GitHubber wants to revive the first Unix in a PDP-7 emulator An IT lecturer from the Australian state of Queensland wants to revive the very first Unix – the version written by Ken Thompson on a Digital Equipment Corporation PDP-7. While the PDP-11 is probably the most famous of the series – a genuine watershed in computer history, and a successful system that sold 600,000 units in its …
It was from UNSW that Lions' Commentary on UNIX 6th Edition snuck out, despite the best efforts of Bell Labs' lawyers.
is that the captchas have got particularly good!
Or, on a more serious note, he's going to need a *very* good OCR program if that's the quality of his source material. Assembly - any assembly - is bad enough to debug when it's in clean text with just typos to debug, but when it's packed with OCR errors as well... he has my sympathies.
---> the one with the hand-entered hex listing in the pocket
tesseract (now FOSS'ed by google, originally from HP?) is pretty flexible for learning dirty text.
How do I know this? I had to OCR my sodding bank statements. Subject Access Request in UK, gets you a pile of photocopies typewritten paper - even if you ask for electronic format .Funnily enough, the same happened with my credit union account. I am assuming for security reasons the machine that accesses the records has no output except a printer. Anyone out there know if this is true?
I had to write some perl code to get sense out of tesseract and this is a clue for anyone else who wants this challenge. tesseract is pretty good at putting boxes around characters, but seems to be unable to detect machine printed lines. I wrote a perl alignment code that put the boxes on lines and grouped chars as a machine would - for numbers and mixed text works very well as word lengths are highly column correlated.
Other than that little code exercise, OCR of dirty paper has greatly improved and linux a good platform to do it on.
P.
I know for a fact that both ken and dmr (RIP) had the original source of the complete system in their archives, on DECtape & cards. Email ken & ask ... the estate of dmr is probably clueless on the subject. I'm pretty certain I still have a complete system original Unics[0] card deck myself (made from a DECtape copy I got at Berkeley, the tape itself is long gone), and I'm certain that I'm not the only one. Unfortunately, us little guys can't share the wealth without the remains of MaBell's lawyers taking us to the cleaners ... you're on your own there ;-)
PDP 7 emulators already exist, see SIMH for a pretty good example.
[0] If you think I mispleled that, you aren't qualified to comment on the subject ...
Dennis Ritchie isn't available to ask for assistance.
We = my esteemed colleague (genius).
Assembler language source code available only in an old, faded, marked up, printed hard copy. Roughly a hundred pages. Scan in, OCR (what a mess), endlessly correct, fix & add comments, endlessly test compile, rinse and repeat. A few month's solid work, hundreds and hundreds of hours, just to get to the starting gate.
Then start the minor modification which was the actual project.
Oh.
A few years ago I corresponded with DMR about precisely this project.
He mentioned having those listings; we pondered whether it would be
practical to get them machine-readable and try to get them loaded and
booting on one of my pdp-15s to recreate an operational 18-bit Unix...
but ultimately decided it was impractical at the time.
And from 2007...
"Post by Dennis Ritchie
Was there a PDP-9 [Unix] port? Mr. Ritchie, sir?
Yes, Ken moved the -7 system to both the PDP-9 and PDP-15
just to try them out. Minimal effort, just some new drivers, no
effort to take advantage of extra features on either. Total time
they actually ran was probably measured in hours or a few
days at most. One of the machines ran a step-and-repeat
camera for making IC masks.
And all that software is presumably long gone along with the PDP-7 version.
Oh, well."
Glad he was wrong. Nice to see those listings have survived and surfaced. Maybe one day I WILL get them running on my pdp-15 :)
Is there a reason they can't scan and post a high resolution image of each page of the code and have multiple people retype it based on what they see? Send out 3-5 copies of each page and diff it to see if everybody agreed on it when they submitted the text, if not, take the answer with the most in agreement. I believe Ancestry was doing something like this for old immigration records that were originally hand-written and weren't able to be OCR'ed. I'm sure a lot of geeks would be happy to retype a page or two in exchange for a mention in the revival documentation for the project.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
