"Peak Document"

That's a term used in the title of the presidential address at the American Historical Association meeting. What McNeill is talking about is the surge of information coming not from documents but from science--especially genetics.

It's a valid subject, of course, but I admit when I first saw it my thoughts went in another direction; the change in sources in the current and coming eras because of digital media.  An example, when I was hired by ASCS people had improved the document management systems involved. The Commodity Credit Corporation board had a permanent secretary and an assistant, the board made decisions based on "dockets" which were systematically filed.  Most decisions within ASCS generated paper documents, memos and letters, all routed through clearance channels and eventually filed in the Secretary's Records or administrator's.   

As a failed historian I was intrigued by the processes.  The paper files didn't capture everything--there was a lot going on in the agency which wasn't fully  documented (particularly the political maneuvers) where the documents were like an iceberg, only a small part visible

By the time I left FSA, this picture was changing.  Partially it was the result of personnel changeover--the institutional memory of the reasons behind practices had been or was being lost.  Partly it was a change of norms--new people and new problems had new ways of doing things, often resulting in faster action but a diminished historical record.  Much of it had to do with automation, both the problems and processes of implementing policy with compers in the county offices and the new powers of communication conferred by new technology.

One example was the "wire notice".  Urgent messages to field offices would be sent by telegraph, which meant going through the telegraph office, therefore required official authorization, and permitted central filing of the message.  Once email arrived, it was possible for anyone to email anything to anyone with no central file. (Of course, this didn't happen immediately.) And for a number of years there was really no system for recording and filing such messages.  Supposedly after 30 years NARS has enforced systems in the agencies, but I'm dubious. 

The bottom line--in the 1970's a historan could look at the official files in the National Archives and do a reasonable history.  I doubt that's feasible for th 2000-2010 perioc