghdad on the grounds that any book destroyed in the fire could simply be reprinted. There are moments when you find out more about someone's scholarship and research habits than you'd ever want to know.
Apparently he was unaware that whereas reprinting might serve to reconstitute his high school library, substantial research libraries contain all sorts of odd things, possibly odd old things, some of which may be sole copies. This goes double for major research libraries, which are textual mathom-houses. Moreover, at the time that some of these odd things were catalogued, they may not have been properly recognized for what they were. (This is, incidentally, why I hate having to do research in closed-stack library systems: You have to take the cataloguer's word on everything.) Anything can turn up there.
Just this year came the news that a big wodge of Tolkien manuscript had turned up in a carton in the Bodleian Library. It hadn't been lost, exactly; but it hadn't occurred to anyone who knew about the material's existence that it might command enough general interest to warrant publication. It has now been published.
It's full of linky goodness and startlement.
