We still use C because no one has really come up with a suitable replacement that lots of programmers know. There is a subset of C++ that is good, in fact preferrable to C, but that is often abused because someone will start expanding that envelope to use more and more C++ features until something breaks. They swear, just a simple template only a one line, then in a month or two they've got full page templates obsfuscating the code to hell and back. So C it is. You know Ada might be ok I'd be willing to a whole new device with it, except that it would be hard to find team members who know it or would be willing to learn it.
(and it's tough to try to get those people to move away from the 70s and start using C99, and then I get complaints about having to use "const" I swear that is true and I'm not making it up)
I've got 20KB RAM now but only 128KB flash for code. Coming from a project with a few megabytes, so everyone gets told to be lean and mean. And still some people copy and paste code the larger system and worry about memory usage just an after thought. Then they complain that they're overflowing the stack and can they have some more please (Oliver! jokes inserted as necessary).
I've considered looking at llvm but there aren't any prebuilt compilers that I've found and the build instructions are way more complicated than gcc and I don't have the time to spend on that.
There were standards and procedures before Therac. The regulation could have been tightened more with more audits of course. And some of the complaints there were kind of ridiculous, like using assembler or a custom OS, things that tons of medical devices still do very extremely good reasons. The problems ultimately were management problems.
Interesting that one important cause of failure was reusing older software that had reliance on some hardware interlocks. Yet today it is practically a religion in most places, even with medical devices, that you must always reuse software and never write anything from scratch. Except that re-use should never short cut the testing, always assume that the tried and true library functions have bugs, and assume that the compiler and operating system are buggy as well (so many major bugs in commercial RTOSs it's not even funny).
Blame the consumer for not asking about security options. If their thermostat is unsecure as an IoT device because it connects to their wifi router, then I wouldn't put any bets about the security of their laptop or smart TV either. The rise of security problems is not necessarily because of IoT security but because there are not so many more things all on the same internet. The security needs to be added even when the consumer is not asking for those features, even if it raises the cost of the products. I think it's good that Apple is encrypting phones and storage by default because the average user would not take that extra step on their own.
To most people the internet is still a new concept. Even people who've grown up with the internet are treating it in naive ways. So right now IoT devices with zero security sounds like a dumb thing, but then look around and see how many wi-fi routers you can see from where you are which are open to everyone or which use WPA. Bad security is *everywhere* because few people take it seriously and few are as paranoid as they should be. Go back ten years and remember how full of security holes everything was, yet the security technology has not really gotten that much better over that period. What has changed is that more devices, companies, and users make use of existing security technology.
Zigbee is old and crusty, the newest version is just strange and bloated and no one has really adopted it. It may die off except that big companies keep demanding Zigbee as a check-off box. The standards of this are new and evolving, and security isn't always there but the device makers are adding it anyway (and if you insist on alliance led standards for security then you'll get crap like WPA as a result when a manufacturer might actually have something better).
Big problem is with the dumb IoT, devices that you really don't need but which want to be on the network and in the cloud. Thermostats and baby cams and such. But tell the Gen Z couple that they want a secure VPN to connect to their baby cam and their eyes will glaze over, since they just want an app on their phone. The makers in this fad market are hoping to monetize all that data in the cloud so that they can finally go IPO and retire before they turn 30.
We really need to split up IoT into categories. After all any device or computer that is addressable on the internet is a part of the IoT, and some people even consider point-to-point bluetooth links to your phone to be IoT just so they can jump on that buzzword. IoT for home automation and baby monitors should not be in the same category as IoT devices for utility infrastructure or scientific data collection.
Yes there are secure IoTs. Problem is with generic devices using generic operating systems with no security added or added as a late afterthought. Ie, "consumer" devices are the ones to beware of. Breaking into the coffee maker isn't giving you any access to your thermostat as they're not connected to each other except for using the same air space. A lot of these are relatively big and bulky devices, full android or linux maybe, with wi-fi networking and all its problems. Cheap devices made by companies with minimal profit margins sold to hipsters and yuppies and gadget-philes, consumers who want the bragging rights who don't care about or understand security. The devices in that case don't talk to each other, but they all talk to the common access point (wifi router) which is a weak link.
I work on devices for utilities, municipalities, and the like. We never used to call them IoT until that term started being the fashion recently. But many of those customers are very insistent upon having good security. This extra panic about security is good and bad, the good thing is that it makes some the bigger customers start demanding security. The security for larger customers is by necessity complicated. Good security is never convenient, it means managing cerfiticate chains, providing temporary authority for field service or installers, auditing, etc.
It's how gamers think too. If it works for him then the fault must lie with the other person's system. No one wants to admit that it worked for them because they were lucky.
Second snag, most companies build or modify some of their ethernet cables. No one in IT is going to splice up a 2 foot HDMI cable for you, but they can do that for ethernet in a couple of minutes. It's very expensive to head to the local ripoff-the-geek store to buy these very simple cables when you can have a spool of cable, a bag of connectors, a crimper, and save money.
People like us would never have built Facebook, because Facebook is a silly idea. Sure, it's a silly idea that makes money but it's still a silly idea. Getting rich is not the same as providing something of benefit to the human race, no matter how much the rich execs try to tell engineers that making money is the only goal out there.
GDB is scriptable. That trumps most things that don't have scripting. It is also portable and available for many different different chips and operating systems. Most IDEs for PCs assume there is no other platform in existence, many IDEs are quickly tossed together from what I can see when I'm forced to use them.IDEs like Eclipse are built on top of GDB precisely because it is portable. I used an IDE in 1981 if we're playing that game.
I don't see the point. Other than that Japan is a seriously humanoid robot obsessed country that should get back on their meds. A robot should do work, making it as human-like as possible is just not very useful and slows down robotic research. These are just better grades of mannequins.
"Slashdot is trying to slam me. Who do they think they are. Have you seen their poll numbers? They're nothing. No one goes to that site anyway. Vote for me. I'll be huuuuge. Now who wants to buy some steaks?"
I'm sticking with the Nigerian prince who sent me some email, he claims to have a full plan to fix the economy, restore jobs, and make Pluto a planet again. Besides he's the only one who actually says what he thinks.
Sounds like Linux. The GUI administration there is awful, and each distribution has a different set of commands and methods for doing administration. Then again if you're on Windows you have to use the registry for a lot of important stuff.
No, but I can avoid companies that use it and criticize them for turning off part of their potential market. I am an adult, I'd prefer that I be treated like an adult. but certainly it's up to those companies to make themselves look bad on their own.
OSX may be broken, but it can still limp along when Windows is twitching in a ditch and being left behind. Seriously, worst UIs in the world all come from Microsoft. People were using Unix to develop before Gates wrote his first BASIC.
I think that instead of people wondering why the hell someone would want OSX they should instead ask what Windows has that would make a developer want to stay.
I'm a developer. I do not want to be the sysadmin for my own machine. It wastes too much of my time. Now granted I do have to do it from time to time on my Mac at work, but it's far simpler than trying to deal with fixing problems in Linux. I used to be a Unix sysadmin, later I used Linux a lot, but these days Linux is just too complicated. Even worse if you have xubuntu or kubuntu so that all of the web searches only give you solutions that work for ubuntu.
And you can get Mac Ports to put other tools on the Mac easily.
If you're forced to have an enterprise machine by the corporate ideology, would you want Windows with a slow and clumsy Cygwin or a native Unix that can run native tools? No, the enterprise people probably won't let you have Linux unless it's a second machine.
No one's really come out with a debugger better than gdb. Everything else tends towards IDEs and cumbersome point and click and rarely any scripting capabilities. Sure gdb is far from perfect but the competition is even further away.
Basically it's Unix while at the same time keeping the corporate people in IT happy. People get nervous around Linux, and many devs with long term experience prefer Unix over the lock-in from Microsoft.
OSX is basically a nice mix between Unix to make developers happy and enterprise software (ie, you can run Office which is almost mandatory for most corporations). Linux is nice too but it lacks a lot of stuff that the enterprise wants, or the developer who doesn't want to waste time micromanaging it. The Mac Book Pro, though expensive, is very nice to use. Probably some Apple patents around which is why other PC laptops have really clumsy touch pads.
Nope, never been to a dance club. All the wrong sorts of people there, frat boys, gropers, drunks. Sort of like a Microsoft event but with less chair throwing.
Slashdot Mirror
We still use C because no one has really come up with a suitable replacement that lots of programmers know. There is a subset of C++ that is good, in fact preferrable to C, but that is often abused because someone will start expanding that envelope to use more and more C++ features until something breaks. They swear, just a simple template only a one line, then in a month or two they've got full page templates obsfuscating the code to hell and back. So C it is. You know Ada might be ok I'd be willing to a whole new device with it, except that it would be hard to find team members who know it or would be willing to learn it.
(and it's tough to try to get those people to move away from the 70s and start using C99, and then I get complaints about having to use "const" I swear that is true and I'm not making it up)
I've got 20KB RAM now but only 128KB flash for code. Coming from a project with a few megabytes, so everyone gets told to be lean and mean. And still some people copy and paste code the larger system and worry about memory usage just an after thought. Then they complain that they're overflowing the stack and can they have some more please (Oliver! jokes inserted as necessary).
I've considered looking at llvm but there aren't any prebuilt compilers that I've found and the build instructions are way more complicated than gcc and I don't have the time to spend on that.
There were standards and procedures before Therac. The regulation could have been tightened more with more audits of course. And some of the complaints there were kind of ridiculous, like using assembler or a custom OS, things that tons of medical devices still do very extremely good reasons. The problems ultimately were management problems.
Interesting that one important cause of failure was reusing older software that had reliance on some hardware interlocks. Yet today it is practically a religion in most places, even with medical devices, that you must always reuse software and never write anything from scratch. Except that re-use should never short cut the testing, always assume that the tried and true library functions have bugs, and assume that the compiler and operating system are buggy as well (so many major bugs in commercial RTOSs it's not even funny).
Blame the consumer for not asking about security options. If their thermostat is unsecure as an IoT device because it connects to their wifi router, then I wouldn't put any bets about the security of their laptop or smart TV either. The rise of security problems is not necessarily because of IoT security but because there are not so many more things all on the same internet. The security needs to be added even when the consumer is not asking for those features, even if it raises the cost of the products. I think it's good that Apple is encrypting phones and storage by default because the average user would not take that extra step on their own.
To most people the internet is still a new concept. Even people who've grown up with the internet are treating it in naive ways. So right now IoT devices with zero security sounds like a dumb thing, but then look around and see how many wi-fi routers you can see from where you are which are open to everyone or which use WPA. Bad security is *everywhere* because few people take it seriously and few are as paranoid as they should be. Go back ten years and remember how full of security holes everything was, yet the security technology has not really gotten that much better over that period. What has changed is that more devices, companies, and users make use of existing security technology.
Zigbee is old and crusty, the newest version is just strange and bloated and no one has really adopted it. It may die off except that big companies keep demanding Zigbee as a check-off box. The standards of this are new and evolving, and security isn't always there but the device makers are adding it anyway (and if you insist on alliance led standards for security then you'll get crap like WPA as a result when a manufacturer might actually have something better).
Big problem is with the dumb IoT, devices that you really don't need but which want to be on the network and in the cloud. Thermostats and baby cams and such. But tell the Gen Z couple that they want a secure VPN to connect to their baby cam and their eyes will glaze over, since they just want an app on their phone. The makers in this fad market are hoping to monetize all that data in the cloud so that they can finally go IPO and retire before they turn 30.
We really need to split up IoT into categories. After all any device or computer that is addressable on the internet is a part of the IoT, and some people even consider point-to-point bluetooth links to your phone to be IoT just so they can jump on that buzzword. IoT for home automation and baby monitors should not be in the same category as IoT devices for utility infrastructure or scientific data collection.
Yes there are secure IoTs. Problem is with generic devices using generic operating systems with no security added or added as a late afterthought. Ie, "consumer" devices are the ones to beware of. Breaking into the coffee maker isn't giving you any access to your thermostat as they're not connected to each other except for using the same air space. A lot of these are relatively big and bulky devices, full android or linux maybe, with wi-fi networking and all its problems. Cheap devices made by companies with minimal profit margins sold to hipsters and yuppies and gadget-philes, consumers who want the bragging rights who don't care about or understand security. The devices in that case don't talk to each other, but they all talk to the common access point (wifi router) which is a weak link.
I work on devices for utilities, municipalities, and the like. We never used to call them IoT until that term started being the fashion recently. But many of those customers are very insistent upon having good security. This extra panic about security is good and bad, the good thing is that it makes some the bigger customers start demanding security. The security for larger customers is by necessity complicated. Good security is never convenient, it means managing cerfiticate chains, providing temporary authority for field service or installers, auditing, etc.
Peter Gibbons, is that you?
It's how gamers think too. If it works for him then the fault must lie with the other person's system. No one wants to admit that it worked for them because they were lucky.
Dumb people never get the feeling that they're working with a bunch of idiots all the time.
I agree. I can drink my bottle of wine in the corner in the dark and not have to share it with anyone and no one tries to talk me back into rehab.
Second snag, most companies build or modify some of their ethernet cables. No one in IT is going to splice up a 2 foot HDMI cable for you, but they can do that for ethernet in a couple of minutes. It's very expensive to head to the local ripoff-the-geek store to buy these very simple cables when you can have a spool of cable, a bag of connectors, a crimper, and save money.
People like us would never have built Facebook, because Facebook is a silly idea. Sure, it's a silly idea that makes money but it's still a silly idea. Getting rich is not the same as providing something of benefit to the human race, no matter how much the rich execs try to tell engineers that making money is the only goal out there.
1) Never trust an executive to tell what makes a great engineer. 2) Never let anyone at Facebook tell you what makes a great engineer.
GDB is scriptable. That trumps most things that don't have scripting. It is also portable and available for many different different chips and operating systems. Most IDEs for PCs assume there is no other platform in existence, many IDEs are quickly tossed together from what I can see when I'm forced to use them.IDEs like Eclipse are built on top of GDB precisely because it is portable. I used an IDE in 1981 if we're playing that game.
I don't see the point. Other than that Japan is a seriously humanoid robot obsessed country that should get back on their meds. A robot should do work, making it as human-like as possible is just not very useful and slows down robotic research. These are just better grades of mannequins.
"Slashdot is trying to slam me. Who do they think they are. Have you seen their poll numbers? They're nothing. No one goes to that site anyway. Vote for me. I'll be huuuuge. Now who wants to buy some steaks?"
I'm sticking with the Nigerian prince who sent me some email, he claims to have a full plan to fix the economy, restore jobs, and make Pluto a planet again. Besides he's the only one who actually says what he thinks.
Sounds like Linux. The GUI administration there is awful, and each distribution has a different set of commands and methods for doing administration. Then again if you're on Windows you have to use the registry for a lot of important stuff.
No, but I can avoid companies that use it and criticize them for turning off part of their potential market. I am an adult, I'd prefer that I be treated like an adult. but certainly it's up to those companies to make themselves look bad on their own.
OSX may be broken, but it can still limp along when Windows is twitching in a ditch and being left behind. Seriously, worst UIs in the world all come from Microsoft. People were using Unix to develop before Gates wrote his first BASIC.
I think that instead of people wondering why the hell someone would want OSX they should instead ask what Windows has that would make a developer want to stay.
I'm a developer. I do not want to be the sysadmin for my own machine. It wastes too much of my time. Now granted I do have to do it from time to time on my Mac at work, but it's far simpler than trying to deal with fixing problems in Linux. I used to be a Unix sysadmin, later I used Linux a lot, but these days Linux is just too complicated. Even worse if you have xubuntu or kubuntu so that all of the web searches only give you solutions that work for ubuntu.
And you can get Mac Ports to put other tools on the Mac easily.
If you're forced to have an enterprise machine by the corporate ideology, would you want Windows with a slow and clumsy Cygwin or a native Unix that can run native tools? No, the enterprise people probably won't let you have Linux unless it's a second machine.
No one's really come out with a debugger better than gdb. Everything else tends towards IDEs and cumbersome point and click and rarely any scripting capabilities. Sure gdb is far from perfect but the competition is even further away.
Basically it's Unix while at the same time keeping the corporate people in IT happy. People get nervous around Linux, and many devs with long term experience prefer Unix over the lock-in from Microsoft.
OSX is basically a nice mix between Unix to make developers happy and enterprise software (ie, you can run Office which is almost mandatory for most corporations). Linux is nice too but it lacks a lot of stuff that the enterprise wants, or the developer who doesn't want to waste time micromanaging it.
The Mac Book Pro, though expensive, is very nice to use. Probably some Apple patents around which is why other PC laptops have really clumsy touch pads.
Nope, never been to a dance club. All the wrong sorts of people there, frat boys, gropers, drunks. Sort of like a Microsoft event but with less chair throwing.