I wonder why "scanning" is required here; it seems to me like the effort to set up and OCR this is greater than the effort of typing it in manually. Its certainly not going to be worse (to an order of magnitude) than toggling the boot loader of that era system anyway.

Yep. Split that into bite sized chunks and use something like Amazon Mechanical Turk, as captcha or some other form of crowd-sourcing.

For quite some time I've been toying with the idea of building CPLD/FPGA recreations of pioneering computers like the Alto, the Lilith, the Symbolics series, the Tektronix storage CRT machines and the PDP-[1~8]. I know people have done it with MSXs and Apple II's, C-64's and even the Cray I, but it would be lovely to be able to gift someone with a simple to use (I understand operating a PDP-x is NOT simple) recreation of this cherished machines.

Of course, when appropriate, an as faithful as practical recreation of the original keyboards (or front panels) should be made available.

Surely this can't be arbitrarily re-licensed as GPL, or were these documents re-licensed at some point?

I think this would fall under the 'Ancient Unix' license that Caldera set up many years back: http://www.tuhs.org/Archive/Caldera-license.pdf (unless you want to be picky about the definition of 'version 1' here).

Edit: I'm wrong, that specifically says pdp11, so the pdp7 code isn't covered.

I'd add that, given the subsequent court cases, it's unclear that Caldera even owned the rights to offer that license.

The scans I looked at are almost illegible. The op codes can be deciphered because there are only a limited number of legal combinations. The variable names are another story.

I wonder how many 'security' holes and exploits are in this original version of Unix.