Solve for "if unknown: do this" and you've solved for the unsolvable
This is extremely ignorant and dangerous when it comes to a complex control system. A lot of people in this discussion are trying to trivialize the efforts involved in programming an automated car. You're also glossing over the capabilities of current sensor systems. Video processors, laser rangefinders, radars, GPS - they're all good and getting better every day, but nowhere near the capabilities of a person's senses of sight and sound yet.
The unexpected events I'm picturing are wildlife or people running into the road in a range too close for the car to stop. Objects falling off of that lawn care flatbed in front of you. Does the software choose to collide with the object? Initiate a risky swerve maneuver? These decisions depend on what the obstruction is and the sensors are simply not capable of identifying them at this point. What about a car in the slow lane occluding a car moving quicker in the fast lane?
These are extremely difficult problems to solve, and you won't see these systems going mainstream until there's a high confidence they're solved. That's why you see incremental steps. Today it's adaptive cruise control that might be capable of either warning or actually using the brakes if an imminent risk is detected. Tomorrow it may be lane holding or caravanning. Make no mistake, I'm not opposed to driverless cars - I'm just saying we're still a long way from replacing the driver.
Admittedly, I'm not a pilot. However, I'm more aware of automatic aircraft landing systems than you'd think;)
There are a couple key differences. First - the collision scenarios are all in a short window of time around takeoff and landing, where you can bet the pilots will be on their toes (plus they are aided by air traffic controllers). If they're on an automatic approach, they aren't going to be sipping coffee and reading a book. Second - pilots are trained professionals who will be far more aware of their surroundings and risks than the typical driver. Cars are already appliances, if we make them more automated we certainly can't count on drivers to keep up their already dismal awareness of what's going on around them.
Automatic cars really will be safer than human drivers some day - but not until they can fully take the human out of the equation. "Fail over to the human driver" won't work on the highway.
Yup, the failover is definitely going to be a big discussion point on this issue. As to programming for "relatively common scenarios" I suppose my point is that there are hundreds of factors at play which can change rapidly as a situation unfolds. There are situations where road conditions dictate whether you try to brake or apply power in an emergency situation. It's a very difficult series of problems to solve. We'll get there eventually, but as the author states, I think it's a bit of a ways off.
Yep, it's going to be a two-pronged process, and I'd put money on the legislative side lagging behind the technical side. I don't know if it would be as clear a stepping point legislatively though. I could definitely see the steps being test cars with active drivers then caravan automatic cars behind a driven pilot car then solo cars with driver ready then magic hover carts. Who knows though, the legislature works in ways that defy all reason.
Boats and planes aren't in a traffic pattern with dozens of others in close proximity. In a plane you don't need to anticipate that *any second now* that Red Bull stunt plane is going to cut you off and hit the brakes. You can count on the captain being able to assess when they should be in manual control of their vehicle and hold them responsible if they aren't. For a car, those situations can change faster than you can reasonably expect a person to assess what's going on and make the right decisions.
There are too many outside factors going into these "details" to be 100% solved. Driving is one of the most fluid and dynamic tasks people take part in. I'm not saying that self-driving cars aren't going to be as competent as a moderately skilled human driver, I'm just saying it's an awful difficult controls problem. Everything solved leads to more problems. People are talking like fully autonomous cars are right around the corner when really we're going to see hundreds of incremental steps before we get there.
You want to install a system in the car with the express purpose of taking the driver out of the equation, but still expect the driver to monitor the autopilot and take over in the event of a fault? I can't see that flying from a legal or legislative standpoint. You'd be asking the driver to take over at the worst possible time: something has happened that the autopilot can't handle and the driver has to take over. The driver who's been relaxing, reading, chatting, doing anything besides monitoring the conditions of the road. It's far too much to ask a driver to make the right decisions in that situation. It's also far too much to ask the legal system to find the driver at fault given what inattentive drivers are able to get away with today.
But you do need to deal with wildlife. I wonder what thought they've put into the programmatic decision making when an accident is unavoidable. Collide with that obstacle? Can it identify a deer from elk or moose? How will it handle swerving? I suspect we'll see cars cruising high-speed on freeways in caravans well before an arbitrary auto-pilot.
A driverless car will certainly be overall more attentive than a human driver, but it also needs to be able to handle the unexpected things a human driver handles. The mundane tasks, sure - but how do you handle things like a tire blowout in a curved section of road with sand on it? As long as there are relatively common scenarios that crop up that a human can handle some reasonable percent of the time that the software can't, it's not ready for prime time. How do you failover when road conditions exceed the thresholds of the car? The software can't simply say "deal with it" and have the driver take over. Driver could have their hands full of coffee and iPads, be sleeping, or otherwise unaware of the situation.
No - in 1999 I was working on a 19" monitor with 1600x1200 pixels and had a keyboard and mouse. "Real" photo editors used wacom tablets as well. I wouldn't go so far as to say it's impossible to create content on a small tablet, but you definitely are giving up a lot of useability when you drop down to the small form factor and lose the keyboard and precise pointing devices. System performance isn't so much an issue anymore, but the screen space and interfaces on the tablet simply don't cut it.
That simply isn't how the world works, though. Sure, you could theoretically create a common UI for computing devices (though the current Windows 8 example is... lackluster...), but that simply can't carry over to everything we use, nor should it. The user end of whatever you're interfacing needs to be derived from that device itself, not an adaptation of some "standard UI." Unless you need timers, a simple button or switch is ideal for that coffee pot. There's no sense forcing a big "unified" UI upon devices that are better served by tactile interfaces. Picture trying to operate a machine or an automobile with LCARS.
What I would like to see in windows is having a distinct zone to click on it instead of having to find the magic spot or hitting the windows key
Like maybe if they put a button down on the taskbar that you could click to open up a menu with all of the the start page functionality? They could label it "Start!"
In more seriousness - my problem with the start/metro page is it takes my entire screen for no purpose. If I'm working on something, I don't want to lose sight of it just to open a new program. Add to that the mouse-over "hot zones" and half-assed adoption of a touch-screen interface and you've got a frustrating environment for regular desktop users.
I've only used Windows 8 on my HTPC and Metro will eventually be pretty decent for that - but I still found it frustrating as hell to set up the first time and have zero interest in using it on my desktop.
For a computer that is used by users, the UI is everything. It's entire worth is your input to it to control it, and it's output to you in response. Everything starts and ends with that. There is nothing more important.
I strongly disagree with this assertion. There have been plenty of platforms that had excellent UIs for their time (Amiga, NeXT, Be) that failed. They didn't live or die by their UI, they lived and died by the ability of their users to do whatever tasks they had: the availability of applications. I have yet to see anything about the common Linux UIs that makes them better or worse than Apple or Microsoft (that is, until Metro came out on the desktop and bombed Windows 8 useability for me).
Now what of commercially successful software that had horrid UIs? MS Word 6 was a mess of buttons and toolbars. Windows 8 asks me to swipe from the side of my non-touch screen to bring up search and other options. If UI were everything we would have stopped using Word and moved to WordPerfect or another competitor back in the 90s. Even the pinnacle of easy to use design in this industry, Apple Computer, was in it's death throes until they came across their killer app: iPods and iPhones. Yet now even the iPhone is losing ground to an utter mishmash of in-congruent UIs from Android devices.
So no, I don't believe your "UI is everything" argument holds water. Price, functionality, and useability all come to play.
To be fair, Windows 8 as an OS has been pretty good, it's just the new Start screen and Metro UI that make it such a pain (seriously... the desktop UI asks me to swipe from the side for more options!?). If you have Windows 8 on a desktop, you can install Classic Shell and be perfectly happy. With any luck, Microsoft will come to their senses and incorporate that functionality natively sometime down the road.
Microsoft and Star Trek... Every other release is a turd.
You're correct in calling out the "smartphone derived" comment - under the hood Windows 8 is just as capable as it's predecessors. It does make me wonder how the Microsoft employees use metro on their desktops because I can't seem to figure out its intended use. As you've seemed to notice, it's just in the way for desktop use, and it certainly doesn't help that MS has gone out of their way to ensure it's the first thing we see when we power up our computers. I haven't tried customizing the win+x menu, but isn't that more or less just bringing back the classic Start menu?
Windows 8 as an OS looks to be just dandy, we're all just scratching our heads as to why MS decided to saddle it with the new start screen. It just feels like an attempt to close off my computer to me...
Boy do you have a bone to pick... The original post you started your random anti-Linux crusade on wasn't even talking about Linux, simply bemoaning the shift to mobile-like UIs in desktop operating systems.
But, since you've gone trucking down this road, would you like to discuss the real reasons why Linux isn't popular on the desktop? You're always talking the UI, but that can only be a tiny part of the picture. The UIs themselves are no more different than switching between Windows and MacOS. Canonical has made the install process for Ubuntu as seamless as a Windows install. A great deal of hardware works fine out of the box. Yes, using Linux can be more technically involved than Windows or MacOS. but what's been holding back Linux has largely been software availability. Businesses want Excel, Word, and Outlook. Creative developers want Photoshop and... whatever else creative developers want. Software developers are whores who will program for anything. They're also the majority of desktop Linux users. Folks at home want whatever they have at work. Availability of video games is also no small factor.
Linux's problem isn't so much useability - it's that it has never had a killer app in the desktop realm to generate a user base around.
Wile I agree with most of what you're saying, I don't agree with your conclusions. There are many things that are either better done on a PC or only able to be done on a PC such as software development tasks, editing of photos and videos. Mobile devices are designed around content consumption, not creation. They are definitely going to be grabbing a lot of lighter weight computer users and drive the PC market down to a fraction of its current size. It will certainly mean the end of the road for a number of PC manufacturers.
However, I simply don't see the PC market decline as being terminal. It will drop down to a new standard and stay there.
Oh come on now, anybody who knows anything knows the smoothest running Xeons are V12s or inline 6 cores.... Or boxers, but you know the first and second rules of boxer core club...
Slashdot Mirror
Solve for "if unknown: do this" and you've solved for the unsolvable
This is extremely ignorant and dangerous when it comes to a complex control system. A lot of people in this discussion are trying to trivialize the efforts involved in programming an automated car. You're also glossing over the capabilities of current sensor systems. Video processors, laser rangefinders, radars, GPS - they're all good and getting better every day, but nowhere near the capabilities of a person's senses of sight and sound yet.
The unexpected events I'm picturing are wildlife or people running into the road in a range too close for the car to stop. Objects falling off of that lawn care flatbed in front of you. Does the software choose to collide with the object? Initiate a risky swerve maneuver? These decisions depend on what the obstruction is and the sensors are simply not capable of identifying them at this point. What about a car in the slow lane occluding a car moving quicker in the fast lane?
These are extremely difficult problems to solve, and you won't see these systems going mainstream until there's a high confidence they're solved. That's why you see incremental steps. Today it's adaptive cruise control that might be capable of either warning or actually using the brakes if an imminent risk is detected. Tomorrow it may be lane holding or caravanning. Make no mistake, I'm not opposed to driverless cars - I'm just saying we're still a long way from replacing the driver.
Admittedly, I'm not a pilot. However, I'm more aware of automatic aircraft landing systems than you'd think ;)
There are a couple key differences. First - the collision scenarios are all in a short window of time around takeoff and landing, where you can bet the pilots will be on their toes (plus they are aided by air traffic controllers). If they're on an automatic approach, they aren't going to be sipping coffee and reading a book. Second - pilots are trained professionals who will be far more aware of their surroundings and risks than the typical driver. Cars are already appliances, if we make them more automated we certainly can't count on drivers to keep up their already dismal awareness of what's going on around them.
Automatic cars really will be safer than human drivers some day - but not until they can fully take the human out of the equation. "Fail over to the human driver" won't work on the highway.
Yup, the failover is definitely going to be a big discussion point on this issue. As to programming for "relatively common scenarios" I suppose my point is that there are hundreds of factors at play which can change rapidly as a situation unfolds. There are situations where road conditions dictate whether you try to brake or apply power in an emergency situation. It's a very difficult series of problems to solve. We'll get there eventually, but as the author states, I think it's a bit of a ways off.
Yep, it's going to be a two-pronged process, and I'd put money on the legislative side lagging behind the technical side. I don't know if it would be as clear a stepping point legislatively though. I could definitely see the steps being test cars with active drivers then caravan automatic cars behind a driven pilot car then solo cars with driver ready then magic hover carts. Who knows though, the legislature works in ways that defy all reason.
Boats and planes aren't in a traffic pattern with dozens of others in close proximity. In a plane you don't need to anticipate that *any second now* that Red Bull stunt plane is going to cut you off and hit the brakes. You can count on the captain being able to assess when they should be in manual control of their vehicle and hold them responsible if they aren't. For a car, those situations can change faster than you can reasonably expect a person to assess what's going on and make the right decisions.
There are too many outside factors going into these "details" to be 100% solved. Driving is one of the most fluid and dynamic tasks people take part in. I'm not saying that self-driving cars aren't going to be as competent as a moderately skilled human driver, I'm just saying it's an awful difficult controls problem. Everything solved leads to more problems. People are talking like fully autonomous cars are right around the corner when really we're going to see hundreds of incremental steps before we get there.
Automatic transmissions, cruise control, back-up sensors, parallel parking assist, adaptive cruise control, lane holding, vehicle following, obstacle avoidance. They're all iterative steps.
I love Sprinters. My sister's still has the Dodge logos all over the outside, but everything on the inside still has the Mercedes logo.
You want to install a system in the car with the express purpose of taking the driver out of the equation, but still expect the driver to monitor the autopilot and take over in the event of a fault? I can't see that flying from a legal or legislative standpoint. You'd be asking the driver to take over at the worst possible time: something has happened that the autopilot can't handle and the driver has to take over. The driver who's been relaxing, reading, chatting, doing anything besides monitoring the conditions of the road. It's far too much to ask a driver to make the right decisions in that situation. It's also far too much to ask the legal system to find the driver at fault given what inattentive drivers are able to get away with today.
But you do need to deal with wildlife. I wonder what thought they've put into the programmatic decision making when an accident is unavoidable. Collide with that obstacle? Can it identify a deer from elk or moose? How will it handle swerving? I suspect we'll see cars cruising high-speed on freeways in caravans well before an arbitrary auto-pilot.
A driverless car will certainly be overall more attentive than a human driver, but it also needs to be able to handle the unexpected things a human driver handles. The mundane tasks, sure - but how do you handle things like a tire blowout in a curved section of road with sand on it? As long as there are relatively common scenarios that crop up that a human can handle some reasonable percent of the time that the software can't, it's not ready for prime time. How do you failover when road conditions exceed the thresholds of the car? The software can't simply say "deal with it" and have the driver take over. Driver could have their hands full of coffee and iPads, be sleeping, or otherwise unaware of the situation.
No - in 1999 I was working on a 19" monitor with 1600x1200 pixels and had a keyboard and mouse. "Real" photo editors used wacom tablets as well. I wouldn't go so far as to say it's impossible to create content on a small tablet, but you definitely are giving up a lot of useability when you drop down to the small form factor and lose the keyboard and precise pointing devices. System performance isn't so much an issue anymore, but the screen space and interfaces on the tablet simply don't cut it.
That simply isn't how the world works, though. Sure, you could theoretically create a common UI for computing devices (though the current Windows 8 example is... lackluster...), but that simply can't carry over to everything we use, nor should it. The user end of whatever you're interfacing needs to be derived from that device itself, not an adaptation of some "standard UI." Unless you need timers, a simple button or switch is ideal for that coffee pot. There's no sense forcing a big "unified" UI upon devices that are better served by tactile interfaces. Picture trying to operate a machine or an automobile with LCARS.
What I would like to see in windows is having a distinct zone to click on it instead of having to find the magic spot or hitting the windows key
Like maybe if they put a button down on the taskbar that you could click to open up a menu with all of the the start page functionality? They could label it "Start!"
In more seriousness - my problem with the start/metro page is it takes my entire screen for no purpose. If I'm working on something, I don't want to lose sight of it just to open a new program. Add to that the mouse-over "hot zones" and half-assed adoption of a touch-screen interface and you've got a frustrating environment for regular desktop users.
I've only used Windows 8 on my HTPC and Metro will eventually be pretty decent for that - but I still found it frustrating as hell to set up the first time and have zero interest in using it on my desktop.
Man I love chocolate sundaes. What do you think, nuts or no nuts?
You're making the flawed argument that common sense had existed before going extinct...
I think you have your statistics backwards...
Remember: Never trust anyone with a user-id over a hundred thousand.
It hurts me that you say this.
For a computer that is used by users, the UI is everything. It's entire worth is your input to it to control it, and it's output to you in response. Everything starts and ends with that. There is nothing more important.
I strongly disagree with this assertion. There have been plenty of platforms that had excellent UIs for their time (Amiga, NeXT, Be) that failed. They didn't live or die by their UI, they lived and died by the ability of their users to do whatever tasks they had: the availability of applications. I have yet to see anything about the common Linux UIs that makes them better or worse than Apple or Microsoft (that is, until Metro came out on the desktop and bombed Windows 8 useability for me).
Now what of commercially successful software that had horrid UIs? MS Word 6 was a mess of buttons and toolbars. Windows 8 asks me to swipe from the side of my non-touch screen to bring up search and other options. If UI were everything we would have stopped using Word and moved to WordPerfect or another competitor back in the 90s. Even the pinnacle of easy to use design in this industry, Apple Computer, was in it's death throes until they came across their killer app: iPods and iPhones. Yet now even the iPhone is losing ground to an utter mishmash of in-congruent UIs from Android devices.
So no, I don't believe your "UI is everything" argument holds water. Price, functionality, and useability all come to play.
To be fair, Windows 8 as an OS has been pretty good, it's just the new Start screen and Metro UI that make it such a pain (seriously... the desktop UI asks me to swipe from the side for more options!?). If you have Windows 8 on a desktop, you can install Classic Shell and be perfectly happy. With any luck, Microsoft will come to their senses and incorporate that functionality natively sometime down the road.
Microsoft and Star Trek... Every other release is a turd.
You're correct in calling out the "smartphone derived" comment - under the hood Windows 8 is just as capable as it's predecessors. It does make me wonder how the Microsoft employees use metro on their desktops because I can't seem to figure out its intended use. As you've seemed to notice, it's just in the way for desktop use, and it certainly doesn't help that MS has gone out of their way to ensure it's the first thing we see when we power up our computers. I haven't tried customizing the win+x menu, but isn't that more or less just bringing back the classic Start menu?
Windows 8 as an OS looks to be just dandy, we're all just scratching our heads as to why MS decided to saddle it with the new start screen. It just feels like an attempt to close off my computer to me...
Boy do you have a bone to pick... The original post you started your random anti-Linux crusade on wasn't even talking about Linux, simply bemoaning the shift to mobile-like UIs in desktop operating systems.
But, since you've gone trucking down this road, would you like to discuss the real reasons why Linux isn't popular on the desktop? You're always talking the UI, but that can only be a tiny part of the picture. The UIs themselves are no more different than switching between Windows and MacOS. Canonical has made the install process for Ubuntu as seamless as a Windows install. A great deal of hardware works fine out of the box. Yes, using Linux can be more technically involved than Windows or MacOS. but what's been holding back Linux has largely been software availability. Businesses want Excel, Word, and Outlook. Creative developers want Photoshop and... whatever else creative developers want. Software developers are whores who will program for anything. They're also the majority of desktop Linux users. Folks at home want whatever they have at work. Availability of video games is also no small factor.
Linux's problem isn't so much useability - it's that it has never had a killer app in the desktop realm to generate a user base around.
Wile I agree with most of what you're saying, I don't agree with your conclusions. There are many things that are either better done on a PC or only able to be done on a PC such as software development tasks, editing of photos and videos. Mobile devices are designed around content consumption, not creation. They are definitely going to be grabbing a lot of lighter weight computer users and drive the PC market down to a fraction of its current size. It will certainly mean the end of the road for a number of PC manufacturers.
However, I simply don't see the PC market decline as being terminal. It will drop down to a new standard and stay there.
Also no matte screens, except for maybe the Verix 530.
Oh come on now, anybody who knows anything knows the smoothest running Xeons are V12s or inline 6 cores.... Or boxers, but you know the first and second rules of boxer core club...
If the write fails, I need the program to go belly up. I see no issue here...