Early Soyuz Spacecraft Had a Peculiar User Interface, Says Developer Charles Simonyi (ieee.org) 104
Tekla Perry writes: When WYSIWIG pioneer Charles Simonyi went to space, he couldn't but help notice the awkward user interface on the rocket's control panel. It was a case of legacy systems, not wanting to change training and documentation, and an emulator that ran Unix on a 386 chip, he reported during a recent discussion on space software held at the Computer History Museum. "They liked the older chips because of radiation resistance and the feature set," he pointed out, noting how operation of the virtual interface was trickier than it seemed. "There are rows and columns," he said, "and you move the cursor over the button and use another button to push the virtual button."
"On the right side," he said, "there are these windows that are numbers you type in by pushing virtual buttons below them. You use the cursor keys to go to the virtual buttons then push an entry button that is virtual." He added: "You can see that even as the technology changes, they want to keep as many things the same as possible."
"On the right side," he said, "there are these windows that are numbers you type in by pushing virtual buttons below them. You use the cursor keys to go to the virtual buttons then push an entry button that is virtual." He added: "You can see that even as the technology changes, they want to keep as many things the same as possible."
In Soviet Russia... (Score:4, Funny)
Re:In Soviet Russia... (Score:4, Interesting)
buttons push you!
You jest, but actually, I remember watching Nova back in the 70's, and the show stated the Soyuz could be remotely controlled from the ground station.
They didn't want any cosmonauts landing the craft in the West.
Re: In Soviet Russia... (Score:2, Funny)
Socialism... people will do anything to escape its poverty.
Re: (Score:2)
You mean to imply that Stalin didn't base his ideology off Marx? Communism is socialism without class distinctions.
Re: (Score:2)
Communism is Socialism without Bananas.
Re: (Score:2)
Marx stated that in "Capitalism", that capitalism is the "transition" state to communism.
Hence, the credit system accelerates the material development of the productive forces and the establishment of the world-market. It is the historical mission of the capitalist system of production to raise these material foundations of the new mode of production to a certain degree of perfection.
Recall my earlier comment:
https://slashdot.org/comments.... [slashdot.org]
From 193x, Marxists (from Socialist party of Great Britain) attacked Stalin, criticized Soviet system under Stalin as "state-capitalism" aka "socialism", and predicted its collapse: http://www.worldsocialism.org/... [worldsocialism.org] http://www.worldsocialism.org/... [worldsocialism.org]
Re: (Score:2)
This comment has been brought to you by the Real socialism has never been tried [kym-cdn.com] club.
Re: (Score:1)
real capitalism was tried and it failed too
so what's the answer
Re: (Score:2)
Socialism... people will do anything to escape its poverty.
I think you mean authoritarianism.
Re: (Score:3)
From a comment made about the Russian space shuttle, they had a policy that one entire flight had to be done without a human before it was deemed safe. Their shuttle (if I recall correctly which I may not) had one unmanned trip before it was cancelled.
So... probably the reason it was able to be fully controlled remotely wasn't because of some random fears but due to this policy.
I believe a rather quick search on YouTube could find the video about the Russian space shuttle. It might be in The Curious Dro
Re: (Score:2)
Re: (Score:2)
You jest, but actually, I remember watching Nova back in the 70's, and the show stated the Soyuz could be remotely controlled from the ground station. They didn't want any cosmonauts landing the craft in the West.
First part is right, second part is either tongue-in-cheek or nonsense. Both the US and Soviet Union recognised that computers were much, much better at flying spacecraft than even the best test pilots (= astronauts). The Soviet Union told their pilots that that's the way it was going to go and just deal with it, the US had endless battles to allow computer control, down to creating systems that allowed pseudo-astronaut control while doing most of the work electronically. Read the book "Digital Apollo" f
Re: (Score:2)
They didn't want any cosmonauts landing the craft in the West.
Name a cosmonaut who escaped to the West?
Remember, cosmonauts have been strictly selected, are patriots, were well treated...
Early Soyuz? (Score:5, Informative)
I don't think they had 386 chips or Unix in the 1960s.
Re:Early Soyuz? (Score:4, Informative)
I don't think they had 386 chips or Unix in the 1960s.
He went to space in 2007, when Soyuz spacecraft had been in use for 41 years.
Re: (Score:3)
Re: (Score:3)
They're not saying it was a 386 running Unix in the 1960s. It was a 386 computer running Unix now that was emulating the 1960s interface.
Re: (Score:2)
Re: (Score:2)
Re:Early Soyuz? (Score:5, Informative)
TFA states
According to Simonyi, it was, “a simulation of an earlier spacecraft that had physical buttons, labeled exactly the same. They wanted to keep the training and all the documentation the same, so they created an emulator that runs on Unix, on a 386 chip.”
Thus, in the mid-2000, the Unix running on a 386 was simulating an operating panel with physical buttons as used in the earliest Soyuz modules of the 1960ies.
Re: (Score:2)
Re: (Score:2)
Re: (Score:3)
Re: (Score:2)
What benefit does that provide over having physical switches?
Re: (Score:3)
First: Weight and wiring complexity. Instead of having that bundle of wires all terminated at the switch panel, you could set computer controlled switch boxes closer to the equipment in question, saving weight and making the wiring a whole lot simpler.
Second: You can provide computer oversight of operator input and disable specific functions based on the phase of flight you are in, where a manual switch panel would need some kind of fail safe, "are you sure" physical barriers for specific features. Yo
Re: (Score:2)
I don't really buy the stuff about wiring and computer oversight. You can easily make a panel controller that only needs a simple serial link to the computer, and probably takes up less space and bulk than a screen. Disabling functions is via a simple lighting system. And in exchange you get all the benefits of a tactile interface.
The third thing, the flexibility, makes more sense. Also the potential for translating the interface into other languages, since a lot of the passengers are not Russian speakers.
Re:Early Soyuz? (Score:4, Interesting)
Most modern airliners have glass cockpits. Can't tell you why it's better, but UI research is taken very seriously by the FAA, all part of keeping the cognitive load during normal flight minimized so that pilots have the most to work with in an emergency.
Fun fact: Southwest Airlines uses all 737s, so all their pilots have the same type rating. Part of their fleet has the legacy physical cockpit (which still has plenty of screens and menu-driven UI), and part has glass cockpits. They got Boeing to make them custom glass cockpits that look as much like the physical layout as possible to save on training costs. It's a more awkward UI because simulating physical controls on a screen is generally a bad plan (it's the sort of mistake "designers" with no job experience make), but it does help new pilots switch back and forth easily.
Actual Soyuz crews make extensive use of tablets alongside the clunky capsule controls. The new Boeing capsule still has physical controls to make older astronauts feel at home. The new SpaceX capsule is entirely glass cockpit.
It's worth pointing out out that the control a pilot uses the most often, whether airline or spacecraft is the checklist, which is main reason Soyuz crews hold tablets - you can cram 100 pages of checklist into a small space easily used with mittens, and that won't lose the page you're on if you let go of it in 0 g. (The center of the yoke on an old-style 737 is a checklist-of-checklists, with a physical slider to keep track of which checklist you're on.)
Re: (Score:2)
Thus, in the mid-2000, the Unix running on a 386 was simulating an operating panel with physical buttons as used in the earliest Soyuz modules of the 1960ies.
That still doesn't make sense since that couldn't have been a 386 *in the Soyuz*. No Soyuz has even had an x86 chip anywhere. Until 2010, it had an Argon 16 [computer-museum.ru].
Re: Early Soyuz? (Score:2)
Soyuz has multiple flight and non flight related computer. UI, as the hearsay says, was at some point ran on a pirated BSD clone. And the flight computer that did actual orientation and propulsion control was just a bunch of analog integrators + Argon 16 for autopilot.
Re: (Score:3)
Re: (Score:1)
* a strange mix of Not Invented Here;
* a big need to copy Western technologies, either because their own tech actaully was sub-par, or was deemed inferior by the higher-uppers;
* and the deficit (aka shortages) of everything, even down to the components.
The shortage situation were simply known as 'the deficit'. While it wasn't as appalling as the shortages in Venezuela, there were any number things that were always in short supply, highly expensiv
Re: (Score:2)
According to Simonyi, it was, “a simulation of an earlier spacecraft that had physical buttons, labeled exactly the same. They wanted to keep the training and all the documentation the same...
And the new craft was called the Soyuz Max.
No changes? (Score:2, Insightful)
I think if the interface changes so that you don't have the old interface, but have to use cursor and push another button to push a virtual button, then the interface has changed hell of a lot. It definately is not the same interface anymore.
Re: (Score:2)
True in the pedantic sense, but not in the practical sense. The new interface has the same buttons, and the same functions as the old one. I.e. you can reuse all of your training material.
You're just using a different method to press those buttons.
Re: (Score:3)
"and you move the cursor over the button and use another button to push the virtual button."
So ... basically every button press has an implicit confirmation step built into it?
I dunno. To me that doesn't sound like a bad thing in a spacecraft.
Not everything needs to have animated GIFs (or be able to be activated if a pen happens to float into a touchscren).
Re:No changes? (Score:5, Insightful)
but have to use cursor and push another button to push a virtual button,
To push a "button" on a GUI application on your desktop, you move your mouse to position a cursor over a virtual button, and then push a button on your mouse to push the virtual button.
As described, this interface doesn't sound that abnormal.
Would the UI be very appropriate in space? (Score:3)
I'm guessing that Simonyi was learning the UI on the ground. I would agree that Simonyi is an expert in UI on Earth - but in space?
I'd be interested in hearing his (and others) comments about how good the UI is in space. I would think that there are requirements and use cases which would make something that works well in 1g without any life-critical functions completely inappropriate in 0g and dealing with life and mission (which is also life) critical.
Re:Would the UI be very appropriate in space? (Score:5, Interesting)
I would think that there are requirements and use cases which would make something that works well in 1g without any life-critical functions completely inappropriate in 0g and dealing with life and mission (which is also life) critical.
That's a very good point. You would not want to be able to miskey operations, and deliberate selection followed by activation by a physical button seems like a good way to avoid this.
Re: (Score:2)
It sounds like he had to move a cursor around using arrow keys or something like that, i.e. a very time-consuming method of selecting a button. I get that this interface predates modern touch screens, but this solution seems to be worse than the button panel it's trying to emulate.
Re: (Score:2)
It sounds like he had to move a cursor around using arrow keys or something like that, i.e. a very time-consuming method of selecting a button.
OTOH it has an implicit confirmation step in each button press. That might be a very good thing in a spaceship (depending on what the functions are)
Re: (Score:2)
Confirmation is good, taking 10 keypresses to emulate 1 button press is not.
Re: (Score:2)
Re: (Score:2)
Yes, and they're awful too. It takes forever to do anything that doesn't have a dedicated physical button. I'm using a Mac as a media server, so I can compare between that and e.g. the crappy STB provided by my ISP. A computer with mouse and keyboard is light years ahead of anything that relies on up/down/left/right for navigation.
Re:Would the UI be very appropriate in space? (Score:5, Interesting)
It's not about the UI, it's about training.
NASA (and everyone else "old school") has a MASSIVE series of manual, after manual, after manual.
Every scenario worked out, like "push these 17 buttons in THIS order if THIS specific thing happens". Especially during the moon shot eras, that knowledge was spread between hundreds, literally hundreds of companies which supplied NASA with parts/infrastructure too. During catastrophes/major errors, those company's experts might be called in.
And this stuff was tested, and tested, and tested.
Then there's the flight simulators, the books, the astronauts that have flown 20 times, etc etc.
It sounds to me like they want to upgrade the hardware, but can't quite upgrade all that legacy stuff yet. And I can see the advantage.. making the hardware switches/control UI virtual, gives loads of advantages. Like improved ground control, etc.
But can you imagine all that stuff being re-written? All that procedure replaced?
Even today, space is very, very unforgiving. You push one button, it's "you're fucked", with no remote chance of rescue.
Re:Would the UI be very appropriate in space? (Score:4, Informative)
This ^
Space-grade interfaces followed the same procedures as flight cockpits, upped to the max. The Soviet Union had a tradition of its own for doing Interaction design, with very interesting concepts and systematic approaches, sometimes better than current Western UX practice - you can go down that rabbit hole if yo know Russian or find one of the few books translated to English.
One interesting outcome is a layout protocol for building the most readable flowcharts I've ever seen -now available in the open source Drakon [wikipedia.org] editor-, which was designed around making it readable to humans, not just being able to capture every possible logical relation.
Re: (Score:1)
But can you imagine all that stuff being re-written?
...at huge expense, as a 5-year project that's currently 15 years behind schedule, by the Russian equivalent of Lockheed-Martin.
The block 1 "Soyuz II" would have cutting-edge parachutes, but the software wouldn't be able to deploy them. Landing without killing everyone is a future enhancement, tentatively scheduled to come with the block 2 capsules in 2024.
Re: (Score:1)
I'm guessing that Simonyi was learning the UI on the ground. I would agree that Simonyi is an expert in UI on Earth - but in space?
I'd be interested in hearing his (and others) comments about how good the UI is in space. I would think that there are requirements and use cases which would make something that works well in 1g without any life-critical functions completely inappropriate in 0g and dealing with life and mission (which is also life) critical.
I would argue that he can't possibly be much of a UI expert anywhere in this galaxy if he doesn't realize the value of maintaining a standard interface. So many UI 'experts' try to validate their existence by constantly changing things in small and peculiar ways and to no benefit. No UI Is perfect the first time but once it has gained traction the cost of changing it is often higher than the cost of leaving it imperfect. This is especially true since no UI will ever be perfect. Mistakes will always be ma
Re: (Score:2)
If he had anything to do with the reveal codes in Word I would dispute your statement that he's an expert in UI on Earth statement.
Re: Would the UI be very appropriate in space? (Score:1)
I'm pretty sure he didn't. By that point committees of committees were 'designing' UIs at Microsoft.
Re: (Score:1)
Re: (Score:2)
It was in both. WordPerfect had codes that looked like XML and you could actually figure out what was going on with the document. There was a section at the bottom of the document that showed the the document with all of the tags while about 4/5's of the window remained for the document itself, though you could change size. You could also edit the document in either area.
In Word it made all of the invisible codes visible and that's about all it did. At least for me it didn't actually help to find out where
Buttons, Zero-G, and i386 (Score:5, Interesting)
Buttons work better than touch screens because you have immediate pre-press tactile feedback. Even our modern smartphones don't allow us to know sight unseen which button we're about to press. Our computer keyboards do! Our 1984 Princess telephone did. Not modern touchscreens. So accuracy while you're in motion is more important than touchscreens. Think of how many times you go "touch this" on a car's navigation system but instead you touch that other thing next to you. Space is deadly.
Second, between takeoff and landing the cosmonauts experience between 0 and ±5G (4.2 on the Soyuz descent Tuesday). That almost makes the previous issue moot. When your hand goes from weighing up to 10lbs to up to 50lbs... and back... you better hope you can anchor your fingers on something and hit THE RIGHT BUTTON and only THE RIGHT BUTTON.
Finally, while I know he's a wiz, you can't tell if it's "Unix on an i386" vs "Linux on an i686" vs Solaris vs Ultrix vs... the external interface provided to the Soyuz is SO LIMITED that unless there's a bootup messages screen (note: there is not) you can't tell. Also most i386 linux based systems will run on higher-end hardware... so just the software messages alone wouldn't really give you 100% certainty.
Still a good story. Soyuz launches have been consistent. I've yet to figure why sometimes they'll do the "short MIT route" and sometimes the 2-3 day route... but hey, I'm not Everyday Astronaut (h/t Tim). One day SpaceX will be ferrying astronauts to the ISS, and my fervent wish is that a)All their missions are successful and trouble free, and b)They sell seats to the Russians for more than $80M each.
Ehud
Tucson AZ US, home of "World View Enterprises", the place that will launch people into the stratosphere with big balloons. Not.
Re: (Score:2)
There are ways to do "touch" interfaces which work fine with gloves, fingers, tree branches or pretty much anything that's opaque that could be space rated.
Besides, I wonder if there are capacitive touch interfaces which would be space rated? Capacitive touch screens are certainly NOT used in aircraft avionics for a reason by the way.
Re: (Score:2)
It's not making a touchscreen work. It's that touchscreens are the wrong interface for space capsules.
Re: (Score:2)
It's not about buttons vs a touch screen. It's about simulating an older interface that used a grid of buttons with a picture of a grid of buttons on a screen. It does seem a bit odd that they couldn't wire up enough GPIO to the computer to use real buttons, but I'm sure they had their reasons.
Re: (Score:2)
Airline Pilot : Hey, this is not what I trained for.
Yeah but the plane is the same general shape so you don't need to bother learning how to fly this particular one. Now get up there and get this thing airborne.
Re: (Score:2)
Airline Pilot : Hey, this is not what I trained for.
Yeah but the plane is the same general shape so you don't need to bother learning how to fly this particular one. Now get up there and get this thing airborne.
It's even more perverse than that. Overseas they toss guys with less than 400 hours of total flying experience in the right seat after about 2 to 3 weeks of type training that includes maybe 20 hours of simulator time. They cannot possibly safely operate such an aircraft without the automation and computing resources being functional with such a low level of experience. The guy in the left seat is a bit better off, having maybe 600 hours... But still, when you have 300 people behind you is that enough? I
Re: (Score:3)
It's a lot more for part 121 now - 1500 hours minimum to warm the right seat of a regional jet in the US. "Real" airlines want a thousand hours of jet time on top of that.
Now, almost everyone gets there by instructing, so we can still debate whether 1500 hours of pattern work in a Cessna 172 is relevant to any other kind of flying...
Re: (Score:2)
Look at the Lion Air crew.. The right seat guy had between 300 and 400 hours and the captain had under 800. They where not in the USA operating part 121. This is what I was talking about, overseas, where the minimum flight experience is often considerably less, and where many want to be pilots from the USA go to get flight time, working hard hours, getting paid in chicken feed.
By the way, I believe the minimum experience is 500 hours for right seat (second in command/first officer) and 1500 for Left sea
newer, better... (Score:5, Funny)
OK guys, welcome to Astronaut school. Now I know some of you will be familiar with the old style UI from videos but i can say that we've been improving things since then, our team has been working real hard on better UIs for some time now and I know you'll all agree there have been leaps ahead in usability.
The latest UI is a SPA written using React and Angula and even has native javascript behind this clean, minimalist icon design that is entirely built using white and grey.
Re: (Score:3)
Thanks, I hate it.
Re: (Score:1)
You jest, but if you really wanted to do this, Svelte would be a better option, particularly if you didn't want to have much power requirements: it's used on over 200,000 simple POS terminals in Brazil...
Re: (Score:2)
It sounds like a similar interface to a Roku. Where you use virtual buttons and it's a pain in ass to type a search in.
WYSIWIG (Score:2, Funny)
What you see is what I get. It's a special kind of blind-folded interface delivery system....
Need to upgrade to Windows 8 (Score:2)
WYSIWIG (Score:2)
At some deep level, it's spectacularly humorous to misspell WYSIWIG, because half of the time—any good neurologist will confirm this—what you see isn't even what you have.
Reminds me of Netflix on my TV (Score:2)
So with the Soyuz interface, you has to position the cursor over a virtual button, then press an actual button to activate. This is very similar to typing in Netflix without an actual keyboard. I use arrow keys to position over a letter on a virtual keyboard, then press a button to type the letter.
It's pretty damned cumbersome. I really ought to look into some kind of wireless keyboard for my TV.