So, Michael (who works in the Archæology library at UT, I believe) apparently mentioned RFIDing all the books in the library to a guy there, and got a lecture on why he was stupid for even talking about it.

The recap Bryan (and I, being nosy) received was: 'Michael, there are twelve million volumes in the library.'

There are numerous reasons why this is a bad objection:

1. Firstly, have the publishers do it, or just do it as books are acquired.
   
2. Every time a book is checked out, RFID tag it. Now the most popular books, at any rate, have been tagged.
   
3. Undergrads are cheap--bordering on free. Unleash them on the stacks.
   

The benefits are many:

1. Even if you only get the 20% most popular books (see item two), you've probably covered the vast majority of activity.
   
2. You don't even need to sort the books anymore. Better, you can sort them by popularity. (Caveat: being able to browse the physical volumes is a very useful capability--just a list of metadata is frequently insufficient. You usually want to browse by subject, not browse random books.)
   
3. If someone fails to return a book, you assemble a strike team and visit its locale. 'Overdue fine? Oh, no, we figure the collateral damage to your apartment during extraction is penalty enough.'
   
4. Now you don't need an expensive television to give you cancer. The books can do it too!
   

I've proposed that we (about twenty Senecans) purchase some twelve million RFID tags and spend a night tagging every volume in the building. Naysayer heed this warning!