[tus666]

> 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.

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.

[tivert]

> 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.

[Sunspark]

> 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.

[anyfoo]

> That 80s Amiga is probably incredibly reliable and robust with its real multi-tasking OS and doesn't require "security patches".

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.)

[Sunspark]

I am sure someone else can explain it better, but basically in the 80s you mostly only had a choice of 8-bit computers (like Apple, C64, etc.), DOS, Macs, Atari ST and Amiga. The 8-bit machines weren't really networking compatible or powerful enough to do more than 1 thing at a time (128k of ram, 1.02 MHz CPU, etc. though I suppose the 16-bit Apple //gs could have been a contender). Macs would have been way too expensive for this purpose. I don't think boards bought Ataris. A DOS PC could have run a multi-threaded program, but if an issue happened then I guess the whole system would go down. So that leaves the Amiga..

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!

[Lio]

> A DOS PC could have run a multi-threaded program, but if an issue happened then I guess the whole system would go down. So that leaves the Amiga.

> 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.

[anyfoo]

> but basically in the 80s you mostly only had a choice of 8-bit computers

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.

[xkcd-sucks]

> What if something goes wrong and all the children roast alive or something?

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"

[Sunspark]

Do you replace your working appliances and vehicles in your home in advance of their failure just in case? Or do you wait, and then spend the money for a repair? Perhaps you could have a backup system in a storage closet?

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.

[tshaddox]

> 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.

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?

[Sunspark]

If you mean in the past, this setup was probably state of the art back in the 80s. Multiple buildings networked together? Doubt that there were many HVAC guys setting that up back then.

If you mean today, sure, but why replace something that has worked so reliably without fail?

[hoten]

The number of layers added to modern programming and computer systems has gone up significantly since the 80s. It's sensible to claim that a kid from the 80s had far less to learn to be effective in the world of the 80s, than a 20s kid would need today.

[tivert]

> The number of layers added to modern programming and computer systems has gone up significantly since the 80s. It's sensible to claim that a kid from the 80s had far less to learn to be effective in the world of the 80s, than a 20s kid would need today.

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.

[paulryanrogers]

My guess is kids today can be effective with less knowledge, depending on what one is trying to do. Some prompt engineering, JS, Python, etc can accomplish a lot.

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.

[ginko]

>Some prompt engineering, JS, Python, etc can accomplish a lot.

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.

[nikau]

The Amiga programmer didn't have Google and stack overflow, they likely were more proficient than most modern devs even as a teenager based on my experience as a kid who read the Amiga rom kernel manuals cover to cover.

[pjmlp]

As 80's coding kid, I am quite confident that kids would be able to master coding Arduinos, ESP32, BBC Micro and Raspberry PIs.

As to be effective in today's world, already knowing that is more than many teenagers know beyond doing likes in social media.

[fiddlerwoaroof]

Ok, but as far as the users are concerned, do the layers add much value to this use-case?

[hoten]

Likely no, but a kid isn't going to be learning low level systems programming in highschool. Their entry point into the world is programming is going to be high-level, as that has less barrier to entry.

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.

[userbinator]

but a kid isn't going to be learning low level systems programming in highschool

There are still plenty doing that; often getting their first exposure via the cracking scene.

[dylan604]

>the computer operates on the same frequency as some of the walkie-talkies used by the maintenance department.

>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.

[quickthrowman]

> You would have to replace everything, probably including the actual heating/cooling systems themselves.

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.

[myth2018]

> A bespoke wireless sensor monitoring and communications system, integrated into an old heating/cooling system

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.

[doubleg72]

That's exactly it.. i work at hospital system and just had to deal with this myself. The HVAC guys were in to upgrade the controllers for the air handling systems. They are basically PLCs that connect to a server like anything else. The issue is interfacing with the system. These pre 2000s air handler units generally communicate with the controller over serial, and there are specific configurations to make it work. We almost had a very costly issue of needing to upgrade the controls in the one system but they ended up getting it going.

[c_o_n_v_e_x]

>I suspect this is what would be difficult to replace.

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.

[doubleg72]

You have no clue what you're talking about. You are correct in that low cost PLC hardware can replace relay and pneumatic controls, but HVAC systems aren't just basic relays and pneumatics. They are microcomputers that communicate with actual protocols. They is no hot cutover or simple handover like you say.

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.

[c_o_n_v_e_x]

Easy, tiger. I was talking about upgrading controls on old equipment, not new equipment.

>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.

[quickthrowman]

> You have no clue what you're talking about. You are correct in that low cost PLC hardware can replace relay and pneumatic controls, but HVAC systems aren't just basic relays and pneumatics.

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.

[guessbest]

I thought it was replaced. The developer posted on discus, but I found the comment on slashdot quoted.

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.

[KerrAvon]

> 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.

Ha, bullshit. Someone decided they liked the Amiga and wanted to play with it. I am sympathetic!

[a-dub]

no that's actually true. in the mid 80s, it was the only personal computer that actually had preemptive multitasking built into the operating system.

serial communications were interrupt driven which didn't lend itself well at all to cooperative multitasking schemes. (gotta service those buffers asap)

[jhallenworld]

Maybe replace it with LoRa... but what device is in each building? I doubt you have to update the heating/cooling systems, they are likely just relay closures.

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..

[doobiedowner]

>they are likely just relay closures.

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.

[paulmd]

Schools are often very challenging RF environments, my uni wifi was constantly overloaded as fuck, even the newly built engineering campus was not great. And they ran a decent amount of APs, but there was also a fuckload of people there on wifi all the time, because lol engineering campus.

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.

[TeMPOraL]

> Or Samsung "Smart Things".. but unlikely to last 30 years..

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.

[jhallenworld]

Well that's what I figure: you buy their thermostat. We don't use ours for hvac control, just water leak monitoring.

Can you imagine what Samsung (or any large corporation) would want for a centralized controller for an entire school district?

[whalesalad]

What makes you say that? A software defined radio could do this pretty easily.

[bcrl]

It's pretty much guaranteed that any software defined radio you build today will be almost impossible to get parts for in 10 years, so your maintenance budget ends up requiring a software developer to upgrade things when parts become obsolete. At least with computers from the 1980s, the hardware is fully understood and can be much more easily repaired. There are already projects that have produced modern replacements for custom chips in machines like the Commodore 64, and there is at least 1 underway for the Amiga's chipset.

[uni_rule]

At that point keep everything else intact and just replace the Amiga there with a Mr. FPGA running the same thing on hardware emulation.

[cobaltoxide]

Reminds me of: "We do these things not because they are easy but because we thought they would be easy."

[uni_rule]

I don't believe one that transmits is anywhere near as cheap as one that recieves.