24 Apr
2008
24 Apr
'08
3:57 a.m.
I thought about it a while back, but the potential for
OCR errors is high,
and so too the frustration level. I'd say, only if someone was to fund the
work :-)
What about a community effort? The sources start on pdf page 6 and go
through pdf page 96. That's only 91 pages. If someone was to OCR these
and place them in some distributed revision control, one page per file,
and 9 people each took ownership over 10 pages each, it wouldn't take
that much effort to get it done. Someone would have to write scripts
to glue the pages properly back into files, but that should be a fairly
minor effort in comparison. Finally someone would have to get build
tools appropriate for processing the files, and feed back the errors
to the contributors to help fix up (or provide tools for contributors
to test their work independently).
I could commit myself to 10 pages if others were willing to come forward
and take a chunk.
the s2 tape in the
PDP-11/Distributions/research/1972_stuff area contains
userland binaries and libraries from 1972, so there's a strong possibility
that the kernel in the PDF document would be able to execute the binaries.
Restoring most of a working 1972 unix software system would be incredible.
Warren
Tim Newsham
http://www.thenewsh.com/~newsham/