Good news: a copy of PDP-6 ITS has been found!

Last year I read in the book The Tomorrow Makers that MIT professor
Gerald Sussman had a 1967 listing of ITS in an office drawer.
Naturally, I asked him if he still has it. Just a few hours ago I heard
from Sussman, and he confirmed he still has the listing which is ITS
version 138. He also has a similar vintage copy of Maclisp verison 97
and a corresponding compiler.

On Tuesday, January 18, 2022 at 11:10:08 AM UTC-8, lars...@nocrew.org wrote:
> Good news: a copy of PDP-6 ITS has been found!
>
> Last year I read in the book The Tomorrow Makers that MIT professor
> Gerald Sussman had a 1967 listing of ITS in an office drawer.
> Naturally, I asked him if he still has it. Just a few hours ago I heard
> from Sussman, and he confirmed he still has the listing which is ITS
> version 138. He also has a similar vintage copy of Maclisp verison 97
> and a corresponding compiler.

And are we going to be able to see either of them?

gah4 wrote:
>> he confirmed he still has the listing which is ITS version 138. He
>> also has a similar vintage copy of Maclisp verison 97 and a
>> corresponding compiler.
>
> And are we going to be able to see either of them?

I think so! Sussman is willing to lend me the listings for scanning.
I'm just about willing to hop on the next plane, but I'll hold off for a
bit. There are also some other options which I'm looking into.

The hard part will be OCR or manual typing. Any volunteers?

Lars Brinkhoff <lars.spam@nocrew.org> writes:
> I think so! Sussman is willing to lend me the listings for scanning.
> I'm just about willing to hop on the next plane, but I'll hold off for a
> bit. There are also some other options which I'm looking into.

There are probably people around MIT who could make a photocopy and mail
it to you. How big a listing are you talking about? ITS wasn't all
that large.

Paul Rubin wrote:
> There are probably people around MIT who could make a photocopy and mail
> it to you.

Right, that's one of the options.

> How big a listing are you talking about? ITS wasn't all that large.

Sussman said around 125 pages. Easy enough to scan, not so easy
to type.

On 1/19/22 1:36 AM, Lars Brinkhoff wrote:
> Sussman said around 125 pages. Easy enough to scan, not so easy
> to type.

A friend of mind got some code working that came from scans of hundreds
of pages of patent documents from the USPTO.

My understanding is that the OCRing got quite close save for some
obvious problems. Said obvious problems were corrected by hand. In the
end he says that the OCR + editing was a lot faster than typing
everything manually.

--
Grant. . . .
unix || die

Lars Brinkhoff <lars.spam@nocrew.org> writes:

Sussman said around 125 pages. Easy enough to scan, not so easy
to type.

If this is the listing that I think it is (that GJS showed to me many
years ago), we're talking about fan-folded lineprinter paper. I would
expect that to take a special scanner wouldn't it?

Alan Bawden <alan@csail.mit.edu> writes:
> If this is the listing that I think it is (that GJS showed to me many
> years ago), we're talking about fan-folded lineprinter paper. I would
> expect that to take a special scanner wouldn't it?

These days you can do very well with overhead cameras. The million or
so books scanned by archive.org were done that way, using high end
consumer digicams on a big mounting frame, with a glass platen to
flatten the pages before photographing and a foot pedal to move the
glass up and down, but a lot of that stuff is to increase scanning
efficiency. There is a photo here:

https://blog.archive.org/wp-content/uploads/2020/07/bookscanning.jpeg

I believe Google Books also uses overhead cameras but they don't flatten
the pages, and instead unwarp them in software. I don't know what their
camera setup looks like.

125 or so pages isn't a crazy amount of typing anyway, if split up
between a bunch of people. So it might be enough to just start with
reasonably careful digicam or phone camera shots. I'd type a few pages
even though I don't feel likely to try to run the software. It would be
interesting to just read the code.

Grant Taylor wrote:
> My understanding is that the OCRing got quite close save for some
> obvious problems.

Ok, I used some free online OCR tools and I suppose I got what I paid
for.

Alan Bawden wrote:
> If this is the listing that I think it is (that GJS showed to me many
> years ago), we're talking about fan-folded lineprinter paper.

I think it must be!

> I would expect that to take a special scanner wouldn't it?

Someone mentioned a "roll scanner". Sounds promising, but some I found
online are around $8,000.

Anyway, MIT Libraries say they can do it. Maybe best to leave it to the
professionals.