search

PDP-10 / its

Incompatible Timesharing System
Other
848 stars 80 forks source link

Gould printer #863

Open larsbrinkhoff opened 6 years ago

larsbrinkhoff commented 6 years ago

CHSGTV; SPGLPT 19
DEVICE; JOBDEV GLP
.GLPR.; GLP NOTICE
GLPSPL; GLPDEV 130
GLPSPL; GLPDRW 4
GLPSPL; GLPFNT 20
GLPSPL; GLPPKO 22
GLPSPL; GLPSPL 228
GLPSPL; GLPSPL NEW
GLPSPL; GLPSPL TODO
GLPSPL; GLPSPT 14
GLPSPL; TS GLPSPL
.GLPT.; BAD GLPT
.GLPT.; GLPT 4
.GLPT.; -GLPT- QUEUE
GLPT; SPGLPT 205
GLPT; SPGLPT 635
GLPT; USGPLT 98
INFO; GLP 22
INFO; GLPT 10
SYS2; TS GLP

larsbrinkhoff commented 6 years ago

@atsampson, I wonder if this is related to VERSA somehow?

larsbrinkhoff commented 6 years ago

Maybe there were more than one kind of Gould printer.

VERSA can make some kind of bitmap for the Gould, right? But I also see a Gould line printer mentioned.

Also, I saw something about MC having a Gould printer, but GLPSPL; -READ- -THIS- says it's for a printer in building 38.

larsbrinkhoff commented 6 years ago

INFO; DIR > says "The Gould electrostatic printer spooler for printing graphics or text files on the printer in building 38."

Oh, just look at INFO; GLP 22.

"There are 2 hardware modes in which the Gould printer can operate: LPT (line printer) mode and bit-image mode."

atsampson commented 6 years ago

The "Miscellaneous Information" section of INFO; GLP has some useful notes:

The Gould spooling software can simulate the XGP printer and has several capabilities which make it more general (e.g. rotation, more fonts, graphics, etc). It uses the same fonts as the XGP, can print TEX files (which the XGP can not print directly).

Maintainers: the GLP device (its a job device) is obtained by assembling AI:SYSENG;XGPDEV > with GLP==1. The :GLP program is obtained by assembling AI:SYSENG;XQUEUE > with GLP==1.

The main spooling program is MC:GLPSPL;GLPSPL > which is a LISP-based system with many support files (ARDS, LPT, XGP, GLPFNT, GLPDRW, GLPSPT, etc.). See the comments in GLPSPL for how to dump out a new spooler. If the spooler dies for some reason (as shown by :P ;JGLP), first see if JLK is around and ask him to look at it, otherwise just run GLPSPL;GLPSPL which will cause a new spooler to get launched. The GLPSPL program gets started automatically when ITS comes up.

VERSA is written in MIDAS, so it probably didn't reuse code from GLPSPL directly, but it's doing essentially the same job.

larsbrinkhoff commented 6 years ago

SYSTEM; IOELEV 432 says:

GOULDP==0 ;DOESN'T HAVE GOULD LPT (ANYMORE)

I think that confirms MC had the Gould at one point.

MITS.S; CONFIG 849 configures MINITS number 42 "PLASMA" with a Gould driver.

larsbrinkhoff commented 4 years ago

If it went away in January 1976, the Gould printer had a short lived affair with MC.

ToTS tape 3100078, .MSGS.; MSG-> 6


MOON@MC 01/20/76 20:44:42 Re: GOULD LPT HAS BEEN FLUSHED
To: * at MC
THE GOULD LINE PRINTER INTERFACE WAS CRASHING THE
SYSTEM (ACTUALLY THE I/O PDP11), SO I FLUSHED IT.
larsbrinkhoff commented 2 years ago

GLPTP is 1 only in SYSTEM; CONFIG 17. In CONFIG 27 from may 1976 it was set to 0.

In SYSTEM; IOELEV, GOULDP is 1 in versions 48-61; timestamps range from 1975-12-05 to 1976-01-22.

Is reappears as part of the RTTY11 (#1837) in November 1976.

larsbrinkhoff commented 7 months ago

I found a document by @macrakis that explains the Gould 5200 printer was part of the Plasma Group facilities. It was driven from a PDP-11 (which also handled Tektronix and Ann Arbour terminals, Grinnell frame buffers, etc). This PDP-11 was sometimes called the "Remote 11", or "RTTY11". See also #1837.

On the ITS side, most of the supporting code was written in Maclisp, and some in MIDAS. The unspooler sent data to the PDP-11 over the GOULD Chaosnet contact. The Gould printer was a 200 DPI raster device, just like the XGP. It accepts most of the same file formats the XGP does, plus also TeX output (DVI perhaps?).

macrakis commented 7 months ago

Credit for most of the software for the Gould printer, the Grinnell frame buffers, etc. goes to John Kulp and Charles Karney, if I'm not mistaken. I've CC'd them so that they can correct the story.

When I arrived in the group, most of this was already working. The connection between the PDP-11 and the PDP-10 when I was there (1977) was not Chaosnet, but a single serial line. The main reliability issue was that the input stage of the receiver card would be blown out periodically by electrical transients. So we kept spare input cards available.

The big visual difference between the Gould and the XGP was that the Gould had clearly defined square (or rectangular?) pixels, whereas the XGP had some sort of additive Gaussian-like pixels. So on the XGP, if you printed

######## ########

you would get a rounded inside and outside corner, so to get a nice result on the XGP, the glyph would look like

# ########

But fonts which were optimized this way for the XGP (the modern equivalent would be device-dependent "hinting") looked terrible on the Gould.

We also ran a downline debugger for PDP-11 programs (I called it HALI) which I hacked together from Radia Perlman's PDP-11 emulator and debugger (RUG) plus a simple load/deposit layer in the PDP-11 communications system.

      -s

On Sun, Feb 11, 2024 at 2:19 PM Lars Brinkhoff @.***> wrote:

I found a document by @macrakis https://github.com/macrakis that explains the Gould 5200 printer was part of the Plasma Group facilities. It was driven from a PDP-11 (which also handled Tektronix and Ann Arbour terminals, Grinnell frame buffers, etc). This PDP-11 was sometimes called the "Remote 11", or "RTTY11". See also #1837 https://github.com/PDP-10/its/issues/1837.

On the ITS side, most of the supporting code was written in Maclisp, and some in MIDAS. The unspooler sent data to the PDP-11 over the GOULD Chaosnet contact. The Gould printer was a 200 DPI raster device, just like the XGP. It accepts most of the same file formats the XGP does, plus also TeX output (DVI perhaps?).

— Reply to this email directly, view it on GitHub https://github.com/PDP-10/its/issues/863#issuecomment-1937843396, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGO6CRA74ESTOEBDYTHF7LYTEKL3AVCNFSM4E6RYD4KU5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCOJTG44DIMZTHE3A . You are receiving this because you were mentioned.Message ID: @.***>

cffk commented 7 months ago

All the credit for the hardware and software belongs to John Kulp. I was a heavy user of the graphics capabilities in this group of MIT graduate students working under Abe Bers. I also would have discussed requirements with John (my office mate at the time). But John was the one to make everything happen. This explains why it took John substantially longer than me to get his PhD.