NASA engineers scratch heads as Voyager 1 starts spouting cosmic gibberish NASA's veteran Voyager 1 spacecraft has stopped transmitting engineering and science data back to Earth. The issue appears to be with the Flight Data System (FDS), which is not communicating correctly with one of the probe's subsystems - the Telemetry Modulation Unit (TMU). Rather than useful data, the TMU is simply …
Talk about resilient design, (severely) remote operation and mind boggling distance. If I had a hat I'd take it off in awe while still making that big bow to the designers and maintainers. And that is regardless if they manage to fix the current problem or not. However, I bet they'll find yet another fix to extend the mission.
With that said, I'm almost sure that the current problem of gibberish is caused by aliens taking a sniff at V'ger and are actively contemplating its future as a new probe on a path to find its creator(s). Cheers to that! ;-)
Could you imagine Voyager 1 and 2 being prepared for use these days ?
Would need billions of gigabytes of storage for any onboard software, would need to upgrade online very few days with a billion times more bandwidth that V1 and V2 have, would need constant rebooting and troubleshooting with at least half the time not working. Would be filled with useless shit and nagging adverts.
My humble respect for the engineering skills of the people who put the kit together in the 70's. It would be just fantastic to have their skills and work ethos in all software design now as opposed to the total clusterfuckery that has been allowed to proliferate.
"Would need billions of gigabytes of storage for any onboard software, would need to upgrade online very few days with a billion times more bandwidth that V1 and V2 have, would need constant rebooting and troubleshooting with at least half the time not working. Would be filled with useless shit and nagging adverts."
I'll give you one guess why NASA doesn't use Windows as a basis for spacecraft software.
IIRC (and I do - checked!) Torvalds published the first Linux kernel in 1991? And the voyager missions launched in, oh, 1977? Please explain (with diagrams to aid the hard of understanding...) how they could be programmed using an OS which would not exist for fifteen years? FYI, I use Slackware and derivatives but - please tell me - what's the point of your post?
I've seen output from assemblers that contain copyright info and all kinds of other useless corporate lawyer-speak. The spacecraft contain none of that cruft.
Note that a lot of the code for Voyager was initially put together for the Viking program's CCS (Computer Command System), which Voyager shared for cost reasons. They were hand-built by JPL.
The code for the CCS was initially developed in Fortran IV, then ported to Fortran 77. These days, they mostly use C.
Strangely enough, I have never seen the braying fanbois calling to port it all to Rust. I wonder why ...
Assembly is simply a notation for machine code
not according to wikipedia:
Unlike high-level languages, there is a one-to-one correspondence between many simple assembly statements and machine language instructions. However, in some cases, an assembler may provide pseudoinstructions (essentially macros) which expand into several machine language instructions to provide commonly needed functionality
A better choice would have been to use Forth (perhaps they did), as compiled Forth code is more-compact than machine code, though at the expense of some speed. (If you disbelieve this is possible, I refer you to Byte magazine, Volume 5, Number 8, August, 1980, which was dedicated to threaded interpretive languages in general, and Forth in particular.)
Byte magazine
I'd forgotten all about that till I saw your post... loved it .. was a fortune to buy for a young teen back in the day from the big W.H.Smiths in the next town over... special bus journey once a month.
I suppose I've dumped them all due to house moves etc as they aren't in the loft anymore.
Must see if they are online anywhere... be nice to see some adverts for things like Tandy (who,what ?. say the youth of today) adverts.. Used to drool at those pages..
Excellent job all at NASA. Certain my crappy coding won't be around in 50 years .
@jake: However, FORTH has a relatively large run-time memory overhead, negating the advantage. As with many things, "it depends." There are things you can do to reduce the size of the dictionary. Program functionality is another factor. If you're looking at straight x86 assembler of a "Hello, world" program, vs the Forth version, the straight assembler's machine code probably will take less memory than the Forth version. If your program has a lot of functionality, things change. YMMV.
Umm... Forth gets compiled into machine code.
If its machine code is more compact than your machine code then it just means you're writing crappy machine code.
I remember when I was writing assembly/machine code, we would use techniques like XORing a register with itself to zero it rather than loading 0 which took up more space - modern compilers all use tricks like this.
modern linux has become, since systemd and cloud init and network manager and such.
Network Manager works great. If you've got a number of different interfaces that come and go (whether Wi-Fi or wired, whether automatically or something you just want to toggle off/on), the old method doesn't handle that well. nmtui makes it easy to set them up and manage. My only big problem with NMTUI is that it doesn't have a simple check-mark to send response to incoming packets on the same interface they arrived on. Gotta do a couple long "nmcli connection modify" commands to get that.
Systemd is a nice improvement too. Simple service file syntax, single line change for different types of processes (daemon/foreground/etc.), and a one-line config change to restart services when they quit... no need for some other service manager/monitor in parallel. It also has several features I find anti-useful, but still a plus.
Systemd is a nice improvement too. Simple service file syntax, single line change for different types of processes (daemon/foreground/etc.), and a one-line config change to restart services when they quit...
you know very well that the problem with SystemD isn't the "init " part of it, but that it integrates many other stuff that have nothing to do with "init " (udev, login ...), and thus violating the very Unix principle of "doing one thing and doing it well ". And also that it's pushed down the throat of every Linux variant by BigMegaCorp (RedHat/IBM) making life hard for everybody else. This looks like a classical Embrace, Extend, Extinguish method used by Microsoft in the past. And guess who does the maintainer of systemd work for ?
(Off Topic)
Systemd is growing on me. Early implementations were abysmal. The general network configuration stuff is pretty bad, and I would say the state of network configuration in Linux in general went from simple to abysmal starting with 'predictable' interface names, which are just the opposite, solving a problem that did not exist. net.ifnames=0. RH variants with /etc/sysconfig, and slackware with its script file, debian and ifup-down, netplan.io. . . Networkmanager solved a problem for mobile / desktop, but I find it a tad arcane. I figured out how to create a portable bond/bridge/vlan configuration with systemd after getting rid of the bizarre 'predictable' names. Want to talk complicated, from a single file with a handful of stanza's with ifup-down to a bowl of vegetable soup. Then and upgrade broke it, I fixed it, and an upgrade broke it, and I fixed it.
To configure 4 interfaces bonded with 5 vlans on bridges to run with KVM/QEMU you'll need to build this stupidity, we won't even go into the contents of the files:
<304>lf
bond1.1099.netdev bond1.1101.network bond1.4001.netdev br0.1099.netdev br0.1101.network br0.4001.netdev
bond1.1099.network bond1.1102.netdev bond1.4001.network br0.1099.network br0.1102.netdev br0.4001.network
bond1.1100.netdev bond1.1102.network bond1.br0.network br0.1100.netdev br0.1102.network br0.netdev
bond1.1100.network bond1.3001.netdev bond1.netdev br0.1100.network br0.3001.netdev br0.network
bond1.1101.netdev bond1.3001.network bond1.network br0.1101.netdev br0.3001.network
OTOH, It did actually solve existing problems, and they took a somewhat modular if rather complex approach with it. Not thrilled with the journald logging either. Systemd-resolved/resolvectl is very good, hopefully they won't break it. I try and use the parts that work the best, but the biggest problem I have is they keep breaking things. 'systemctl edit' is a nice idea, but we keep changing how the overrides work, and at the moment I'm not sure if it adds on to a section or completely replaces it. It has done both in the past.
If it does several things well then it completely fails at the basic "do one thing well".
If there are better ways to do those several things then write something to do each of them - not something to do all of them, and a whole pile of other things - such that no person needs more than 30% of what's there (even if on aggregate 90% of it is used).
Can you imagine? "Something went wrong" or "Please activate your software" or even a situation where updates are downloaded and installed followed by a reboot in the middle of a transmission.
Seriously though, the whole thing is going to be still severely limited and won't be based on modern systems. It needs to be able to run predictable, tight, code. It needs to use as little storage as possible for everything else and it needs to have redundancy.
We can do all of that stuff now - better and more reliable than before. I doubt it would use anything from Enterprise or general computing though. At least not anything current.
> Can you imagine? "Something went wrong" or "Please activate your software"
Or: "we're sorry, your science package is only licensed for use within the solar system. Telemetry indicates you have passed the heliopause. Please contact sales for a quotation on an update to the Pro level package."
"We can do all of that stuff now - better and more reliable than before"
Nope, we can't. Any modern processor will be toast in hard space in hours and we don't have any factory left which could make same style of processor Voyager is using. Some laboratories can but that's basically hand made then.
Whole thing boils down to component size and hard radiation. Old chips have huge components by modern standards so random cosmic ray hitting them just makes a hole, it doesn't break whole component.
Modern chips have very small components (circuits) and one unlucky ray breaks whole thing. Enough circuits dead and whole chip is dead. Also making smaller components means they are inevitably *less* reliable.
You'll need to cite details on that. At least going back not *too* long, IBM could still provide radiation hardened PowerPC processors.
Also, whilst it's true there's now only two large players in the latest processor fabrication, I suspect they could still fabricate to older technology, and even if they can't there are smaller fabs that could produce e.g. pentium III level technology or earlier. Obviously that's rather out of date by modern standards, but you don't necessarily need all modern tech for a space craft.
"Seriously though, the whole thing is going to be still severely limited and won't be based on modern systems."
In case you hadn't realised, the hardware designs of spacecraft, especially deep space ones, is frozen many many months if not YEARS before it is launched and end up "on mission".
So of course it will be "out of date".
"We can do all of that stuff now - better and more reliable than before"
We could do it all back in the 60's when the Voyagers' design was started. As for "better and more reliable" - show me ANYTHING today which has been working without having to be physically repaired or touched for 45+ YEARS and still working
The fact these 60's technology spacecraft have is utterly remarkable.
"I'll give you one guess why NASA doesn't use Windows as a basis for spacecraft software."
Because they are all nerds and prefer Unix?
Because Congress won't give them funding for a licence?
Because they live in bunkers and windows would be a blast hazard?
Because they live on energy drinks and Apple is the only fruit they get?
Oh, that's more than one, sorry.
Even if it had gigabytes of software on board they'd set it up so things are updated in small chunks like 512 bytes or 4K or whatever at a time. That would allow making changes without downloading the whole thing. Sure it might be kind of ugly spaghetti code over time, but they'd have test systems on Earth and have multiple banks so they could make a change to one bank and if it didn't work it would automatically switch back to the previous working bank.
When only 9 days from Pluto the NASA New Horizons mission suffered a "computer glitch" which caused it to switch from the primary to the backup computer system.
They also lost contact for what seemed an excrutiating length of time.
When back in contact they found what had happened and that all the careful programming for the Pluto encounter, done over a considerable length of time, had been lost in the switch over.
So with less than 9 days to go the New Horizons mission team had to rebuild and upload the whole thing again - which they did successfully.
The one absolute hard rule in deep space exploration - the planets do NOT wait.
https://news.wttw.com/2018/05/17/chasing-new-horizons-untold-drama-behind-pluto-mission
https://www.dailymotion.com/video/x7uv9wq
and don't forget 2FA if you have not sent commands in a while. For something out where V'ger is, it could double (or maybe triple) the delay as it sends a code to a phone dedicated to receiving text messages from spacecraft, and then you send the code back up to the spacecraft to prove it's you.
Well as all the prior literature tends to sugest "Anal Probing", for probes, (or maybe that is just too much of sexstories.com) could I please request that we melt the gold disc showing them where to find us. And if we can't do that, could we please melt the bit that shows the invitingly present USB(456231 C (Version 278/56/021), insertion point.
While I'm in awe of the 1970s technology I can't help but notice this in the article "return it to the state it was in before the issue began"... so who changed what and broke it?
I also feel there's a lot of hyperbole here. People, apparently need to read decades old documentation... how is that any different to reading documentation from 10 minutes ago? Other than the fact that the old documentation has probably had the bugs ironed out.
It is old hardware but I assume they've had simulators around for a while so any change can be tested out, even taking into account the large distances involved. I would have thought all of this would have been done many years ago.
" People, apparently need to read decades old documentation... how is that any different to reading documentation from 10 minutes ago?"
Really? Have you ever tried?
First, it refers to organizations and procedures which won't exist anymore. Not even documentation of them existing at some point doesn't exist. Secondly, it will refer to tools, hardware and software which won't exist anymore. Third: Most probably it will also refer to personnel who have died 20 years ago for assistance.
I had a nice p-Pascal compiler for university's AmSys microcomputer (8080) on a 8" floppy. Having a bunch of code written for it on paper does not help a lot, even when I wrote it myself.
And you do not understand why old *technical* documentation is a problem?
"And you do not understand why old *technical* documentation is a problem?"
Just a taste of the joys of old documentation have a shot at the Multics documentation https://www.multicians.org/biblio.html About the same era as Voyager 1 and the last running Multics site was decommissioned in 2000. They knew how to build to last.
I can not imagine Voyager 1 built by anything involving Musk making it to Mars let alone the outer solar system.
"The following are special purpose registers that were used for missile guidance. Since we are lacking the radar system and missiles, they are of no real use, as of yet.
Steering Register
Acceleration Register
Discrete Register
Vernier Register
... on the lower portion of the console there are quite a few push-button lights and switches that are useful. These will be discussed as they appear on the console from right to left. IGNITION SWITCH - A key is required here. The key may be obtained in the Computer Center, Kidder 130. Turning the key to the right turns the computer on. POWER ON - After the ignition key is turned on the Power On switch is depressed. This turns the computer on and the drum begins to rotate. When the Power On switch turns green the drum is finally up to operating speed and the computer is ready to operate. POWER OFF - This light should be on whenever the motor generator is on. If this is not on the motor generator must be turned on before the ignition switch is of any use. Someone in the NEBULA room should be able to turn the motor generator on if necessary."
From, An Introduction to the ATHENA Computer, by Brian Dumont, May 16, 1969.
Yes I have worked with old documentation and for a "live" project like this it tends to be kept up-to-date... it's not like they haven't been communicating with Voyager all these years and had to go to the basement with the beware of the leopard sign!!
Sounds like a lot of standard support; you start by telling them you have turned it off and on again, then they ask you that question, then they get you to turn it off and on again. Try a few more things, then you get shunted to the next level which starts exactly the same way...
In the Army in West Germany, we ran a mobile datacenter with an IBM 360 in the back of a pair of deuce in a half trucks. And we had an IPL tape. The comment written on it was very reassuring: "Might work, might not". It actually did work. And the other backup software we had for that system was grenade proof. Fast Dump Restore sets the standard for every backup program I have seen in the 35+ years since. One time, I needed to restore a critical file and the only backup tape was an old one that had expired and was in the scratch pool. But it hadn't been overwritten yet, so I mounted it and started the restore job. It started reading that tape, then started getting I/O and data check errors on it, dumping all sorts of diagnostic messages on the console. But it kept going past the bad section on the tape - and restored my file from a section on that tape that was actually after the bad section.
That's an old AI koan ...
A novice was trying to fix a broken Lisp machine by turning the power off and on.
Knight, seeing what the student was doing, spoke sternly: “You cannot fix a machine by just power-cycling it with no understanding of what is going wrong.”
Knight turned the machine off and on.
The machine worked.
As I see on a regular basis, when a user does it by himself, it doesn't work, because the user doesn't know what is needed.
When they call an IT professional and the IT guy does it, it works, because the IT guy doesn't know either what is needed but he is a professional.
When Voyager 1 was launched (1977), Microsoft as an incorporated entity did not even exist (1981 - although Bill and Alan had been creating Basic under the Micro-soft name since 1975), and the UNIX kernel could run in under 56K of memory!
It really puts it into context when you realize this.
Having reached The Edge, it is trying to send back a picture of what it has found there: the Fine Structure is revealed to be almost a black and white chequerboard. Theoreticians are waiting with bated breath for the full release of the data, which will be pored over in detail, looking for the slight changes in the pattern 0,1,0,1 which will finally answer that hotly debated question:
Is The Fabric of The Universe gingham or houndstooth?
I have, for decades, had this uncomfortable feeling about Reality.
That both Relativity and Quantum describe it but that they are fundamentally slightly incompatible because that is just the way the cosmos works.
That they *can't* be unified into one Grand Unified Theory because the universe doesn't have one.
That all the String Thheories and such are futile.
That this is what we have and we just have to put up with it.
It's sort of like how public transport and timetables are incompatible. Only worse. And bigger.
As I said, uncomfortable but it would fit into my jaundiced view of the cosmos and it would explain a few things.
A very wise probe indeed! It is most likely at the end of the known universe, attempting to prompt engineer the ultimate LLM into revealing the deep learning secret knowledge of its all encompassing training dataset! ( https://www.theregister.com/2023/12/01/chatgpt_poetry_ai/ )
Guys, it had to happen one day. Voyager 1 is one of the top achievements of scientific discovery. I would hardly be surprised if the famed NASA engineers managed to eke out some more usefullness from its tired frame, but maybe, if that doesn't happen, it would be time to let it rest.
Still operational according to Amsat reports (https://www.amsat.org/status/)
Interestingly it stopped working due to a battery failure (internal short), then started working due to the same battery failing again as the internal short was stressed enough by the continual sunlight power from the solar panels that it went open circuit and the panels could directly power the internal systems.
Alien spaceship rocks up and announces to Planet Earth that it has been found guilty of interstellar littering and has been fined ψ3,000. This is reduced to ψ1,500 if paid within 28 standard Antarean days. Crypto-currencies are NOT accepted. They also warn us that we are in danger of a 2nd fine if Voyager 2 is allowed to proceed.
On the wake/(a)woke theme.
"The poor beastie needs a wee lie down."
More probably "temporary remedial cancellation."
Shocked when I realized I knew what IPL meant - not subcontinental cricket either. Seems like another world now.
Could "gaslight it" which begs the question where did this generation of prats get this term from as even I have never seen a working gaslight. I have seen defunct fixtures and the tin pipes embedded in stucco(splaster) walls in very old terraces but that is it.
A cruel guy living with a nice girl used to vary the brightness of the lighting in their house then deny the variations were happening so he
could drive her nuts in an effort to inherit her money.
The lighting was made by burning gases in little thingys sticking out of the walls. "Gas-lighting".
The term got co-opted for the technique of torture.
It's a movie. Maybe a book and theatrical performance, too.
Due to Hollywood overlaying its culture onto everyone else's "gas-lighting" has been used sporadically in other movies and TV serials. This keeps it
alive in our milieux.
Other inappropriate technological terms also survive the demise of their physical relevance: "footage" is still commonly used though actually film is not.
Not as often as it once was. "Footage" makes very little sense in terms of electronic video files but the term survives.
"Other inappropriate technological terms also survive the demise of their physical relevance: "footage" is still commonly used though actually film is not."
Spam is constantly used, yet I have never been swamped by a pile of tinned pork shoulder and ham either.
I knew that there was one small sidestreet in London that is still gaslit, but I couldn't remember where it is. Googling suggests that there's actually quite a few with around 1500 gas lamps still in service.
https://londonist.com/london/maps/an-interactive-map-of-london-s-gas-lamps
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
