ITS pictures

[larsbrinkhoff]

I have tried to find pictures of ITS machines. I haven't found much.

[larsbrinkhoff]

So it's probably the AI lab moby memory. In some photos it's seen near the PDP-6 console. In others, it's not there so maybe moved nearer the AI PDP-10.

[larsbrinkhoff]

A frame from https://archive.org/details/AIfilms/x01-eye.mp4, 24 seconds in. Is this the AI PDP-10?

[larsbrinkhoff]

Tom Knight wrote about the picture with Horn:

It’s the AI KA-10, probably before we were running ITS on it. The KA was on the left, with the moby memory, the original 16K DEC memory, then the SC disk controller. At an angle, the PDP-6 was next, first the AC bay, then the processor, then the 340 display.

Early on, before the KA pager was installed, the PDP-6 ran time sharing, while the KA was stand alone, being used for pager mods and moby memory debugging. It also was used for some robotics work, and I’m sure BKPH was doing some stand alone robotics work on the machine. The display was part of a temporary A/D and D/A controller for robotics work. It was not around for very long.

Once the pager was installed, ITS moved to the KA and the PDP-6 was used for the standalone robotics stuff.

[larsbrinkhoff]

AI lab PDP-6, from a 1981 article. I'm not a hardware expert, but it looks to me to be gutted.

[jh95468]

What makes you think it's been gutted?

The cabinet doors are open, and what you see is the machine's backplane, where all the logic modules are interconnected by wires. The modules themselves would only be visible through the doors on the back side of the cabinets.

Typically you would keep the doors closed. Otherwise the airflow coming from underneath the false floor wouldn't make it to the top rows of modules, which might overheat. All that's missing is a Tektronix 'scope hooked in to that bed-of-nails backplane....

It looks pretty normal to me. Even the stacks of coke bottles against the back wall. That's how the 9th floor was. Clearly not a staged-for-marketing photo. /Jack Haverty

[larsbrinkhoff]

Sorry, it looked to me that the open door was not a door but the inside of the cabinet.

[larsbrinkhoff]

Moby memory at CHM: https://www.computerhistory.org/collections/catalog/102731715

The object is a memory unit. The front panel door contains a control panel , which has several indicator lights, switches, dials and a "Percent of normal gauge". The control panel is divided into four areas labeled, "Address register", "Data register", "power control", and "Test control". On the blue part of the door "Go Check" is handwritten in pen. There is a Mass. Inst. of Tech. (MIT) sticker as well as a sticker reading "PROP. of U.S. GOVT. USN". Taped in the Data register section there is a fortune cookie fortune reading, "A good memory does not equal pale ink." There are another two fortunes taped to the interior of the control panel. One reads, "Vanity plays lurid tricks with our memory". The other reads, "You will hear pleasant words which will remain in your memory". Taped to the lower section of the interior door there is a paper with a matrix regarding quick-latches and cable and pin connections. On the power supply the is a partially filled in table reading "Measure with machine stopped". There are five repair tags from A.L. laboratory. The machines is densely wired and populated with circuit boards of various sizes.

[larsbrinkhoff]

The KL-UDGE. See #622.

[larsbrinkhoff]

AI PDP-6, source: https://www.computerhistory.org/collections/catalog/102657022

[larsbrinkhoff]

Article from The Lewiston Daily Sun - Jan 23, 1967.

High resolution photo from https://www.theatlantic.com/photo/2017/02/50-years-ago-a-look-back-at-1967/516174/

[larsbrinkhoff]

Another view of the PDP-6.

[larsbrinkhoff]

Tom Knight's comment quoted above says the PDP-6 sat at an angle, so I think it's that arrangement we see here.

The bright box behind someone's head could be the Systems Concepts DC-10.

[larsbrinkhoff]

Possibly Richard Schroeppel playing on behalf of Machack VI in a Boston chess tournament.
EDIT: Schroeppel confirms it's him.

Source: https://youtu.be/zOP1LNr70aU?t=433

[larsbrinkhoff]

Sussman, PDP-6, Spacewar consoles. Maybe DEC 16K memory? No moby memory.

In this video https://youtu.be/2MYzvQ1v8Ww?t=175, Sussman says the number to that phone is 6765, which is fib(20).

[larsbrinkhoff]

Eugene Charniak at Datapoint 3300. Black KA10 in the upper right corner.

[larsbrinkhoff]

DM's pager panel. This photo and the one immediate below courtesy of J. Noel Chiappa.

Source: https://gunkies.org/wiki/File:DM-10.jpg

[larsbrinkhoff]

E&S Matrix Multiplier.

[SimHacker]

There should have been a little numbered knob that lets you adjust the number of Zorks allowed to run at once. (It would usually be set to “2”.)

-Don

On 6 Oct 2019, at 16:01, Lars Brinkhoff notifications@github.com wrote:

DM's pager panel.

https://user-images.githubusercontent.com/775050/66270344-8d792580-e852-11e9-8771-1c783f7ef5f4.jpg — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/PDP-10/its/issues/425?email_source=notifications&email_token=AAIDL3BIC3FUSNIWFXLRGQ3QNHVURA5CNFSM4C4PRO7KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAOKZJA#issuecomment-538750116, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIDL3CNZGBUWY7ZQVGERADQNHVURANCNFSM4C4PRO7A.

[jh95468]

I'm wondering why running more than 2 Zorks (or anything on Muddle) would be an issue.

By the era of Zork (mid/late 70s), the DM PDP-10 had a lot of memory and disk space (by the standards of that time). The processor was the same, but programs like Zork spend most of the time idle, just waiting for the user to type something. Also, Muddle had evolved a lot to include mechanisms for dealing with limited memory, e.g., much of the Zork code would have been in memory shared among all active Zorks, so having several of them wouldn't consume much additional memory or swap space.

We (DM) spent a lot of effort to get big programs, like Zork, to be very efficient in use of memory and other physical resources, but I suspect much of that work never got documented or published in journals et al.

/Jack Haverty (MIT-DM 1970-1977)

[larsbrinkhoff]

@bawden was kind enough to send me some photos of panels he has.

---

The Systems Concepts DM-10 pager.

---

Tom Knight's TK-10 terminal controller.

---

Some kind of memory. The panel matches schematics for a DEC Type 164. And it's possibly the same type as pictured above: https://github.com/PDP-10/its/issues/425#issuecomment-461700097

---

Last but not least, the ML console!

[larsbrinkhoff]

@masswerk says every large PDP came with a special "DEC chair". I think I can see it a few times in the photos posted here.

[larsbrinkhoff]

I think this is the DMCG PDP-6 because 1. Licklider, 2. it's not the AI PDP-6.

[jh95468]

I concur. That's the DM PDP-6.

The door in the back wall led to a workshop (for repairing Fabritek memories!) and the area where Multics lived. The DM PDP-10 console was in the same line of cabinets, to the left of the PDP-6 along with ARM-10 memory, SC pager, TTY controller (aka Morton Box), and tape drive. There was another row of cabinets behind the PDP-6/10 row, containing disks and controller, more memory, and the ARPANET IMP.

[larsbrinkhoff]

Thanks @jh95468. I got hold of Jeff Harris in the Licklider photos, and he pointed me to an interview with Al Vezza:

our PDP-6 running the ITS (Incompatible Timesharing System) didn’t have a disc operating system,so we bought discs and a smart freshman, Scott Cutler, adapted the Artificial Intelligence Labs’ PDP-6 disc operating system to our discs.

This is the first I heard that the DM PDP-6 had disks. Is this true, or do Vezza have the 10 and the 6 mixed up?

[larsbrinkhoff]

I got some information from Scott Cutler. He wasn't sure about the exact timeline, but he said he used the new ML PDP-10 for testing disk code, and transferred this code to the DM computer. Since ML arrived in late 1971 and has ITS ported during 1972, we can infer that the DM machine was the PDP-10, not the PDP-6. This is also consistent with other people's recollections that the DM group relied on DECtape storage in the early days.

[jh95468]

I joined Lick's group in early 1970. At that time, the main machine used by the group was the PDP-10. There were no disks yet, everything was done with DecTape and some paper tape. All the user terminals were on the 9th floor in the computer room. I don't remember the exact timeline either, but over the next year or so much effort was put in to adding memory, tty controller, and disks to the PDP-10.

In 1970/71, the PDP-6 was next to the PDP-10 "main" bay (with the switches and lights) to the right of the PDP-10, and all of the memory, tapes, and eventually disk controller were installed in two parallel lines of racks. Underneath the raised floor, there was a rat's nest of cables interconnecting all the memories, processor(s), and I/O systems. Since the PDP-6 and -10 were of similar architecture, it was possible to move memory, disk controllers, etc., from one to the other by "simply" moving the cables. Moving the cables was a tedious, messy, unpleasant job working under the floor. But I do recall several major projects to move things around as we got more hardware into the picture. The huge steel I-beam under the floor which was put in when Multics occupied the space made that work especially difficult.

When I was there, circa 1970, the PDP-6 was no longer used for timesharing. But it was still very useful for system debugging, and a small amount of memory was usually kept on the -6 to make that possible. New ITS features, and new hardware, were often debugged using the -6 running ITS in standalone mode. That gave lots of system debugging time while still keeping the timesharing ITS running for general use by the group.

I don't recall exactly which things were debugged on the -6 rather than directly on the -10. My guess is that the IMP software was developed on the -6, but that something like the memory mapping hardware had to be developed on the -10. It was fairly straightforward to move tapes back and forth so you could debug on the -6, then try it on the -10, and if necessary go back to the -6, and repeat until done. It might involved unplugging/plugging some cables too.

The DM disk controller was physically in the back row at the left end, and the ML machine was installed directly to its left. The PDP-6 was quite far away at the far right end of the front row. I can't remember exactly what was there when the first disk arrived, but if the ML machine was present it surely would have been easier to use it rather than routing all those (very expensive) cables to get to the PDP-6.

I'm not sure how much of this kind of detail would have been known by Management (Al Vezza and Lick). We moved things around a lot as needed for debugging, repair, etc. So it's possible that the PDP-6 was officially supposed to be used for the disk integration, but Scott and others noticed it would be much easier to just use the not-yet-operational ML machine so they just cabled it in. Mike Brescia (user MB) was another of the system hackers around that timeframe, he may remember more; I'll ask him. Also Allen Brown (ALB) if anyone knows how to contact him.

I do remember one nasty bug that was found in that disk code. When we eventually removed the DecTapes and their ITS code, the disks stopped working. &^$*&^%$^&%$..... It turned out that the disk code accomplished some cleanup work at the end of a system call by JRSTing into the middle of the DecTape code, which of course was no longer there.

What you read in the official reports and management summaries isn't always quite what actually happened.

Fun times.

[eswenson1]

ALB has an entry in the MIT Alumni Directory. It lists an email address for him, as well as his linkedin page. I won't post here due to privacy concerns, but I'm happy to forward a message to him. If you have access to the Alumni Directory, Jack, you can look up his entry there and contact him directly. If you can't do that, I'm happy to forward a message to him and he can reply to you.

[eswenson1]

I sent a message to ALB pointing him to this thread.

[jh95468]

Thanks Eric, I just sent a similar message to Mike Brescia.  I can probably get in to the Alumni data, just have to remember how.....

/Jack

On 1/31/20 10:51 AM, Eric Swenson wrote:

I sent a message to ALB pointing him to this thread.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/PDP-10/its/issues/425?email_source=notifications&email_token=ABLZOWCL7FDTS26DHYC3X33RARXKFA5CNFSM4C4PRO7KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKPUC2I#issuecomment-580862313, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABLZOWBZVWKWIPCCJ63SB2LRARXKFANCNFSM4C4PRO7A.

[larsbrinkhoff]

ML KS10 at @davemcguire's Large Scale Systems Museum in Pittsburgh. Photo by @hanshuebner, with permission.

[larsbrinkhoff]

Jeff Del Papa posted these yesterday and today. He says the first is from the AI 340, and the second is a 16K memory from ML.

Sources: https://www.reddit.com/r/vintagecomputing/comments/fdpy2s/status_lights_digital_decus_model_30_display_from/
https://www.reddit.com/r/iiiiiiitttttttttttt/comments/foyeuk/one_rack_try_8_or_more_status_panel_for_memory/

[larsbrinkhoff]

PDP-6 room with moby memory. From http://www.bitsavers.org/magazines/Computer_Design/196710_Fabritek_MT.pdf

[larsbrinkhoff]

From https://people.csail.mit.edu/bkph/phw_copy_demo.shtml

Horn, PDP-6, moby memory, 340.

[larsbrinkhoff]

Robert Weissberg using the E&S LDS-1 and tablet. Lebling identified the location as "ninth floor, just to the left of the DM PDP-10."

[SimHacker]

Nice sweater! It was COLD in that room.

[larsbrinkhoff]

@atsampson commented:

That matches what Tom Knight said - "The KA was on the left, with the moby memory, the original 16K DEC memory, then the SC disk controller. At an angle, the PDP-6 was next". The DC10 isn't in that photo but it's been added in the one with BKPH below. 81-mathlab.mp4 shows the moby memory, two cabinets to the right, then the third in the process of (de-)construction. That said, the three narrow cabinets all have different indicator panels: https://user-images.githubusercontent.com/775050/87186255-7a1c4e80-c2eb-11ea-83ec-a00877159769.jpg The left one's a 16k memory, but what are the other two?

I believe the middle panel is a Type 164, and quite possibly the same as @bawden's photo.

According to TK, the right panel would be the Systems Concepts DC-10. Here is the 81-matlab picture.

[larsbrinkhoff]

Better version of Moses and Winston photo in https://github.com/PDP-10/its/issues/425#issuecomment-461419973

Now we can see the AI PDP-6 in the background.

Source: http://blogs.wgbh.org/on-campus/2014/5/28/mit-commemorates-50-years-computing/

[davemcguire]

On 10/20/20 5:05 AM, Lars Brinkhoff wrote:

Better version of Moses and Winston photo in #425 (comment) https://github.com/PDP-10/its/issues/425#issuecomment-461419973

Now we can see the AI PDP-6 in the background.

And a GT-40, beautiful!

       -Dave

-- Dave McGuire, AK4HZ New Kensington, PA

[larsbrinkhoff]

This used to sit on MIT-MC. "Who done ITS, where is ITS, when was ITS, how to do ITS, why is ITS?"

[masswerk]

BTW, sorry for my confusion regarding your post. For whatever reason, I was thinking that this had been posted by @sohkamyung, but it had been you all along. (Apparently, I had been looking at too many posts at once?)

Best, Norbert

On 15.03.2021 13:25, Lars Brinkhoff wrote:

This used to sit on MIT-MC. "Who done ITS, where is ITS, when was ITS, how to do ITS, why is ITS?" [1]

-- You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub [2], or unsubscribe [3].

Links:

[1] https://user-images.githubusercontent.com/775050/111153089-c95d0200-8591-11eb-81bb-efde41ca1083.jpg [2] https://github.com/PDP-10/its/issues/425#issuecomment-799377855 [3] https://github.com/notifications/unsubscribe-auth/AAHBNC546JESC6N6G5OD2NLTDX4DBANCNFSM4C4PRO7A

[ams]

I feel that this should be on a t-shirt ....

[bawden]

What's the date on that newspaper? I'm guessing 1983 or 1988 (both years had a Tuesday October 25).

[larsbrinkhoff]

Alan, I think 1983. And according to information in https://github.com/PDP-10/its/issues/181#issuecomment-267783313, MC went away in September 1988.

[larsbrinkhoff]

DM console photo from Chiappa. Note it says "KA10" instead of "PDP-10".

Source: https://gunkies.org/wiki/File:KA10FrontPanel.jpg

[larsbrinkhoff]

TK posted this photo of the AI PDP-6 console. Source: https://twitter.com/TomKnightSynBio/status/1480635229001953280

[davemcguire]

On 1/11/22 1:20 AM, Lars Brinkhoff wrote:

TK posted this photo of the AI PDP-6 console. tk-pdp6 https://user-images.githubusercontent.com/775050/148891468-ae261951-1852-42f8-a06c-c3b2c1e38b79.jpeg Source: https://twitter.com/TomKnightSynBio/status/1480635229001953280 https://twitter.com/TomKnightSynBio/status/1480635229001953280

So very sad to see a "trophy console" like this. :-(

          -Dave

-- Dave McGuire, AK4HZ New Kensington, PA

[hanshuebner]

Am Di., 11. Jan. 2022 um 17:23 Uhr schrieb Dave McGuire < @.***>:

So very sad to see a "trophy console" like this. :-(

...says the man who has three warehouses full of ancient computers and considers renting a fourth one :D

Happy new years!

-Hans

Message ID: @.***>

[davemcguire]

On 1/11/22 11:25 AM, Hans Hübner wrote:

So very sad to see a "trophy console" like this. :-(

...says the man who has three warehouses full of ancient computers and considers renting a fourth one :D

Hi Hans, don't confuse LSSM with someone else, nearly everything we have at the museum is complete, not "rip the console out and trash the rest"!

And no rent, we own these buildings. It's the only way to keep it all safe.

We just need to get more people involved. But as you know, it's hard work, and nobody is interested in hard work.

Happy new years!

Same to you, I hope you will be able to come back to visit soon!

         -Dave

-- Dave McGuire, AK4HZ New Kensington, PA

[larsbrinkhoff]

Bill Gosper hacking LIFE on the PDP-6. Email from Steven Levy, who wrote "here's a photo of Gosper hacking LIFE that I got from Mike Beeler."

[johnlorentzson]

Picture of KATIA. Does it look like it's running ITS?

I know this picture was discussed several years ago, but I felt I should add to it. I spoke to one of the people in the picture and he's certain it wasn't ITS. He says Stacken did run ITS for a short while, but not in that machine room. He also says that computer only ran TOPS-10.