All this password brouhaha is silly. If you have physical access, you can always do whatever the heck you want -- for all we know, you can unplug the control wiring from the PLC and run the machine from a pushbutton panel. That has always been the case. All one needs is a button on the PLC that you have to press, perhaps twice, to indicate that you're local and want to let a connection from your laptop access the PLC's administrative functions. Otherwise, if you're doing it from a cental office of some sort, there is no reason at all not to have your public key uploaded to all the PLCs during commissioning. Passwords fill a need that IMHO doesn't exist in real scenarios in a plant. Either you have local access, or you have access to system documentation that includes the file with public key. Anything else invites an "electrician" to cause a lot of damage.
I'm partial to HP gear, and I always claimed that it has quite decent TCO even in very small scale deployments (we have 5k worth of gear, not 40M). People who buy Cisco must be getting a lot of free pussy or something.
I use RDP, and it does perform like crap, if you compare it to a scenegraph-shipping system. I've tinkered with 3D scenegraph shipping over ISDN B-channel (64kbps) and it worked pretty damn good in an interactive application -- way better than any similarly good looking modern application would perform over any common technology (RDP, X11, VNC, whatever).
It's not inefficient given that the GPUs don't even blink when faced with rendering it. What's inefficient is that there's no common way of representing such "effects" to the underlying display system -- a.k.a. scenegraph.
Unless you live next to a major thoroughfare, it doesn't matter. What's important is number of cars per capita, and population (and thus car!) areal density
It's not an unlikely way to do espionage you clod, it's the simplest way to do it. What's simpler than having direct access to all the communications infrastructure, accessible from anywhere in the world?
Yup, even when you a-priori know in which couple hundred lines to look. In a large application, like you'd find in a router, it's demonstrably impossible of a task unless they use something safer than C -- and even then it'd take a formal method approach.
I hate to rain on your parade, but there's plenty of network equipment that can transparently re-terminate a TCP/IP connection -- whatever TCP protocol layer stuff is seen by the either of the endpoints of the link is generated by the network equipment, not the endpoint. This is a must on any long-latency connection. It is also necessary on wireless links, but, unfortunately, most wireless networking equipment isn't that clever. In late 90s an acquaintance of mine was involved in R&D on a custom wireless networking system and it did just that, otherwise TCP/IP performance was unbearable.
It's inefficient because the TCP protocol can't but assume that packet loss is due to network saturation. On wireless links this is practically hardly ever the case.
It does describe the situation. Basically, IP protocol is deficient because there's no way for network equipment to tell the recipient why a packet was lost (and that it was, indeed, lost). TCP assumes that packets are lost because the network is saturated. On wireless links this assumption is most often wrong.
Newsflash: every nontrivial device around you is leveraging the branch of mathematics called algebra. The "algebra" you've learned in grade school, if that's what you're referring to, is quite far removed from what a mathematician would consider algebra. The grade school version does everything to obscure the matters, and only gives you a peek at a very particular application area of algebra.
Well, all common desktop OS implementations of TCP support SACK (selective acknowledgments) so the big retransmits (6-20 in your example) are an issue only if the TCP/IP stack doesn't implement RFC 2018. Said RFC is 25 years old...
Sure everyone makes mistakes. But the deal is that there are lessons to be learned from such mistakes. Autopsies are the only way to let you learn that lesson.
The brainless mind wank that comprises most of twitter is just digital wallpaper.
That's like saying that eBay sucks because the last seller ripped you off. I hope you see the error of your ways. Hint: twitter is just the facilitator. The content comes from, well, wherever you choose it to come from. If you read tabloids, must you also complain about them all the time?
I'm not so sure about that. The problem might be that VHS has some, um, leniency as to how long a head switch takes. I bet the head switch obliterates the teletext scanline...
Drawing commands ha ha ha. X's "drawing commands" are so useless these days that the various toolkits that abstract X away usually dispense with such "drawing commands" and pretend they are not there. Look at your browser: all it does is render html to a bitmap, whether it's gecko- or webkit-based. Any modern Qt or GTK application does the same thing, because falling back to X's "drawing commands" kills performance vs. rendering to bitmaps.
I don't think anyone gives a rat's ass about window managers under any platform other than X11. It's a self-fulfilling prophecy of sorts. Nobody cares about X11 in its present form. It's a bunch of useless crap, it offers inherently no functionality that any modern application would care for except input, blitting and network transparency. Oh, and OpenGL extensions. Frameworks would just as happily dump support of this crap and go to a better alternative, were such to exist. I'm not commenting on the state of Wayland APIs, they may be crap too for all I know (because I don't).
Where I work, even very low impact systems have configuration management, auditing and snapshot/rollback functionality... the rigour you claim is something that can be grasped in a week by any sysadmin out there. No magic to it.
Well, but those add/move/changes should follow an automated workflow that is approved, and any deviations need to be controlled and approved as necessary. IOW anything anyone will be typing at the console is to fill in the fields, not playing at the command prompt.
All this password brouhaha is silly. If you have physical access, you can always do whatever the heck you want -- for all we know, you can unplug the control wiring from the PLC and run the machine from a pushbutton panel. That has always been the case. All one needs is a button on the PLC that you have to press, perhaps twice, to indicate that you're local and want to let a connection from your laptop access the PLC's administrative functions. Otherwise, if you're doing it from a cental office of some sort, there is no reason at all not to have your public key uploaded to all the PLCs during commissioning. Passwords fill a need that IMHO doesn't exist in real scenarios in a plant. Either you have local access, or you have access to system documentation that includes the file with public key. Anything else invites an "electrician" to cause a lot of damage.
I'm partial to HP gear, and I always claimed that it has quite decent TCO even in very small scale deployments (we have 5k worth of gear, not 40M). People who buy Cisco must be getting a lot of free pussy or something.
Now try it with a modern application. You know, anything that uses WPF, for example, and uses it for more than mere semi-static controls.
I use RDP, and it does perform like crap, if you compare it to a scenegraph-shipping system. I've tinkered with 3D scenegraph shipping over ISDN B-channel (64kbps) and it worked pretty damn good in an interactive application -- way better than any similarly good looking modern application would perform over any common technology (RDP, X11, VNC, whatever).
It's not inefficient given that the GPUs don't even blink when faced with rendering it. What's inefficient is that there's no common way of representing such "effects" to the underlying display system -- a.k.a. scenegraph.
Just be glad your 16 month old didn't figure out how to open all the locks on the front door and go for a stroll in the middle of the night :)
Unless you live next to a major thoroughfare, it doesn't matter. What's important is number of cars per capita, and population (and thus car!) areal density
It's not an unlikely way to do espionage you clod, it's the simplest way to do it. What's simpler than having direct access to all the communications infrastructure, accessible from anywhere in the world?
Yup, even when you a-priori know in which couple hundred lines to look. In a large application, like you'd find in a router, it's demonstrably impossible of a task unless they use something safer than C -- and even then it'd take a formal method approach.
I'd have thought that the entire goal was to compile and install it, otherwise the source code is kinda pointless.
I hate to rain on your parade, but there's plenty of network equipment that can transparently re-terminate a TCP/IP connection -- whatever TCP protocol layer stuff is seen by the either of the endpoints of the link is generated by the network equipment, not the endpoint. This is a must on any long-latency connection. It is also necessary on wireless links, but, unfortunately, most wireless networking equipment isn't that clever. In late 90s an acquaintance of mine was involved in R&D on a custom wireless networking system and it did just that, otherwise TCP/IP performance was unbearable.
It's inefficient because the TCP protocol can't but assume that packet loss is due to network saturation. On wireless links this is practically hardly ever the case.
It does describe the situation. Basically, IP protocol is deficient because there's no way for network equipment to tell the recipient why a packet was lost (and that it was, indeed, lost). TCP assumes that packets are lost because the network is saturated. On wireless links this assumption is most often wrong.
Newsflash: every nontrivial device around you is leveraging the branch of mathematics called algebra. The "algebra" you've learned in grade school, if that's what you're referring to, is quite far removed from what a mathematician would consider algebra. The grade school version does everything to obscure the matters, and only gives you a peek at a very particular application area of algebra.
Well, all common desktop OS implementations of TCP support SACK (selective acknowledgments) so the big retransmits (6-20 in your example) are an issue only if the TCP/IP stack doesn't implement RFC 2018. Said RFC is 25 years old...
Sure everyone makes mistakes. But the deal is that there are lessons to be learned from such mistakes. Autopsies are the only way to let you learn that lesson.
I wouldn't consider it the most humane thing to do, but hey, if it'd teach some dumb fucks a lesson and make them less hypocritical: why not.
Well, at least he/she can reclaim some methane and claim going green and helping stave off global warming. Manure, mmm.
The brainless mind wank that comprises most of twitter is just digital wallpaper.
That's like saying that eBay sucks because the last seller ripped you off. I hope you see the error of your ways. Hint: twitter is just the facilitator. The content comes from, well, wherever you choose it to come from. If you read tabloids, must you also complain about them all the time?
I'm not so sure about that. The problem might be that VHS has some, um, leniency as to how long a head switch takes. I bet the head switch obliterates the teletext scanline...
Drawing commands ha ha ha. X's "drawing commands" are so useless these days that the various toolkits that abstract X away usually dispense with such "drawing commands" and pretend they are not there. Look at your browser: all it does is render html to a bitmap, whether it's gecko- or webkit-based. Any modern Qt or GTK application does the same thing, because falling back to X's "drawing commands" kills performance vs. rendering to bitmaps.
I don't think anyone gives a rat's ass about window managers under any platform other than X11. It's a self-fulfilling prophecy of sorts. Nobody cares about X11 in its present form. It's a bunch of useless crap, it offers inherently no functionality that any modern application would care for except input, blitting and network transparency. Oh, and OpenGL extensions. Frameworks would just as happily dump support of this crap and go to a better alternative, were such to exist. I'm not commenting on the state of Wayland APIs, they may be crap too for all I know (because I don't).
Where I work, even very low impact systems have configuration management, auditing and snapshot/rollback functionality... the rigour you claim is something that can be grasped in a week by any sysadmin out there. No magic to it.
I thinker too, but man, if you need to ask around on slashdot for inspiration, you're in a real dry rut :)
Well, but those add/move/changes should follow an automated workflow that is approved, and any deviations need to be controlled and approved as necessary. IOW anything anyone will be typing at the console is to fill in the fields, not playing at the command prompt.