[Simh] Questions about PDP-7

[Bob Supnik]

SimH stores 18b data, whether disk or DECtape, right justified in a 32b 
container.

As for ASCII data... "it depends." DEC's own software for the PDP-7 used 
a mishmash of character sets, including FioDEC, Baudot, and the early 
form of ASCII associated with the KSR-33. The official assembler didn't 
even allow for storage of ASCII characters, only of FioDEC, Baudot, and 
BCD (Analex printer code).

The assembler shown is clearly not the standard PDP-7 assembler. For one 
thing, it has lower case characters. So you'd either need to have the 
assembler listing to know how ASCII is stored, or you'd need to find 
some representative code sequences to see how text is handled.

That said, there are four possibilities:

- two characters per 18b word. The questions would be whether 8b or 7b 
characters, and whether right or left justified.
- five 7b characters per pair of 18b words, left justified. This is 
equivalent to PDP-10 ASCII storage and was used on the PDP-15 for its 
line printer.
- three characters per 18b word. If there is no sign of upper-case text, 
then it's possible that the character set was reduced to 6b and packed 
three per word. Take 040-137 as the character set, subtract 40, let UC 
represent LC alphabetic characters.
- one character per 18b word. Very wasteful, but possible.

/Bob

On 2/25/2016 12:00 PM, simh-request at trailing-edge.com wrote:
> Message: 1
> Date: Thu, 25 Feb 2016 08:04:01 +1000
> From: Warren Toomey<wkt at tuhs.org>
> To:simh at trailing-edge.com
> Subject: [Simh] Questions about PDP-7
> Message-ID:<20160224220401.GA8375 at minnie.tuhs.org>
> Content-Type: text/plain; charset=us-ascii
>
> Hi all, recently we unearthed a set of pages that contain assembly listings
> of the PDP-7 Unix kernel and some associated user-mode programs. See the files
> 0*.pdf athttp://www.tuhs.org/Archive/PDP-11/Distributions/research/McIlroy_v0/
>
> There's a move afoot to see if we can bring this up on a real PDP-7 and also
> on SimH. I've set up a mailing list if anybody wants to help out, and my
> initial Github repository is athttps://github.com/DoctorWkt/pdp7-unix
>
> I've got some PDP-7 questions and also some SimH PDP-7 questions. What is
> the format of disk blocks that SimH stores for the 18-bit systems? In other
> words, how are the 18-bit words in each block stored in the bytes of the
> host systems? At some point we will need to write a tool to generate a
> filesystem, so this information will be crucial.
>   
> A general PDP-7 question: how are ASCII strings stored in memory? Are
> they 7-bit or 8-bit? I'm assuming that two ASCII characters are stored
> in one 18-bit word, but which bits go where? If I had the string "abcd",
> which bits go where in the two 18-bit words?
>
> Thanks in advance for all your help, and also thanks to RMS and others
> who wrote Simh as it's been an invaluable tool to keep the PDP-11 and
> Vax Unix systems going.
>
> Cheers, Warren