Back when 2Mbits was fast, I used to hook subsystems together with 5V RS232esque serial. The interface chips generally supported a synchronous mode (I.E. same byzantine start+data+parity+stop frame format, but with a clock line). So hook up the clock line, and set it to 2Mbit/s. You could multi drop it to and talk to multiple boards in a rack with a little added protocol.
Async RS232 always seemed like a poor tradeoff. A bucket load of signals (DTR, DCD, DSR, RI, RTS,CTS, RTR) in addition to the important ones (tx, rx, gnd, chassis). Then there was a bucket load of complexity with oversampling receivers to sync to the async frame. Why couldn't two of the bucket-o-signals have been a source clock? Let the modem handle the async complexity and send the recovered clock on the wire.
Anyway, to the point or your post, When anything fast or busy was happening I always found it to be a good plan to stick a micro behind the ACIA and handling buffering as it's only job. You couldn't trust OTS hardware to do it well.
If you aren't designing your own PCBs, this might not be an option.
Why is an FTDI serial driver needed? USB has had a serial port protocol as part of its base spec and Windows has a default driver for things declaring themselves to be a serial port. I have several devices that work in this manner.
Why would a vendor of a basic USB-Serial port converter bother writing a driver?
what he meant was that the apple computers he has bought since have all failed within 6 years.
seriously, the rate the mobos go bad should be alarming. who cares if it's encased in aluminium if it just breaks some subsystem or another when it wants to. just today hd controller broke on one over here.. solution? buy a new mac of course.
It's the batteries that go bad first. But that is common to all laptops. The choice of replacement is based on other factors.
Curing blindness sounds admirable - but at what cost? Don't forget the unseen side effects! The first Google hit on CRISPR side effects is: http://phys.org/news/2015-10-c...
Without a suitable blindness cure, the side effects will remain literally unseen.
Islam does not work well with Freedom of choice. There's only one way, Conformity. It started as a political movement, and so it continues. Religion is more economical for control than sheer force. People are born without religion, 99.999% is adopted from their parents. There is no choice involved. The parents are always right.
The methodist doctrine didn't stick in my case. Are the ratios the same when you aren't under threat of death if you choose atheism? Or am I one of the 0.001%?
2) iMessage which handle text messaging and FaceTime, encrypts "End to End". The "End to End" encryption generates an public private key pair where the private key never leaves the senders phone and which can't be determined by Apple engineers even if Apple has thousands of years and unlimited CPU cycles. "End to End" scares lawenforcement at the very highest and lowest levels. iMessage appears to be U breakable, to such a degree that Apple's lawyers feel perfectly comfortable walking into any court anywhere in the world and telling the judges, sorry, can't help you. Stunning is all I can say.
You're envisaging RSA when you say "public private key pair". Given what we know, it's probably Diffie Hellman, which is a public key establishment protocol that establishes a shared symmetric secret key. There is no key pair.
I upgraded a fedora box two days ago by Googling for the right commands and typing them in. I then typed shutdown -r now. It couldn't talk to init or something like that and refused. So I tried typing reboot. Nope. A quick Googling revealed systemd to be at the root of this.
SysV init is indeed horrible, but it didn't break shutdown. System programmers need a hippocratic oath.
Many of the other commenters just showed their smug attitude by saying "You're wrong" without bothering to put in the effort to explain the efficient primality test as the reason.
>The only redeeming feature of IBE is that it's so obviously academic wank that the industry has stayed away in droves.
Nope, some of us in industry have a turgid knob for IBE too. It solves specific problems exceedingly well. It provides a way to do key distribution amongst things you control while not having to trust the intervening infrastructure and not having to do as much computation at the endpoints.
The GCHQ M-S scheme has been around for a while. It's a well engineered IBE scheme compared to many of the schemes coming from academia. I certainly wouldn't use it when a third party was the KDC, but that's not what it's for. It was a contender for the key management in some standards that would be very widely deployed, but lost out to more conventional PKI schemes due to people being masochists for using things that have failed consistently in the past.
Common compared to other primes? What an ignorant statement! A proportion of approximately 1/log(x) of the natural numbers less than x is prime. There are well over 2^74000000 primes less than this newly-discovered one, but only 48 of them (assuming none between the previously-discovered one and the new one) are Mersenne primes.
Right, but pick a specific form that you can search along. If you searched along the odd numbers, incrementing by two each time, you would search through more cases than searching through (2^n)-1 increasing n by one each time. (2^n)-1 is a rich vein compared to say (2^n)-3 for incrementing n, or n_(i+1) = (n_i) + 2 for incrementing i.
>In a special class of extremely rare prime numbers known as Mersenne primes.
I understood that Mersenne primes are not rare, they are common compared to primes that don't fit the 2^n -1 form. Hence searching for Mersenne primes is a more efficient way of finding big ones.
That is not true. PCI-DSS set a very low standard for security but the whole purpose is to establish minimum security controls.
The reason they keep getting hacked is, transaction data or credit related personal data is really valuable, and PCI-DSS sets such a low standard.
PCI and the card standards they set and the compliance they regulate is the direct cause of all the insecurity. It all happened on their watch. Setting low standards isn't an excuse, it's the cause. It's their fault.
We would be better served if the job was handed over to IEEE. P802 got screwed the first time around when they asked the government for help and got given WEP by the NSA, but they wised up and have since proven capable of secure and implementable standards to sound engineering standards.
My family has a small business. But I have a day job as a security engineer. I design security circuits and work in cryptographic standards.
So I got exposed to the PCI-DSS specs when I was implementing the point of sale system for my family's business and many of their requirements ran counter to security. They should have concentrated on more specific details of how computers handle personal details and card details.
PCI-DSS, the security standards for payment processing have nothing to do with security. There is a veneer of 'we are doing this for security', but none of it makes sense. This is why we keep seeing PCI-DSS compliant systems getting hacked and revealing card and personal details by the million.
It's a pretty good assumption. The odds are in favour of there not being some magic afterlife, on the grounds that there has been not one shred of evidence to suggest such a thing might exist.
Slashdot Mirror
Logarithmic TV FTW !
I'll see your logarithmic and raise you an exponential TV.
Back when 2Mbits was fast, I used to hook subsystems together with 5V RS232esque serial. The interface chips generally supported a synchronous mode (I.E. same byzantine start+data+parity+stop frame format, but with a clock line). So hook up the clock line, and set it to 2Mbit/s. You could multi drop it to and talk to multiple boards in a rack with a little added protocol.
Async RS232 always seemed like a poor tradeoff. A bucket load of signals (DTR, DCD, DSR, RI, RTS,CTS, RTR) in addition to the important ones (tx, rx, gnd, chassis). Then there was a bucket load of complexity with oversampling receivers to sync to the async frame. Why couldn't two of the bucket-o-signals have been a source clock? Let the modem handle the async complexity and send the recovered clock on the wire.
Anyway, to the point or your post, When anything fast or busy was happening I always found it to be a good plan to stick a micro behind the ACIA and handling buffering as it's only job. You couldn't trust OTS hardware to do it well.
If you aren't designing your own PCBs, this might not be an option.
Why is an FTDI serial driver needed? USB has had a serial port protocol as part of its base spec and Windows has a default driver for things declaring themselves to be a serial port. I have several devices that work in this manner.
Why would a vendor of a basic USB-Serial port converter bother writing a driver?
The process was fast, professional, and like nothing ever seen before.
What a complete load of crap.
Indeed. They got caught.
what he meant was that the apple computers he has bought since have all failed within 6 years.
seriously, the rate the mobos go bad should be alarming. who cares if it's encased in aluminium if it just breaks some subsystem or another when it wants to. just today hd controller broke on one over here.. solution? buy a new mac of course.
It's the batteries that go bad first. But that is common to all laptops. The choice of replacement is based on other factors.
Curing blindness sounds admirable - but at what cost?
Don't forget the unseen side effects!
The first Google hit on CRISPR side effects is:
http://phys.org/news/2015-10-c...
Without a suitable blindness cure, the side effects will remain literally unseen.
The number 5 key on my current desktop gaming PC just failed on me today. But WASD still works, so I'll be fine.
I fully expect to replace my laptop more frequently than every 6 years.
The only computer I genuinely used as a primary machine for more than 6 years was the Apple ][+ which lasted for 10. Things have changed since then.
With a paper-based OTP they are probably more secure than ANYTHING you can buy on the open market.
Wow. The least practical key management scheme is the most secure? Let's hope no one manages to recover the pad.
I know I was stoned to death by my parents when I left the Catholic Church.
What doesn't kill you makes you stronger. Oh wait..
Islam does not work well with Freedom of choice. There's only one way, Conformity.
It started as a political movement, and so it continues. Religion is more economical for control than sheer force.
People are born without religion, 99.999% is adopted from their parents. There is no choice involved.
The parents are always right.
The methodist doctrine didn't stick in my case. Are the ratios the same when you aren't under threat of death if you choose atheism? Or am I one of the 0.001%?
No. It's interesting.
Laser printers were infecting and reinfecting 68K macintoshes over the network when I was in college in 1990.
2) iMessage which handle text messaging and FaceTime, encrypts "End to End". The "End to End" encryption generates an public private key pair where the private key never leaves the senders phone and which can't be determined by Apple engineers even if Apple has thousands of years and unlimited CPU cycles. "End to End" scares lawenforcement at the very highest and lowest levels. iMessage appears to be U breakable, to such a degree that Apple's lawyers feel perfectly comfortable walking into any court anywhere in the world and telling the judges, sorry, can't help you. Stunning is all I can say.
You're envisaging RSA when you say "public private key pair". Given what we know, it's probably Diffie Hellman, which is a public key establishment protocol that establishes a shared symmetric secret key. There is no key pair.
True story:
I upgraded a fedora box two days ago by Googling for the right commands and typing them in. I then typed shutdown -r now. It couldn't talk to init or something like that and refused. So I tried typing reboot. Nope. A quick Googling revealed systemd to be at the root of this.
SysV init is indeed horrible, but it didn't break shutdown. System programmers need a hippocratic oath.
Argh. My html got messed up. For shame! For shame!
Learning something by making stuff up in a forum and waiting for people to correct it has a high social cost. Please find another way to learn.
There's only a high social cost if you care what other people think. I had confused that series with . An AC did eventually explain the reason for searching for Mersenne primes being the efficiency of the primality test. I'm very familiar with primality tests for key search algorithms, which need to select uniformly amongst primes for the security of the keys to be maintained, because I'm an implementor of crypto.
Many of the other commenters just showed their smug attitude by saying "You're wrong" without bothering to put in the effort to explain the efficient primality test as the reason.
Thank you. That makes sense. I learned something.
>The only redeeming feature of IBE is that it's so obviously academic wank that the industry has stayed away in droves.
Nope, some of us in industry have a turgid knob for IBE too. It solves specific problems exceedingly well. It provides a way to do key distribution amongst things you control while not having to trust the intervening infrastructure and not having to do as much computation at the endpoints.
The GCHQ M-S scheme has been around for a while. It's a well engineered IBE scheme compared to many of the schemes coming from academia. I certainly wouldn't use it when a third party was the KDC, but that's not what it's for. It was a contender for the key management in some standards that would be very widely deployed, but lost out to more conventional PKI schemes due to people being masochists for using things that have failed consistently in the past.
Common compared to other primes? What an ignorant statement!
A proportion of approximately 1/log(x) of the natural numbers less than x is prime.
There are well over 2^74000000 primes less than this newly-discovered one, but only 48 of them (assuming none between the previously-discovered one and the new one) are Mersenne primes.
Right, but pick a specific form that you can search along. If you searched along the odd numbers, incrementing by two each time, you would search through more cases than searching through (2^n)-1 increasing n by one each time. (2^n)-1 is a rich vein compared to say (2^n)-3 for incrementing n, or n_(i+1) = (n_i) + 2 for incrementing i.
>In a special class of extremely rare prime numbers known as Mersenne primes.
I understood that Mersenne primes are not rare, they are common compared to primes that don't fit the 2^n -1 form. Hence searching for Mersenne primes is a more efficient way of finding big ones.
That is not true. PCI-DSS set a very low standard for security but the whole purpose is to establish minimum security controls.
The reason they keep getting hacked is, transaction data or credit related personal data is really valuable, and PCI-DSS sets such a low standard.
PCI and the card standards they set and the compliance they regulate is the direct cause of all the insecurity. It all happened on their watch.
Setting low standards isn't an excuse, it's the cause. It's their fault.
We would be better served if the job was handed over to IEEE. P802 got screwed the first time around when they asked the government for help and got given WEP by the NSA, but they wised up and have since proven capable of secure and implementable standards to sound engineering standards.
My family has a small business. But I have a day job as a security engineer. I design security circuits and work in cryptographic standards.
So I got exposed to the PCI-DSS specs when I was implementing the point of sale system for my family's business and many of their requirements ran counter to security. They should have concentrated on more specific details of how computers handle personal details and card details.
>PCI (Payment Card Industry)-compliant servers
PCI-DSS, the security standards for payment processing have nothing to do with security. There is a veneer of 'we are doing this for security', but none of it makes sense. This is why we keep seeing PCI-DSS compliant systems getting hacked and revealing card and personal details by the million.
How do you know?
It's a pretty good assumption. The odds are in favour of there not being some magic afterlife, on the grounds that there has been not one shred of evidence to suggest such a thing might exist.