STM32F0 and SAMD21... so it's not cheap and can't be.
Is the connector standardized for CAN? Otherwise they could have picked one that's easy to install by crimping, like RJ11. We've already been there with the Dallas one-wire networks: either use RJ11 to have power, power ground, signal and signal ground, or RJ45 because those connectors are more readily available and some extra pins are there just in case. Or maybe use an audio connector, for convenience and robustness, although those are more trouble to make up your own cables.
But there are other standards for a reliable low-speed low-wire-count low-compute-power network. But differential signaling is a must, and higher voltages help to make it more robust too.
A worthwhile next step would be to get an open core design for one of these incorporated into a next-gen Risc-V based microcontroller. Then all the makers could get behind it, just to support the open-IP ecosystem.
Remember when RepRaps used RS485 between components? (e.g. https://reprap.org/wiki/Extrud...) And there have been smart stepper motors. I kindof thought that idea was going to take off, early on, but most seem to have decided it's cheaper to centralize the logic and the stepper drivers on one board. But that doesn't scale to larger machines. If CAN has an advantage over RS485 for that, it might make some sense; but I still think one micro ought to be enough to implement it; and if it's not, then CAN is probably the wrong choice.
Wireless is popular, but every device needs power so nothing can really be disconnected for the long term, unless it runs from solar power. (Batteries either have to be plugged in to recharge, or else they are environmentally unsustainable. Or both.) And there is the ongoing suspicion that RF exposure might cause health problems too. Whenever that risk finally hits the majority's radar, which technology is going to be in position to be the next contender? LiFi could be fairly easy I think.
I had an idea years ago to incorporate optical fiber into every power cable and every power outlet (simply standardize the position on the plug, relative to the other 3 prongs, assuming a choice of fiber technology such that precise alignment isn't necessary), so that when you plug anything in, you get networking at the same time. But that's a chicken-egg problem.
Alternatively, find a way to make one of the powerline networking standards cheaper. We can't get away from in-wall wiring to power stuff; so, one way or another, the network and the power wiring ought to be combined, IMO.
I can't even get qemu to emulate a raspberry pi faster than the pi itself on a Core i7 processor. I don't suppose going the other direction would work any better.
When I dreamed of having an intelligent computer a decade or two ago, I never dreamed that it could only be accomplished by sending queries to some big corporate-controlled cluster and getting responses back. I don't want to use Siri or Echo, because of this spying which is so far inherent to AI, and because Amazon and Google exist mainly to sell us stuff, to exploit us and get us to buy more of something. When open-source AI is capable of doing something useful, then I will run it on my own machine.
But can we ever expect an AI to get anything done without communicating? A lower standard: can we expect it to communicate to the extent necessary to get something done, but still respect our privacy? To have a positive answer requires an AI with ethics. It's probably more work for the AI to understand what is necessary to respect the user's privacy (like a good friend would do) than to answer the questions we ask of it.
I think it will blow over. But I'm bummed because I wanted to buy one; guess I'd better wait until they have solved it in production rather than by recalling or by just hoping for the best. OTOH it's not "open" enough is it... Sony is trying to get their stuff supported in mainline Linux, so is a Sony phone my best chance of both having "flagship" specs AND running a real convergence-oriented Linux OS on it in the future? It's just a matter of time until Ubuntu and Plasma Mobile will be able to move on from that libybris on top of Android hackery... I hope. I want modern Wayland, modern Qt, the ability to plug into a monitor, and total freedom with software.
I'm getting by with my original Note for now; there's still nothing much wrong with it, other than being slow, not lasting as long as I'd like every day, and needing the occasional reboot. I had to replace the board with the micro-USB connector once; it got too loose and wasn't making good enough contact to charge reliably.
Now if they had just made the new Note with a replaceable battery like my old one, that would make it much easier for anyone who already bought it and doesn't want it igniting in his pocket.
Why aren't the batteries standardized by now anyway? EU should have tried to make that happen, right after the micro-usb charging standard. They could keep growing in capacity, but keep using a few standard sizes. I don't care if the phone ends up a mm or two thicker because of that.
But fine, distributing drivers with binary blobs is OK, while this little license incompatibility between two open-source projects is a big deal. Whatever, dudes.
Prismatic/thin battery sizes could have been standardized years ago, so that the replacements would fit a huge number of devices. Presumably it's just a racket: the manufacturer gets a chance to sell you an overpriced replacement or a new device, your choice.
I actually use Linux, OS X and occasionally Windows at work. (I work on Qt, so I need to be able to test bugs and features on every supported platform.) I use the same keyboard for all of them, on a KVM. It's not so hard to get capslock to be the control key on all of them. (On Linux though, that means remapping it on both the console and in X11 and Wayland globally, not just in some desktop environment's control panel.) But then after using OS X I have grown to like the idea of keeping control (as used on the terminal) separate from command (as used to copy and paste text and other such things, even in the terminal window). I'd like to find a way to make that consistent everywhere. I think I will find a way on Linux. But first I want to have an ergo keyboard with a real command key. (Thus the ergodox, with its customizable firmware.)
For now though, on a regular ergo keyboard, capslock is control, and control is also control on Linux and Windows, whereas on OS X I map capslock to control and control to command and the windows key to command, so that I can develop portable muscle-memory: I use the same key for control-C or command-C regardless of OS when I want to copy text, and the same capslock key for control-C to stop a running program, for example. And the windows key is closer to where the command key is on a MacBook Pro, so if I hit that one by accident it still works, whereas the windows key has no other use on OS X.
It IMO doesn't make as much sense to remap the windows key, because it has its own uses in various Linux desktops and window managers now. So, the more bucky keys, the better; it seems I actually use 5 of them now: control, command, alt, window and compose. Oh, and shift, of course. The window key is still the least useful though. Can't decide whether I prefer using window-mouse-drag or alt-mouse-drag for moving windows around. There is some inconsistency about that between openbox, awesome and weston.
I was using a model-M at home, with no Windows key, but my wife complained that she couldn't sleep, so that's the trouble with that. But then I had to try to use the two alt keys more effectively. The right one has to be compose if you need compose, whereas if you also have a menu key, you could use that for compose. I live in Norway now, and don't want to get used to Norwegian keyboards because they are too different (lots of odd and unnecessary changes), because I don't actually write much in the Norwegian language, and because I will never be able to switch completely due to having lots of old hardware. So I use compose for writing the 3 extra characters ø å and æ. I think it's an intuitive and extensible way to handle all the accented characters, and everyone should have a convenient way of typing them, even if they are only needed in rare cases.
After I switch to ergodox, I wonder if it's going to get really hard to use other keyboards though.
I have ordered a couple variants of these; haven't received either of them yet. But it sounds good in theory. And there are no IP issues: it's an open-source design, with open-source firmware, completely programmable so that each user can customize his own layout. (For the first time I'll have a keyboard I really "own".) The keyboard can still be made flat (although usually the two halves are positioned independently), so any laptop manufacturer could even start shipping with this kind of layout. I hope that will happen at some point.
The main points are that the two halves are rotated to a better angle to reduce strain, and also that the thumbs are able to operate several keys each, not just the spacebar.
What to do with capslock is such a minor issue compared to everything else that's wrong with QWERTY. This is why, despite being quite a fast typist, and not suffering particularly much wrist pain, I nevertheless feel that it's time to try something more efficient. I do have some shoulder pain sometimes, so figured maybe that's from having to hold my hands too close together in front of me. It's worse when I use a laptop too much, whereas at work I've been using an old Microsoft 4000 ergo keyboard for years.
Now if they would just ship...;-)
An similar alternative is keyboard.io. From one side, I wish I'd waited for their crowdfunding campaign to get started, because it's gorgeous, and I love the wooden-case idea. From another side, I think theirs is going to be less customizable: every key has a different shape, so you can't have printed rearrangeable keycaps, e.g. if you want to try colemak you have to ignore the legends on the keys, or get keys without legends. And they are definitely not going to ship until the middle of next year, either.
See also the story about the Amiga HVAC controller which someone else already linked to...
If the system is expected to run for 25 years, then it should still be able to run its own compiler in 25 years too. As long as it's not underpowered, it should be able to build its own software. So do the development that way too. Although I do wonder if flash storage is up to a 25-year lifespan.
But on the flip side, it amazes me that some not-so-geeky people manage to make use of Arduinos. The whole package is a not-so-thick veneer over plain Atmel C programming, and yet it opened up the chance for many times more people than would have ever managed with the Atmel-provided IDE. And then there's Processing for an even less-geeky approach.
Web technology is just stupid... all the usual languages are bad ones, and even if you manage to find a cool language to develop in, you still have to use it to translate everything to XML and Javascript at the end. Web 3.0 will hopefully happen eventually. The language of the future should be declarative and imperative at the same time, extensible at the metalanguage level so that it can be adapted for every task on both client and server, and with elegant syntax too. I don't enjoy web programming as much as writing applications and frameworks, not because I couldn't learn enough about the technology but because it's kindof disgusting. And now it's hard to switch away from "the browser" to something else, because the replacement would have to be more compelling on so many levels that people would actually use it, and avoid getting corrupted too early with commercials and spying and malware and crap. It should be clean and elegant and there should be some kind of self-enforcing social contract that keeps it that way. I suppose that part is impossible though...
Experience has shown that cell division is just a precursor to further growth.
How about we go back to having an unambiguously named War Department which is in recess whenever there is not war (which should be most of the time), and abolish the rest of the agencies that are determined to make enemies of everyone possible? Policing the people should be a matter for the states anyway.
I worked at Nokia from 2011-2012. Everyone was saying then that the reason for Elop (who was otherwise so useless) was to devalue Nokia enough that it would be a good deal for Microsoft. And here we are... the other shoe drops. But there will be a third shoe when he becomes CEO of Microsoft. They deserve each other.
The same strategy enabled high-EER air conditioning: use a small compressor which runs most of the time plus a larger one to handle peak cooling loads, rather than an even bigger compressor which cycles on and off frequently.
Siri? Google Glass and the apps that run on it? Google Voice turning your voicemails into emails as fast as you receive them? Turn-by-turn directions developed independently by several companies? This stuff used to be called AI.
A course that covers the history of software would be about as useful as, say, music appreciation or art history. It could be taught in a non-CS department just to get some of that liberal arts flavor, maybe even count as one of the required humanities credits.
That's a good idea actually. It just takes a lot of space relative to the cargo it can carry, but the sky is big so what the heck. I suppose it's just not going to be fast enough for passenger transport.
Maybe if the envelope was e-ink, you could make it rise by turning it black (to absorb sun) and fall by turning it white. Or use a tether to raise and lower the cargo so that you don't have to completely land in order to "drop-ship" something. Not that it would work so well when the wind is blowing...
Yeah it's been a while hasn't it... I think I've read at least 80% or so of the days since sometime in the 90's, it might've been 97; it's like smoking must be (not that I've tried), addictive and fills the awkward moments. You guys threw a cool party "get sloshed with slashdot" at a silicon valley Linux conference in 1999 I think (celebrating the Andover money I suppose); I and a friend from the local LUG rented a minivan and drove from Phoenix for that (well, for the conference, not only the party;-). I hope the site is not going downhill in any way after your leaving. But I can appreciate the need to find some other meaning to life after such a long time in one gig.;-) So good luck with whatever you end up doing next, and thanks for all the news and entertainment for all those years. As they say, time flies like arrows...
Slashdot Mirror
STM32F0 and SAMD21... so it's not cheap and can't be.
Is the connector standardized for CAN? Otherwise they could have picked one that's easy to install by crimping, like RJ11. We've already been there with the Dallas one-wire networks: either use RJ11 to have power, power ground, signal and signal ground, or RJ45 because those connectors are more readily available and some extra pins are there just in case. Or maybe use an audio connector, for convenience and robustness, although those are more trouble to make up your own cables.
But there are other standards for a reliable low-speed low-wire-count low-compute-power network. But differential signaling is a must, and higher voltages help to make it more robust too.
A worthwhile next step would be to get an open core design for one of these incorporated into a next-gen Risc-V based microcontroller. Then all the makers could get behind it, just to support the open-IP ecosystem.
Remember when RepRaps used RS485 between components? (e.g. https://reprap.org/wiki/Extrud...) And there have been smart stepper motors. I kindof thought that idea was going to take off, early on, but most seem to have decided it's cheaper to centralize the logic and the stepper drivers on one board. But that doesn't scale to larger machines. If CAN has an advantage over RS485 for that, it might make some sense; but I still think one micro ought to be enough to implement it; and if it's not, then CAN is probably the wrong choice.
Wireless is popular, but every device needs power so nothing can really be disconnected for the long term, unless it runs from solar power. (Batteries either have to be plugged in to recharge, or else they are environmentally unsustainable. Or both.) And there is the ongoing suspicion that RF exposure might cause health problems too. Whenever that risk finally hits the majority's radar, which technology is going to be in position to be the next contender? LiFi could be fairly easy I think.
I had an idea years ago to incorporate optical fiber into every power cable and every power outlet (simply standardize the position on the plug, relative to the other 3 prongs, assuming a choice of fiber technology such that precise alignment isn't necessary), so that when you plug anything in, you get networking at the same time. But that's a chicken-egg problem.
Alternatively, find a way to make one of the powerline networking standards cheaper. We can't get away from in-wall wiring to power stuff; so, one way or another, the network and the power wiring ought to be combined, IMO.
Reminds me of Mystery Science Theater 3000.
Maybe that explains it...
CONFIG_DOUBLE_STUF
cause you know more is always better...
I can't even get qemu to emulate a raspberry pi faster than the pi itself on a Core i7 processor. I don't suppose going the other direction would work any better.
When I dreamed of having an intelligent computer a decade or two ago, I never dreamed that it could only be accomplished by sending queries to some big corporate-controlled cluster and getting responses back. I don't want to use Siri or Echo, because of this spying which is so far inherent to AI, and because Amazon and Google exist mainly to sell us stuff, to exploit us and get us to buy more of something. When open-source AI is capable of doing something useful, then I will run it on my own machine.
But can we ever expect an AI to get anything done without communicating? A lower standard: can we expect it to communicate to the extent necessary to get something done, but still respect our privacy? To have a positive answer requires an AI with ethics. It's probably more work for the AI to understand what is necessary to respect the user's privacy (like a good friend would do) than to answer the questions we ask of it.
So it's literally a burning platform.
I think it will blow over. But I'm bummed because I wanted to buy one; guess I'd better wait until they have solved it in production rather than by recalling or by just hoping for the best. OTOH it's not "open" enough is it... Sony is trying to get their stuff supported in mainline Linux, so is a Sony phone my best chance of both having "flagship" specs AND running a real convergence-oriented Linux OS on it in the future? It's just a matter of time until Ubuntu and Plasma Mobile will be able to move on from that libybris on top of Android hackery... I hope. I want modern Wayland, modern Qt, the ability to plug into a monitor, and total freedom with software.
I'm getting by with my original Note for now; there's still nothing much wrong with it, other than being slow, not lasting as long as I'd like every day, and needing the occasional reboot. I had to replace the board with the micro-USB connector once; it got too loose and wasn't making good enough contact to charge reliably.
Now if they had just made the new Note with a replaceable battery like my old one, that would make it much easier for anyone who already bought it and doesn't want it igniting in his pocket.
Why aren't the batteries standardized by now anyway? EU should have tried to make that happen, right after the micro-usb charging standard. They could keep growing in capacity, but keep using a few standard sizes. I don't care if the phone ends up a mm or two thicker because of that.
I happen to like using ZFS.
But fine, distributing drivers with binary blobs is OK, while this little license incompatibility between two open-source projects is a big deal. Whatever, dudes.
Seriously, 3 years of unfulfilled temptation? Such a vial wouldn't survive very long on the Mir.
But then, vodka doesn't benefit so much from aging.
Prismatic/thin battery sizes could have been standardized years ago, so that the replacements would fit a huge number of devices. Presumably it's just a racket: the manufacturer gets a chance to sell you an overpriced replacement or a new device, your choice.
I actually use Linux, OS X and occasionally Windows at work. (I work on Qt, so I need to be able to test bugs and features on every supported platform.) I use the same keyboard for all of them, on a KVM. It's not so hard to get capslock to be the control key on all of them. (On Linux though, that means remapping it on both the console and in X11 and Wayland globally, not just in some desktop environment's control panel.) But then after using OS X I have grown to like the idea of keeping control (as used on the terminal) separate from command (as used to copy and paste text and other such things, even in the terminal window). I'd like to find a way to make that consistent everywhere. I think I will find a way on Linux. But first I want to have an ergo keyboard with a real command key. (Thus the ergodox, with its customizable firmware.)
For now though, on a regular ergo keyboard, capslock is control, and control is also control on Linux and Windows, whereas on OS X I map capslock to control and control to command and the windows key to command, so that I can develop portable muscle-memory: I use the same key for control-C or command-C regardless of OS when I want to copy text, and the same capslock key for control-C to stop a running program, for example. And the windows key is closer to where the command key is on a MacBook Pro, so if I hit that one by accident it still works, whereas the windows key has no other use on OS X.
It IMO doesn't make as much sense to remap the windows key, because it has its own uses in various Linux desktops and window managers now. So, the more bucky keys, the better; it seems I actually use 5 of them now: control, command, alt, window and compose. Oh, and shift, of course. The window key is still the least useful though. Can't decide whether I prefer using window-mouse-drag or alt-mouse-drag for moving windows around. There is some inconsistency about that between openbox, awesome and weston.
I was using a model-M at home, with no Windows key, but my wife complained that she couldn't sleep, so that's the trouble with that. But then I had to try to use the two alt keys more effectively. The right one has to be compose if you need compose, whereas if you also have a menu key, you could use that for compose. I live in Norway now, and don't want to get used to Norwegian keyboards because they are too different (lots of odd and unnecessary changes), because I don't actually write much in the Norwegian language, and because I will never be able to switch completely due to having lots of old hardware. So I use compose for writing the 3 extra characters ø å and æ. I think it's an intuitive and extensible way to handle all the accented characters, and everyone should have a convenient way of typing them, even if they are only needed in rare cases.
After I switch to ergodox, I wonder if it's going to get really hard to use other keyboards though.
I have ordered a couple variants of these; haven't received either of them yet. But it sounds good in theory. And there are no IP issues: it's an open-source design, with open-source firmware, completely programmable so that each user can customize his own layout. (For the first time I'll have a keyboard I really "own".) The keyboard can still be made flat (although usually the two halves are positioned independently), so any laptop manufacturer could even start shipping with this kind of layout. I hope that will happen at some point.
The main points are that the two halves are rotated to a better angle to reduce strain, and also that the thumbs are able to operate several keys each, not just the spacebar.
What to do with capslock is such a minor issue compared to everything else that's wrong with QWERTY. This is why, despite being quite a fast typist, and not suffering particularly much wrist pain, I nevertheless feel that it's time to try something more efficient. I do have some shoulder pain sometimes, so figured maybe that's from having to hold my hands too close together in front of me. It's worse when I use a laptop too much, whereas at work I've been using an old Microsoft 4000 ergo keyboard for years.
Now if they would just ship... ;-)
An similar alternative is keyboard.io. From one side, I wish I'd waited for their crowdfunding campaign to get started, because it's gorgeous, and I love the wooden-case idea. From another side, I think theirs is going to be less customizable: every key has a different shape, so you can't have printed rearrangeable keycaps, e.g. if you want to try colemak you have to ignore the legends on the keys, or get keys without legends. And they are definitely not going to ship until the middle of next year, either.
See also the story about the Amiga HVAC controller which someone else already linked to...
If the system is expected to run for 25 years, then it should still be able to run its own compiler in 25 years too. As long as it's not underpowered, it should be able to build its own software. So do the development that way too. Although I do wonder if flash storage is up to a 25-year lifespan.
It must have been a champagne supernova in the sky.
It would be appropriate if the medal was made of Field's metal.
But on the flip side, it amazes me that some not-so-geeky people manage to make use of Arduinos. The whole package is a not-so-thick veneer over plain Atmel C programming, and yet it opened up the chance for many times more people than would have ever managed with the Atmel-provided IDE. And then there's Processing for an even less-geeky approach.
Web technology is just stupid... all the usual languages are bad ones, and even if you manage to find a cool language to develop in, you still have to use it to translate everything to XML and Javascript at the end. Web 3.0 will hopefully happen eventually. The language of the future should be declarative and imperative at the same time, extensible at the metalanguage level so that it can be adapted for every task on both client and server, and with elegant syntax too. I don't enjoy web programming as much as writing applications and frameworks, not because I couldn't learn enough about the technology but because it's kindof disgusting. And now it's hard to switch away from "the browser" to something else, because the replacement would have to be more compelling on so many levels that people would actually use it, and avoid getting corrupted too early with commercials and spying and malware and crap. It should be clean and elegant and there should be some kind of self-enforcing social contract that keeps it that way. I suppose that part is impossible though...
Experience has shown that cell division is just a precursor to further growth.
How about we go back to having an unambiguously named War Department which is in recess whenever there is not war (which should be most of the time), and abolish the rest of the agencies that are determined to make enemies of everyone possible? Policing the people should be a matter for the states anyway.
(the term from Snow Crash)
I worked at Nokia from 2011-2012. Everyone was saying then that the reason for Elop (who was otherwise so useless) was to devalue Nokia enough that it would be a good deal for Microsoft. And here we are... the other shoe drops. But there will be a third shoe when he becomes CEO of Microsoft. They deserve each other.
The same strategy enabled high-EER air conditioning: use a small compressor which runs most of the time plus a larger one to handle peak cooling loads, rather than an even bigger compressor which cycles on and off frequently.
Siri? Google Glass and the apps that run on it? Google Voice turning your voicemails into emails as fast as you receive them? Turn-by-turn directions developed independently by several companies? This stuff used to be called AI.
A course that covers the history of software would be about as useful as, say, music appreciation or art history. It could be taught in a non-CS department just to get some of that liberal arts flavor, maybe even count as one of the required humanities credits.
That's a good idea actually. It just takes a lot of space relative to the cargo it can carry, but the sky is big so what the heck. I suppose it's just not going to be fast enough for passenger transport.
Maybe if the envelope was e-ink, you could make it rise by turning it black (to absorb sun) and fall by turning it white. Or use a tether to raise and lower the cargo so that you don't have to completely land in order to "drop-ship" something. Not that it would work so well when the wind is blowing...
Yeah it's been a while hasn't it... I think I've read at least 80% or so of the days since sometime in the 90's, it might've been 97; it's like smoking must be (not that I've tried), addictive and fills the awkward moments. You guys threw a cool party "get sloshed with slashdot" at a silicon valley Linux conference in 1999 I think (celebrating the Andover money I suppose); I and a friend from the local LUG rented a minivan and drove from Phoenix for that (well, for the conference, not only the party ;-). I hope the site is not going downhill in any way after your leaving. But I can appreciate the need to find some other meaning to life after such a long time in one gig. ;-) So good luck with whatever you end up doing next, and thanks for all the news and entertainment for all those years. As they say, time flies like arrows...
Seriously I remember hearing it back in the 70's. Enough already. Transistors are so small that human hair isn't even a reasonable comparison to make.