S/34, S/36, and S/38 as well...
I think we still had 5-1/4" when the AS/400 came out in 1988. I know for sure some of the early models had 3.5". Heck, we were happy when we got 1/4" QIC tape!
2 Likes
When I arrived at Ft. Riley, KS in 1984, as a Honeywell mainframe technician, we had to clear out a conex that still had parts stored in it for the old and departed 360/370's they had prior. I recall the iron core memory cards, of which I had only learned of and never seen in school, being quite the attraction by all involved. I think they were maybe 32K cards?
I did score a nice old engineer's tool case that still had a few decent IBM specific repair tools left in it, including a nice oiler for printers and card punch/readers. FWIW, I repaired DD card punch/readers that were used with the Honeywell, as well as CDC 40MB 5-platter removable pack drives (we had 2 stacks of 4 of these in our computer van), STC vacuum tape drives, and lots of really old comms gear.
Operationally, the Honeywell application as written was lacking so the unit ported their existing 360 applications (all COBOL) onto the S/34, then subsequently S/36, and so I heard a few years after I left there in 1988, eventually the AS/400. Used as their primary front end, where the Honeywell system issued by the US Army served only as the primary interface to the Army centralized systems.
And here we are today discussing the difficulties encountered trying to make Alexa turn on the damn lights.
3 Likes
I only worked on paper tape when I was going through my (2nd) MOS training in 1984.
I got a 3-day pass for getting 100% on both the written and practical test for the Decision Data Card Punch/Reader. They offered the pass because since the school started (maybe 2 years?) they hadn't had anyone get 100% on both for that particular piece of gear. Tried to weasel out of it and I went up the COC to the School Commandant to get my damn pass.
Sadly, I think if you put a DD and 5 IBM card punches in a lineup I could pick out the DD.
1 Like
Eric,
And here I thought I was the only techno-nerd dinosaur in this forum when I started, and certainly the only typesetting nerd! My first job out of college, in 1970, was working as a FORTRAN and Assembly Language programmer for Radnor Graphic Arts, the main typesetting firm for TV Guide, which had its headquarters down the street. In the cellar was a cavernous room full of manual- and paper tape-driven hot lead machines, about thirty or so of them, both Merganthaler Linotypes and the part-for-part identical Harris Intertypes, along with a couple of Monotypes for headlines, posters and such. Upstairs, we had an IBM 1130 mini computer with a special-order 8-track paper tape punch (which tape I could read, too, along with being able to read 80-column punched cards, also at a glance) and, at the start, two Merganthaler Super-Quick optical line-by-line phototypesetting machines, which were so much faster than the hot lead machines in the cellar that they were doing more and more of the work. Within a short time, Merganthaler installed a monster at our site, their page-at-a-time phototypesetter, the Linotron 505, and the first installed at a customer site. The fonts were stored on large film negatives, and the image was created by projecting the actual character images on film. 1970 doesn't seem all that far away in time, to me, and what's most amazing is that I was amazed and impressed with that very primitive technology.
Regards,
Jeff
PS: Just to stay a bit on topic, I'm pleased to note that, as I've been trying inexpensive, generic A19 Zigbee lamps, the Linkind- and nearly identical, Samsung-branded SmartThings lamps appear to have fast CPUs, sufficient memory and, based upon preliminary results, appear to be reliable repeaters and, so far, with four of the former and three of the latter installed, have caused no problems with my once-delicate Zigbee mesh. Clearly, they're made in the same factory and I expect the model number difference is due to factory branding for Linkind and Samsung.
I remember the DD reader/punch as we had a few of them at remote workstations when I worked as Lead Systems Programmer at Rutgers U., starting in 1972. As for sadly/miraculously permanent but useless memories, I can still rattle-off the model numbers of the IBM mainframe gear we used, including our special-order 2,000 card-per-minute 2501 optical card reader and our 2540 card read/punch/sort monster, which used brushes with relatively high voltage bias, to read the cards electro-mechanically. The reader was prone to jam and, when clearing the jam, one removed the multi-column brush matrix, which often gave one a nasty shock. I just found the operator's manual online and they mention damaging the brushes, but not the fact that they remained energized whilst the system was powered-up, as you can see.
Jeff
No issues. I leave it paired as a secondary controller, and when I delete ghost devices, etc, I don't change anything on the HE side. More to the point - since there is no way to sync the primary from the secondary on HE, I think you HAVE to have the HE side up and running to do it.
I missed some of that cool gear. I started in 1976 and we were already using Autologic APS CRT phototypesetters. Some of our shops had APS-4 and I was lucky enough to start in one of the shops that had the APS-5. We had a fairly custom interface in one of the PDP 8/e systems to drive the APS through the paper tape reader port. The editing and H&J was done on another 8/e system. The APS worked off a roll of photographic paper so each batch of pages had to be developed in a photoprocessor - LogEtronics PC-18 and later the smaller PC-13. A few of our shops had a cool feeder so the paper would be cut and fed right into the LogE. We had to do it manually - advance the paper and put the APS offline, run in the darkroom, open the door on the APS, cut the paper with a quick slice of a razor blade, take the roll over to the LogE, set the roller on a little holder rack and feed the paper in, close the light-tight top, take the empty roller back the APS, thread the paper through a slot, spin it enough to hold, close the door, exit the darkroom, put the APS back on-line.
2 Likes
Unless the originals have been replaced, you likely have metal electrical boxes. The addition of a brass switch plate should make for a good radio shield. If you experience Zwave issues with those switches, you might want to consider going with plastic.
2 Likes
Reminds me of visiting my former in-laws in the Philippines a few years ago. They had mesh screens covering every window and terrible cell phone reception inside the house.
I used to discreetly move the mesh from one window a couple times a day to maintain contact with the outside world. Mosquitos be damned.
4 Likes
I don't think poor cell reception should cause the breakup of a relationship. I think you should relax your standards.
4 Likes
1 Like
One more post in the 'time flies' vein...
When I saw the title of this hilarious video Here's What Happens When an 18 Year Old Buys a Mainframe my first thought was of putting one of these 1 gigabyte beauties in my basement:
I was on the team that designed just the power/thermal controls for that 400hz, chilled water fed beast. The control hardware alone resided on a half dozen motherboards-- one per subsystem, each populated with a half dozen cards of custom LSI (yikes, another archaic term) logic and analog circuits, supervised by a herd of Intel 8048's and 8039's. They fed monitoring data to an embedded 'service processor'-- the repackaged guts of an IBM 4331 (a bonafide mini-mainframe in its own right) running a 'hacked' VM/370 pinned in main memory, its dedicated vm's tasked with processor resets and recovery, logic and power monitoring, and ultimately telling humans what to fix when things went wrong. The goal was to self-diagnose any machine fault (correlating logic errors to power transients if possible), automatically call home and dispatch a CE with the right parts in hand to minimize downtime. In reality lots of customers even then were too security conscious to allow automated outbound calls from their data centers (and many had a resident customer engineer anyway) but it was a noble goal.
Sobering to think that the tech in that YouTube video (4 generations newer than my first project) is scrap-- and that slim cabinet is now today's icon for a mainframe computer.
1 Like
Actually, most of the original switch boxes are ceramic, and, in any case, the switches in question are certified for use with metallic switch plates as the faces of the switches are plastic and the antennas are placed to take advantage of this. Thanks, Jeff
Tony,
Wow! IBM's first water-cooled system was the 360/91. An amusing story, which has never been substantiated that I've discovered, is that IBM released the specifications of what would have become their fastest system to date but secretly never planned to produce it. CDC, it is said, sued them as their claim was that the announcement alone was designed to retrieve market share from their own super computer efforts. As a result, according to the story, IBM produced ten of them, and one wound-up at Princeton U. Now, Rutgers and Princeton were separated by a mere forty minute drive, we became their emergency back-up site and they became ours. As a result, all of the Systems Programmers met periodically and more-or-less had the run of each others' machine rooms. The console was enormous and was driven by the guts of a 360/30 with direct main memory access. They had 4MB of modified 2365 memory in several, special boxes called BOMs, which included the same sort of a rack as used in a pocket door. The modified, quarter meg core boxes, which were beefed-up for speed, were bare and on wheels, and had a matching fixture for the rack in the boxes. When there was any indication of impending memory failure, a spare 2365 in a room, with another 360/30 used to constantly test it, would be rolled-out and the potentially failing quarter meg would be rolled into the room and hooked to the 360/30 tester.
The entire system, including the memory, as I recall, was water-cooled. There were two huge water coolers and (occasionally leaky) plumbing under the floor with finned radiators in all of the CPU boxes, channel boxes, etc. As a result, it was cool enough in the machine room to require sweaters, at least. Fun times, indeed!
Jeff
PS: That video, which has always given me chills, was produced by SHARE, of which I used to be a member.
Ha! I remember hearing stories about the term BOM. That acronym faced a mandatory phase-out (so the story goes) due to some alleged incidents involving air shippers and bills of lading (I also heard that when T.J. Watson got wind of the DoD's ballistic missile project he lobbied for them to include the 'C' in ICBM).
1 Like
Here's a neat bit of history. SHARE was formed in 1955 as an IBM mainframe users' group (and one of the founders, the great mathematician Dr. May Risch Kinsolving was one of my Math profs in college). The name was chosen to imply that the members would be sharing software, knowledge, etc., but some time after the founding, someone decided that, if interpreted as an acronym, it would be The Society to Help Alleviate Reduntant Effort!
1 Like
I have no doubt there is a basis for your CDC story. In the '80s anti-trust litigation era there was a ton of scrutiny re: competitive practices, leading to a very formalized alpha/beta/C-test/announce rollout protocol to avoid the perception of predatory action.
1 Like
Is there any way we can split this thread into Zigbee vs. old guys (which includes me) reminiscing about how we [mis-]spent our youth (perhaps in Lounge)? I'm just trying to think about the new user having Zigbee issues finding this thread a year from now and wondering what the heck we're talking about.
1 Like
Yes! Let's do it. Why don't you think of a name for the thread and open it? Let us know... PS: I still remember most of the 1130's twenty nine instructions and, if pressed, could almost certainly still write code in 1130 Assembly Language if I had a couple of manuals in front of me (they're all online, in any case).
Jeff Broido
Old Fart (seventy two) Retro Hippie
