Gary Jones replied; "And Commodore was easy to work with back then. When we asked for documentation, they sent us a stack of documentation about four feet high. They were willing to tell us everything about their machine. Since we had to design some custom hardware to go inside, it really helped to know exactly how everything worked."
"It just turned out that it was a good machine. The things that make a machine good for playing games also tend to make it good for processing and displaying data, because you've got some of the same problems. You need a very efficient, very fast operating system, and the Amiga has that and very little overhead too. That's what makes it nice; we don't load down the system running the overhead; we can just process the data."
"Most of our customizing is hardware customizing. The Amiga operating system is flexible enough that we have to drop into assembly only once in a while to initialize some of the special boards we use, but otherwise the operating system is fine; we don't do anything unusual with it. We use it just like it is and build hardware for our interfacing requirements because we have to pull the data out of the data bus in this building, process it, and put the data back in."
[0] http://obligement.free.fr/articles_traduction/amiganasa_en.p...
I suspect this is what would be difficult to replace. A bespoke wireless sensor monitoring and communications system, integrated into an old heating/cooling system. Modern system just would not integrate into this, without a lot of work. You would have to replace everything, probably including the actual heating/cooling systems themselves.
I wonder if it just integrated into a pre-existing control network:
> The Commodore Amiga was new to GRPS in the early 1980s and it has been working tirelessly ever since. GRPS Maintenance Supervisor Tim Hopkins said that the computer was purchased with money from an energy bond in the 1980s. It replaced a computer that was “about the size of a refrigerator.”
> ...
> A Kentwood High School student programmed it when it was installed in the 1980s. Whenever the district has a problem with it, they go back to the original programmer who still lives in the area.
> ...
> Hopkins said the system runs on a radio frequency that sends a signal to school buildings, which reply within a matter of seconds with the status of each building. The only problem is that the computer operates on the same frequency as some of the walkie-talkies used by the maintenance department.
If programming it was a school project for an 80s kid, might as well try to make its replacement a school project for a 20s kid. If a kid did it back then, the integration can't be that hard.
You need to find a decision-maker to agree to the liability of letting some kid implement a system today. What if something goes wrong and all the children roast alive or something? Then the parents would sue. It'll never happen today, people are terrified of being wrong so no is always the easiest answer.
Also, yes integration would be hard today. This same pinhead decision maker would require it to run on Windows for "support" and "security".
That 80s Amiga is probably incredibly reliable and robust with its real multi-tasking OS and doesn't require "security patches".
What a shame Commodore dropped the ball on it, it was ahead of its time.
>project for a 20s kid.
I'm honestly surprised that the system hasn't been commandeered by some other enterprising student with some cheap boafang from amazon just trying to trigger with key presses.
You wouldn’t need to replace the actual boilers/chillers/air handlers/terminal units, just the temperature controls/control wiring/control relays/control computer. There is mechanical equipment out there that is still controlled by pneumatic controls and that equipment can be retrofit to use digital controls.
Replacing the mechanical units may be beneficial, condensing boilers are extremely efficient and so are newer chillers/etc.
Ofc I'm just speculating here, but based on the article's description I suspect this whole wireless integration could be stripped. Maybe in the 80s they saw the need for a central computer running all the logic, but I suspect that nowadays this computer could be replaced with programmable controllers, one at each building. Industrial controllers are very reliable and there are available professionals for support.
Nah, anything driven by relays and or pneumatic controls can be replaced with modern, low cost PLC hardware. It is common for old chemical plants to have their control systems replaced while still operating. Control is handed over from the old system to the new system (hot cutover) one control loop at a time. All you need is adequate planning.
Source: I literally just lead a project to upgrade the PLC controls on a few air handler units and build entire server and network infrastructure. We had to run Ethernet to each of these from the network closets because they previously had went to a Windows XP box in the maintenance guys office. There were many challenges, one being getting the configuration off a couple of the old PLCs to make the new ones even talk to them.
https://tech.slashdot.org/comments.pl?sid=7543887&cid=499003...
>> When the Amiga system originally went in it was controlling well over 100 buildings throughout the district, including the entire GRCC campus at the time. The Amiga replaced the head-end of the system, which was experiencing expensive hardware failures every year ... and you couldn't get parts for that mini-computer on e-bay. It is essentially acting as a huge database (schedules, configurations, control programs, history, etc.), system manager, and monitoring system ("head-end") for the remaining 19 buildings HVAC systems. If the Amiga goes down, the buildings will continue to operate using the configurations last received, with most of the individual device controls being able to be manually overridden inside each building, albeit with less energy efficiency. What you will loose is the ability to change schedules/custom control code/configurations and the ability to centrally monitor the performance of the buildings.
>> Each building has one or more local control systems, and those systems communicate back to the central head-end over radio-modem (there was no district-wide network back then). Schedule and other control changes are sent to the buildings and alerts/reports are sent back. That old equipment in the buildings, even older than the Amiga, is what dictates the radio communications link. They incorporate specific protocols for keying up the radio that are not directly compatible with a newer serial to Ethernet type device that would seem like a logical replacement.
>> The control systems themselves gather temperatures, both inside and outside the building, look at trends and do predictive control of the equipment to accomodate scheduled use of various areas of each building. For the day, this was very advanced building control and offered significant energy savings, as well as comfort in the buildings.
>> Over time, as buildings have been updated, sold or replaced, the local controls withing those buildings have been replaced with newer/more modern controls that communicate with newer central control systems. Replacing these controls that are local to the buildings is what is responsible for the majority of the cost I would say.
>> As far as the Amiga system itself, I believe most of the components are still the original. The hard drive may have failed twice over the years, requiring a rebuild from backups. They did pick up or have donated a few Amiga systems to use as parts as needed, but the system has proven to be very resilient. Obviously, Monitors, Keyboards and Mice can only take so much use without needing to be replaced. Without this, the system likely would have become inoperable and unservicable many years ago, or been incredibly expensive to keep running.
>> From a technical stand point, the Amiga was selected because at the time it was the only "Personal Computer" (PC) that had a true pre-emptive multi-taskng operating system. It needed to be able to handle multiple processes simultaneously, including interfacing with the systems, maintaining settings in the database, monitoring the system as well as support for both local and remote access to the system simultaneously. Basically, its capabilities fit the need. While for nostalgia reasons I would hate to see it go, it has been 30 years and I think the system has done its job. Replacing a building's control system doesn't happen overnight, and when you are talking 19 buildings with ancient (yes I am calling myself ancient I guess) control systems, it is going to take money and time. The payback in energy savings, comfort and safe control of the buildings though I think justifies the cost.
Ha, bullshit. Someone decided they liked the Amiga and wanted to play with it. I am sympathetic!
Also each school is probably on the net these days, so replace with some Wifi IOT device.
Or Samsung "Smart Things".. but unlikely to last 30 years..
You get it, yup.
>Also each school is probably on the net these days, so replace with some Wifi IOT device.
What I use are Ethernet to Bacnet gateways. Each building gets a gateway device with Bacnet nodes connected over a serial daisy chain. If the HVAC device isn't networkable, serial devices do the relay flipping. IT folks get the data into different applications.
Integration and maintenance are expensive.
Is there anything that actually works with it directly? I have a growing number of connected devices around me, and none of them can talk to SmartThings, except floor heating controllers that do so by half-broken integration with Tuya...
Trying to figure out (without success) how to hook A/C units to SmartThings, I learned that apparently Samsung/ST is abandoning their existing "we'll figure it out for you in the cloud somehow" architecture, moving towards "please buy our hub/edge device, and run some Lua scripts on it" architecture, which actually doesn't feel like an upgrade.
https://www.theverge.com/23801118/imax-movie-palm-pilot-oppe...
No documentation, of course, so a metric ton of black-box testing to find out what did what.
In the end we managed to design a pretty nice interface/HMI, and a API so that it could be monitored online.
I wonder if they received the funding for replacing it?
Just receiving a funding doesn't imply the need to burn the money.
Any replacement using wifi +- cellular and some javascript pile of dependencies is likely to work less reliably: for any budget of b=1 million it'll cost a multiple x>1 of that.
If you start now, and finish before it breaks, you don't have to have downtime.
This would have been done by an enthusiastic high school student with both programming, electronics, maybe some RF knowledge too. Radio amateur? I find these stories kind of sad. Clever system built by someone talented, at a low cost -- replacement with a modern system costed in the millions. Those talented people still exist but they wouldn't get a look in with most RFPs.
The intermediate step was to emulate the old systems software on Windows XP machines until all the other signal systems could be upgraded.
https://web.archive.org/web/20201112040545/https://www.smh.c...
Impressive!
It is indeed impressive, but I hope there's a contract for this. Everything else about this story sounds insanely incompetent, so I... I worry.
In principle there's nothing wrong with a Motorola 68K running a system like this, but the rest of this is damned.
However, it's still cool that it's running. I love seeing old tech that's still just running. In the early 2000s, one of my clients was running a woodcutting system off a ZX Spectrum. They wanted to replace it, but the cost to rewrite their custom software would have been so much that they ended up buying a few used Spectrums as backup.
However, I am anxiously awaiting my Spectrum Next...it seems to be inching towards a delivery date.
The 2015 bonds were approved: https://grps.org/reimagine/
Here are the GRPS munis, issued in early 2016 totaling about 134m: https://emma.msrb.org/IssueView/Details/ES360882
Unfortunately, no luck finding any interesting tidbits on this Commodore or whether it was replaced. Maybe theres an RFP out there?
Man, I'll build it for $750k - will buy a few Raspberry Pis and we're good to go...
Keep it classy, America.
Moreover, this story is very old, is probably no longer true, but continues to resurface from time to time.
I think a novel benefit of open source is that it shields engineers from the often fatal consequences of being on the wrong side of a billionaire. As a consequence, there are now a couple of generations of engineers who can't relate to what life in their occupation was like before the open source movement went mainstream.
this is happening more and more that I see something in my reddit RSS feed first, then hours later it pops up on HN
Another example from today https://news.ycombinator.com/item?id=37078047
https://www.reddit.com/r/todayilearned/comments/15ncz6j/til_...
Assumptions about IoT availability in terms of bluetooth/BTLE or 2.4/5GHz might be challenging for sure. Especially inside whatever HVAC metal/concrete cage etc (just run ethernet). Still though I mean it's running on a legacy UHF (?) system right now.
There are some industrial ISM bands at like 900 MHz or 432 MHz or something, and you can get modem boards that run on that. The 432 MHz aren't too fast but you just need to send "thermostat X turn on" and get a confirmation back. Which probably should be signed commands so some some yahoo can't have fun with the thermostat, even on this system, hopefully. But still, in a lot of cases probably a commodity solution would work.
serial communications were interrupt driven which didn't lend itself well at all to cooperative multitasking schemes. (gotta service those buffers asap)
>but HVAC systems aren't just basic relays and pneumatics.
Old ones are.
Feel free to withdraw your overly aggressive comment any time you'd like.
The person you’re replying to was saying that the old method of HVAC/process control, prior to digital controls, was pneumatics and relays. Old HVAC control systems were indeed pneumatic systems, I’ve seen some in the wild in the last few years in the US.
Can you imagine what Samsung (or any large corporation) would want for a centralized controller for an entire school district?
If they do nothing now, instead of taking a bet the new system will be more stable than the old one (a bet against the 30+ years odd) you don't have to have downtime.
No... It's known by the guy who wrote it 40 years ago that happens to still live there and want to chip in to fix it
> maintained
Well, they said parts are difficult to find now. I would not consider something with no replacement parts well maintained
> taking a bet the new system will be more stable than the old one
What? It's a gamble to leave it as it is with a bus factor of 1 and rickety old computer parts. Setting up a replacement doesn't imply this one goes away. If the new one is worse, they can have a plan to switch back to the old. If this computer catches on fire, they have to pray they can find a working replacement that the original implementor can install.
So yes, it's well know by this person and his company. I'd be surprised if in 30 years he didn't document the system and teach it to his employees.
> Setting up a replacement doesn't imply this one goes away.
Agreed, it generally implies a replacement project that goes on for years before being declared a failure at taxpayer expense.
> Well, they said parts are difficult to find now. I would not consider something with no replacement parts well maintained
The parts are "difficult" to find, yes. So if you say there are "no replacement parts", you are moving the goalpost from "possible" to "impossible" to obtain, while they are available for money on ebay. A few thousands will get you many vintage computer parts!
BTW spares for the new system will also be necessary, if you plan to have it work for another 30 years. They may be more expansive.
> It's a gamble to leave it as it is with a bus factor of 1 and rickety old computer parts
We do not know how many employees he has, and whether the replacement parts are "rickety and old". If you are used to maintaining vintage system off ebay parts, you will see lots of perfects parts available for about everything.
> If this computer catches on fire, they have to pray they can find a working replacement that the original implementor can install.
They say they have 3 spares, so no prayer required. Documentation and backups are likely available.
I just don't understand how so many people are afraid of the existing and want to rewrite everything from scratch. Looks like Resume-Driver Development to me: good for the developer and manager, bad for the company and its clients.
I'm not so sure. There's a huge electronics/hardware hacking scene nowadays fueled by Raspberry Pis and Arduinos. It might actually be easier today for a kid to learn what he needed to know.
I mean, back in the 80s at best the kid probably only had the manual, a couple of programming books, and maybe a magazine subscription for reference.
JS is appearing everywhere, and Python isn't far behind. Both seem approachable for modest projects. Basic had its charms, yet also becomes difficult to maintain as it grows -- not unlike many modern stacks. I suppose the primary difference is the stack is deeper, with many more layers.
I recall a friend going deep with Basic in the 90s while I was more pragmatic. I only knew Batch scripting, Kilk n Play, and various other scripting and gaming tools. Yet I had already made several prototype games, more functionality, 3D models and character animations, and even learned some BBcode and HTML to help mod a gaming forum. He had ... some very unimpressive screen drawing demos.
The problem is you can’t just deploy this and forget it without regular security updates (with the likelihood of breakage in compatibility - python libs are particularly bad with this)
That said, the original radio based system is probably extremely easy to mess with if you have a basic SDR. It’s just that no one bothered so far.
As to be effective in today's world, already knowing that is more than many teenagers know beyond doing likes in social media.
My point is that kids in the 80s interested in computers likely had the skillset to do this job. Kids interested in programming today, they are more likely to be qualified to make the school a website, not program its heating system. I don't doubt some would be able to do it, but the field has grown so much that beginners necessarily must focus on specific areas, and I bet that is not often low level enough to do this job.
There are still plenty doing that; often getting their first exposure via the cracking scene.
I generally agree with some of your comment, but this sentence confuses me a bit.
I'd argue that the only reason that it does not require security patches is because it's not attached to any network. Though, probably the only actual reason would be that nobody is looking too closely at the whole system, because that wireless communication system is a network. And we really were not good at keeping computers sufficiently secure from any network in the 80s...
(Besides, what does "real" multi-tasking OS mean here? And why is multi-tasking relevant for this job? It seems like in this case a simpler OS would be more reliable and robust, while still being able to perform the job.)
The Amiga was the most performant system, its components (i.e. audio, video, etc.) were all able to work independently of the CPU. It also came with 3 networking APIs built-in (one of which was TCP/IP) which probably helped a lot with the coding.
It was also the only machine with a true multi-tasking OS. Everything else was either co-operative or non.
If an issue happened with one piece, it wouldn't have brought down the whole system because everything was able to run independently.
As for security patches, I was expressing a bit of frustration there because I have seen over and over a Windows "security patch" introduce bugs or break functionality. You didn't have that problem back then with the Amiga. It just worked.
I agree the Amiga was a good choice for the time to administer multiple machines.. and it came with decades of support from the original programmer!
> It was also the only machine with a true multi-tasking OS. Everything else was either co-operative or non.
> If an issue happened with one piece, it wouldn't have brought down the whole system because everything was able to run independently.
The Amiga was an amazing systems and the best choice they had available to them at the time but...
it's easy to forget it lacked memory protection. That's one of the design decisions that made it run so well on the limited hardware of the time.
It also meant that if one of your multitasking programs went down there was a very high likelihood it would bring the rest of the system down too.
Ah, that was my confusion. I thought we would compare with systems nowadays, to replace the Amiga.
> A DOS PC could have run a multi-threaded program, but if an issue happened then I guess the whole system would go down
Eh, not really. DOS was for almost all intents and purposes as single threaded as the other systems. Segmentation allowed for some sense of relocation, and this was used for TSRs as "background programs", but that's not much of a departure from any of the other systems you mentioned.
> As for security patches, I was expressing a bit of frustration there because I have seen over and over a Windows "security patch" introduce bugs or break functionality. You didn't have that problem back then with the Amiga. It just worked.
Yeah, but that was only because there simply was no untrusted networking. All the other systems fared the same in that regard. And had the Amiga survived, it would need exactly as much security patching as the other systems.
The real risk here seems to be "the HVAC stops working, the person who implemented it is indisposed/gone, and there is no entity to recover money from to get someone else to fix it"
In the US, school taxes to a school are generally from the local neighbourhood and not the city at large, so advocating for spending to replace something that isn't broken might mean needing to take currently allocated funds from other programs if some contractor says they can only give you a new control system if you agree to let them replace all the boilers as well.
I guess I sometimes agree with this sentiment, but this seems like an inappropriate place to invoke it. Is it really so insanely risk averse to get professional HVAC people to install and maintain the HVAC system for a large public facility?
If you mean today, sure, but why replace something that has worked so reliably without fail?