The difference is that Theo has acted in a way in the past that has caused us to route all communications from him directly to the lawyers. It's not to do with divulging secrets. It's to do with past behavior.
True. The difference is that if a NetBSD developer emailed me to ask about using RdRand in the kernel (A thing I would know about) I would happily enter into a technical discussion and help them out. If Theo emailed, I would have to refer the email to the lawyers.
If Theo hadn't systematically pissed off everyone in large corporations that he's come in contact with, they might have written some drivers. Linus is pragmatic, manages a team of experts well and the so the corporations are happy to work with him.
I thought that little used operating systems were less vulnerable because fewer hackers would target them compared to popular, mass market operating systems such as Linux and MacOS.
Read the PCI-DSS specifications. They will tell you what the card processors want vendors to adhere to. However being compliant involves ticking the yes box on the "Yes I am Compliant" tick box on the PCI web site.
That could be relevant. It would be easy to test. Hook up a bench power supply to the rails of the circuit board. See if it works better than with the wall wart.
>even the ARM holdings group has been talking about having "dark silicon"
'Dark silicon' is a stupid idea. Going down that route is just evidence that you're too stupid to do power management correctly on chip.
BPM (Bad Power Management): Have a power controller that turns things on and off as they are needed. GPM (Good Power Management): On chip units know when they are needed and turn themselves on and off as required.
You simply cannot get BPM to scale with complexity. You hit a complexity wall because a central command entity can't know the details of supply and demand between every entity on chip. So behaviors are wrapped up into 'power states' and OSs, drivers and applications get dumped with dealing with power states and bugs aplenty arise.
GPM just requires than individual entities ensure that they are on when they are needed. It's a relatively easy task to identify the things an entity uses and the things than use an entity. Tie the power switches to entities a this level and power management happens from the bottom up. Nothing in left on necessarily.
Dark silicon is just an aspect of BPM, whereby you duplicate functions or add specialized functions (think CISC) that aren't used all the time, so they can be turned off when not used and so the power density across the surface of the chip goes down. If power density is a problem, modulate the clock with temperature rather than doing bloody stupid things like dark silicon. It is an admission that you have more transistors available to you than you know how to use well.
>cheap electroytic capacitors have a tendency to degrade and fail over time.
Not significantly over 2 years and you don't use electrolytics in the in IF/RF signal path in a 2.4 & 5.8GHz radios. I don't think electrolytics are it.
I have my suspicions about the noise figure of LNAs changing over time. There are some very highly strung, teeny weeny transistors in LNAs (Low Noise Amplifiers) right in the signal path.
During my career electronic product design it has generally been used to describe flash memory during the many times that supply has gone on allocation.
There is an appropriate engineering approach to making a robust system. Being robust both legally and technically is eminently achievable. Especially for the money that medical devices cost.
Slashdot Mirror
>And has the power to tax...
No, to license.
You don't have to pay if you don't have a telly.
Or sell off the 'kernel development' part of the operation, and call it 'Flagilent'
If HP are in charge, it must be time to fire 20% of the Linux Foundation workforce.
802.1X is an essential security feature?!
How did we survive before EAPoL?
That was not the nature of the exchange as I remember it.
The difference is that Theo has acted in a way in the past that has caused us to route all communications from him directly to the lawyers. It's not to do with divulging secrets. It's to do with past behavior.
True. The difference is that if a NetBSD developer emailed me to ask about using RdRand in the kernel (A thing I would know about) I would happily enter into a technical discussion and help them out. If Theo emailed, I would have to refer the email to the lawyers.
The moral here maybe that if you're starting a new software product you have to put equal attention into these two things.
Software? I design cryptographic hardware for a living you insensitive clod!
If Theo hadn't systematically pissed off everyone in large corporations that he's come in contact with, they might have written some drivers.
Linus is pragmatic, manages a team of experts well and the so the corporations are happy to work with him.
I thought that little used operating systems were less vulnerable because fewer hackers would target them compared to popular, mass market operating systems such as Linux and MacOS.
>using almost exclusively standard equipment and materials already needed to make conventional chips.
So that will be a small $5,000,000,000 for a state of the art semiconductor factory then.
http://www.tomshardware.com/news/intel-fab42-14nm-cpu-factory,14545.html
It looks like they've copied my website and are therefore infringing my copyright.
But I won't be suing them because I don't mind, because I'm not Apple.
Read the PCI-DSS specifications. They will tell you what the card processors want vendors to adhere to.
However being compliant involves ticking the yes box on the "Yes I am Compliant" tick box on the PCI web site.
Actual compliance is optional.
You obviously don't know what I do for a living.
As I said in another thread, this would be easy to test with a bench power supply a wall wart.
That could be relevant. It would be easy to test. Hook up a bench power supply to the rails of the circuit board. See if it works better than with the wall wart.
>even the ARM holdings group has been talking about having "dark silicon"
'Dark silicon' is a stupid idea. Going down that route is just evidence that you're too stupid to do power management correctly on chip.
BPM (Bad Power Management): Have a power controller that turns things on and off as they are needed.
GPM (Good Power Management): On chip units know when they are needed and turn themselves on and off as required.
You simply cannot get BPM to scale with complexity. You hit a complexity wall because a central command entity can't know the details of supply and demand between every entity on chip. So behaviors are wrapped up into 'power states' and OSs, drivers and applications get dumped with dealing with power states and bugs aplenty arise.
GPM just requires than individual entities ensure that they are on when they are needed. It's a relatively easy task to identify the things an entity uses and the things than use an entity. Tie the power switches to entities a this level and power management happens from the bottom up. Nothing in left on necessarily.
Dark silicon is just an aspect of BPM, whereby you duplicate functions or add specialized functions (think CISC) that aren't used all the time, so they can be turned off when not used and so the power density across the surface of the chip goes down. If power density is a problem, modulate the clock with temperature rather than doing bloody stupid things like dark silicon. It is an admission that you have more transistors available to you than you know how to use well.
With adaptive rate and power control, I wouldn't suspect the output path first. I would look to the input path first. That's the delicate bit.
>cheap electroytic capacitors have a tendency to degrade and fail over time.
Not significantly over 2 years and you don't use electrolytics in the in IF/RF signal path in a 2.4 & 5.8GHz radios.
I don't think electrolytics are it.
I have my suspicions about the noise figure of LNAs changing over time. There are some very highly strung, teeny weeny transistors in LNAs (Low Noise Amplifiers) right in the signal path.
You're probably right. But I'd get out as soon as it gets there.
During my career electronic product design it has generally been used to describe flash memory during the many times that supply has gone on allocation.
Nah, Willamette Valley. It's like Napa, but you can afford to live there.
Work,
Home,
Work,
Home, Supermarket, Wine bar,
Work,
Home,
Work,
Home,
Home,
Work,
etc.
>magnetic field signature is computed based on the monitored compass output
In what way is this no 'near field communications'
http://en.wikipedia.org/wiki/Near_and_far_field
There is an appropriate engineering approach to making a robust system. Being robust both legally and technically is eminently achievable. Especially for the money that medical devices cost.