Umm, pray tell, what has heating up (using microwaves) got to do with ionization? Radiation sickness happens because key elements of cells get damaged. The chemistry of our cells is not designed to deal with intense interference with structures of various complex molecules. Ionizing radiation is called such because it can dissociate molecules into ions - that's a rather significant alteration, the original molecule ceases to exist.
The sunburn is a very specific reaction to DNA damage and only DNA damage. The DNA is sensitive to non-ionizing UV-B radiation; new chemical bonds are formed when a photon is absorbed and doesn't get converted to heat. This happens without ionization having occurred. The DNA is simply, by design, susceptible to this kind of radiation, and the direct damage mechanism of DNA is a part of a large family of photochemical reactions. In terms of sensitivity per photon, the DNA is about 2 orders of magnitude less sensitive (~0.1% eff.) than the silver halide you'd find in a film emulsion (~20% eff. undoped).
How about the idea that having a bunch of lame-ass mooches, trolls, and flamers causing nothing but drama increases the stress level of developers and causes them to abandon projects entirely?
It's not that hard just to ignore them. Heck, I'd say that online it's only so much easier to ignore them than in face-to-face situations. Censorship is a slippery slope. He should know better.
I'm talking here only of stupid firmware/software. I didn't let myself suffer long enough for the hardware aging to come into play. After a few months I've had enough.
If the system can't be taken down for maintenance, in pieces if necessary and with redundance if necessary, then the initial design was incompetent.
This, a thousand times this!! Google updates their systems constantly, constantly deploying both new hardware and new software. Somehow we can google things without seeing a "down for maintenance, we've got 5000 storage boxes to upgrade" page. And I'm pretty damn sure that whatever infrastructure google runs their search engine on would make even a large SAP deployment something to laugh at.
Nope, the learning curve is not steep! If it were steep, it'd mean you can learn very quickly. Learning curve is not a hill for crying out loud. Steep hills are hard, steep learning curves mean easy (fast) learning!
The screen-oriented workflow that is pretty much enshrined into the SAP Basis runtime's design is something that puts SAP firmly in the 20th century, usability-wise. There's no sane way to retrofit a SAP system into a modern, object-oriented UI. By object orientation I mean almost anything you use today: the file shell (explorer, finder,...), any "editing" application where you manipulate objects (vector drawings in office suites or illustration packages, modern CAD,...). In a usual deployment of a system like SAP, you can't, say drag the PO you're working on to a "desktop" to keep it there for easy reference, you can't tag things, there's no object-agnostic history of what you've been doing, etc. SAP is really just a green-screen-oriented design that keeps getting shoehorned into modern presentations, but the basic workflow is well understood to be nightmarish from the human efficiency standpoint.
This is very insightful. SAP's biggest advantage (integration) is its biggest disadvantage, too. In software engineering, tight coupling in such a big system has been repeatedly shown to lead to disasters. By any reasonable metric of software structure, SAP is a disaster.
You have the right gut feeling. The card is not flaky, it's a quite reliable and standardized piece of hardware. It's the crypto infrastructure that's built around it that sucks. One of the reasons it sucks is because, due to its proprietary and closed nature, it's hard to learn about it. So even people who work for the cable company simply don't understand what they're really doing, and there's much voodoo involved - voodoo that's a stand-in for knowledge that's inaccessible without you signing an NDA...
The EE in me finds the "multiple tuner" part to be perhaps the most ass-backwards one. I'd hope they do the tuning digitally and there's just one wideband IF signal chain that feeds into an ADC, with the tuning done inside of an all-digital ASIC. Demodulating QAM in an ASIC is really cheap.
I've had their 2nd generation box and it was a joke in terms of reliability and UI ergonomics. At home we have a couple of eyeTVs hooked up to an iMac and it works just fine.
In a WiFi solution the wake-up packet filtering can be done in the radio hardware, the microcontroller can be asleep. A receive radio subsystem doesn't consume all that much power.
Powerline ethernet costs way more than WiFi. You need a whole bunch of components to interface with the power line - components that don't exist in WiFi or Bluetooth solutions. These days the entire radio is integrated on one chip with a microcontroller. There's usually about 10 external components needed to get it on the air if you assume that suitable power supply voltage is available.
You're completely detached from real economics of the thing.
WiFi in the socket means that you need to replace the whole light fixture, as the entire thing has to be UL listed as a unit, you can't just swap out a socket for something else. That usually costs big money: a permit, a licensed electrician to do it, materials, etc. A light bulb has a much smaller regulatory footprint, and you're bound to sell it in much higher volume than light fixtures.
The electronics of course don't give a flying fuck about the heat, because in such a small unit there's no parts that degrade over time with heat, it's as simple as that - at least as long as you're talking about what adding WiFi would entail. I don't know if you realize what a modern WiFi solution looks like. It's a chip, a couple tiny external components, and an antenna. You'd be swapping the existing power supply controller chip for a WiFi chip. Chip count stays the same. I'm sure that at a small added cost you could get rid of the electrolytics in the bulb's power supply as well. That way you end up with a design that has can be made to work "forever".
Again, the cost of the chip is nothing compared to the cost of changing anything more than a lightbulb. It's basically light bulb or nothing.
X10 is almost useless in a modern home where almost anything you plug into the outlet has a switching power supply, including every light bulb. It wasn't designed to cope with that. There are much better and higher bandwidth protocols. They leverage modern signal processing. It lets them perform better than X10 in spite of being orders of magnitude higher bandwidth!
I think it won't be too long till the added cost of Wi-Fi in a bulb will be a couple of USD. Remember: a Wi-Fi chip has a microcontroller inside of it. That microcontroller should be enough to run Wi-Fi and a simple mesh network. It doesn't need a full-blown webserver, but even that could be done on a micro. The volume lets you optimize the heck out of everything. It would cost $0.0 in materials to have this chip control the light that already needs to have a power supply built into it anyway. In fact, the added on-top-of-Wi-Fi chip cost might even be negative if they let the Wi-Fi chip do power regulation in firmware as well. That way you lose the voltage regulation silicon, so that may cut some cost. Why the heck not, every modern industrial servo drive does all of the control in software anyway, and it's a couple control loops worth more complex than a power supply:)
The problem is that Snowden actually HAS proof to back up his claim
So far the government tries to weasel out pretending like it isn't true, and, last time I checked, "releasing" fiction doesn't get you in legal hot water. So far we only have Snowden's word that the "proof" (the documents) are true. Further government action will be required for us to have some substantiation as to whether it's fiction or fact. So far everything looks like yes, the documents are true, and Snowden is certainly in hot water, but the government is in orders of magnitude more hot water.
he broke several laws in releasing this information
The point is: if the information is made up, he broke no laws. Writing fiction is writing fiction, plenty of people do it publicly and get paid for it, even! If the information is true, though, then whatever laws he broke doesn't take the government out of hot water. They are admitting to the accuracy of the information by bringing charges against the guy. That makes the so-far only alleged data collection practices suddenly undeniable. In fact, they must be true for any charges to be with merit, the government has no choice but to admit to the truth of the information in order to bring charges.
The government had no way of covering it up even if they wished to. As I've said, they were in the hole many times their worth. At least in the almost-failed Euro countries, you've had debt on the order of GDP. In Iceland, GDP was a joke compared to what was lost.
Slashdot Mirror
Umm, pray tell, what has heating up (using microwaves) got to do with ionization? Radiation sickness happens because key elements of cells get damaged. The chemistry of our cells is not designed to deal with intense interference with structures of various complex molecules. Ionizing radiation is called such because it can dissociate molecules into ions - that's a rather significant alteration, the original molecule ceases to exist.
The sunburn is a very specific reaction to DNA damage and only DNA damage. The DNA is sensitive to non-ionizing UV-B radiation; new chemical bonds are formed when a photon is absorbed and doesn't get converted to heat. This happens without ionization having occurred. The DNA is simply, by design, susceptible to this kind of radiation, and the direct damage mechanism of DNA is a part of a large family of photochemical reactions. In terms of sensitivity per photon, the DNA is about 2 orders of magnitude less sensitive (~0.1% eff.) than the silver halide you'd find in a film emulsion (~20% eff. undoped).
How about the idea that having a bunch of lame-ass mooches, trolls, and flamers causing nothing but drama increases the stress level of developers and causes them to abandon projects entirely?
It's not that hard just to ignore them. Heck, I'd say that online it's only so much easier to ignore them than in face-to-face situations. Censorship is a slippery slope. He should know better.
I'm talking here only of stupid firmware/software. I didn't let myself suffer long enough for the hardware aging to come into play. After a few months I've had enough.
If the system can't be taken down for maintenance, in pieces if necessary and with redundance if necessary, then the initial design was incompetent.
This, a thousand times this!! Google updates their systems constantly, constantly deploying both new hardware and new software. Somehow we can google things without seeing a "down for maintenance, we've got 5000 storage boxes to upgrade" page. And I'm pretty damn sure that whatever infrastructure google runs their search engine on would make even a large SAP deployment something to laugh at.
Nope, the learning curve is not steep! If it were steep, it'd mean you can learn very quickly. Learning curve is not a hill for crying out loud. Steep hills are hard, steep learning curves mean easy (fast) learning!
You need to know the OS you're installing on though, or be prepared to learn it as you go.
?! I think you'd be crazy to install anything as complex and under-documented as SAP BASIS without really knowing your OS!!
The screen-oriented workflow that is pretty much enshrined into the SAP Basis runtime's design is something that puts SAP firmly in the 20th century, usability-wise. There's no sane way to retrofit a SAP system into a modern, object-oriented UI. By object orientation I mean almost anything you use today: the file shell (explorer, finder, ...), any "editing" application where you manipulate objects (vector drawings in office suites or illustration packages, modern CAD, ...). In a usual deployment of a system like SAP, you can't, say drag the PO you're working on to a "desktop" to keep it there for easy reference, you can't tag things, there's no object-agnostic history of what you've been doing, etc. SAP is really just a green-screen-oriented design that keeps getting shoehorned into modern presentations, but the basic workflow is well understood to be nightmarish from the human efficiency standpoint.
This is very insightful. SAP's biggest advantage (integration) is its biggest disadvantage, too. In software engineering, tight coupling in such a big system has been repeatedly shown to lead to disasters. By any reasonable metric of software structure, SAP is a disaster.
I don't know about their internal architecture, but I do agree that the UI on my 2nd generation was absolutely horrendous.
You have the right gut feeling. The card is not flaky, it's a quite reliable and standardized piece of hardware. It's the crypto infrastructure that's built around it that sucks. One of the reasons it sucks is because, due to its proprietary and closed nature, it's hard to learn about it. So even people who work for the cable company simply don't understand what they're really doing, and there's much voodoo involved - voodoo that's a stand-in for knowledge that's inaccessible without you signing an NDA...
The EE in me finds the "multiple tuner" part to be perhaps the most ass-backwards one. I'd hope they do the tuning digitally and there's just one wideband IF signal chain that feeds into an ADC, with the tuning done inside of an all-digital ASIC. Demodulating QAM in an ASIC is really cheap.
I've had their 2nd generation box and it was a joke in terms of reliability and UI ergonomics. At home we have a couple of eyeTVs hooked up to an iMac and it works just fine.
There's a long way between the antenna and the microcontroller :)
But, arguably, those light bulbs, due to volume, will have the most vulnerability-scrutinized software stack ever :)
I've recently bought some 70W-equivalent LEDs and they weren't very heavy. Our desktop lamp didn't complain.
In a WiFi solution the wake-up packet filtering can be done in the radio hardware, the microcontroller can be asleep. A receive radio subsystem doesn't consume all that much power.
Powerline ethernet costs way more than WiFi. You need a whole bunch of components to interface with the power line - components that don't exist in WiFi or Bluetooth solutions. These days the entire radio is integrated on one chip with a microcontroller. There's usually about 10 external components needed to get it on the air if you assume that suitable power supply voltage is available.
You're completely detached from real economics of the thing.
WiFi in the socket means that you need to replace the whole light fixture, as the entire thing has to be UL listed as a unit, you can't just swap out a socket for something else. That usually costs big money: a permit, a licensed electrician to do it, materials, etc. A light bulb has a much smaller regulatory footprint, and you're bound to sell it in much higher volume than light fixtures.
The electronics of course don't give a flying fuck about the heat, because in such a small unit there's no parts that degrade over time with heat, it's as simple as that - at least as long as you're talking about what adding WiFi would entail. I don't know if you realize what a modern WiFi solution looks like. It's a chip, a couple tiny external components, and an antenna. You'd be swapping the existing power supply controller chip for a WiFi chip. Chip count stays the same. I'm sure that at a small added cost you could get rid of the electrolytics in the bulb's power supply as well. That way you end up with a design that has can be made to work "forever".
Again, the cost of the chip is nothing compared to the cost of changing anything more than a lightbulb. It's basically light bulb or nothing.
Too heavy? What the fuck?
X10 is almost useless in a modern home where almost anything you plug into the outlet has a switching power supply, including every light bulb. It wasn't designed to cope with that. There are much better and higher bandwidth protocols. They leverage modern signal processing. It lets them perform better than X10 in spite of being orders of magnitude higher bandwidth!
I think it won't be too long till the added cost of Wi-Fi in a bulb will be a couple of USD. Remember: a Wi-Fi chip has a microcontroller inside of it. That microcontroller should be enough to run Wi-Fi and a simple mesh network. It doesn't need a full-blown webserver, but even that could be done on a micro. The volume lets you optimize the heck out of everything. It would cost $0.0 in materials to have this chip control the light that already needs to have a power supply built into it anyway. In fact, the added on-top-of-Wi-Fi chip cost might even be negative if they let the Wi-Fi chip do power regulation in firmware as well. That way you lose the voltage regulation silicon, so that may cut some cost. Why the heck not, every modern industrial servo drive does all of the control in software anyway, and it's a couple control loops worth more complex than a power supply :)
The problem is that Snowden actually HAS proof to back up his claim
So far the government tries to weasel out pretending like it isn't true, and, last time I checked, "releasing" fiction doesn't get you in legal hot water. So far we only have Snowden's word that the "proof" (the documents) are true. Further government action will be required for us to have some substantiation as to whether it's fiction or fact. So far everything looks like yes, the documents are true, and Snowden is certainly in hot water, but the government is in orders of magnitude more hot water.
he broke several laws in releasing this information
The point is: if the information is made up, he broke no laws. Writing fiction is writing fiction, plenty of people do it publicly and get paid for it, even! If the information is true, though, then whatever laws he broke doesn't take the government out of hot water. They are admitting to the accuracy of the information by bringing charges against the guy. That makes the so-far only alleged data collection practices suddenly undeniable. In fact, they must be true for any charges to be with merit, the government has no choice but to admit to the truth of the information in order to bring charges.
The government had no way of covering it up even if they wished to. As I've said, they were in the hole many times their worth. At least in the almost-failed Euro countries, you've had debt on the order of GDP. In Iceland, GDP was a joke compared to what was lost.
It wasn't exceptional evil, just wholly everyday stupidity, concentrated in one spot.