Why the hell do you presume 50% efficiency? That's absolutely atrocious. We can get 85% from ATX switchmode supplies, and those are so under spec'ed that the current folk wisdom recommends supplies rated at twice what any sensible single-socket pc consumes, not to mention the economies of scale.
Furthermore, your estimate of the amount of power needed to remove the heat is incredibly pessimistic. Think of all the pad mount transformers around large buildings. Those are 800-2000 kVA, and every single one I've seen has been passively cooled. Thermal management in large-scale power electronics systems is not a new problem, and is a well-developed industry in itself. The only place heat dissipation might be a concern is in the batteries themselves. This is, of course, the storage efficiency: the one consideration you decided to neglect!
Considering the charging problem, the obvious approach is to charge in series and discharge in parallel. Assuming a single drive motor, a plurality of prismatic Li-ion cells could be connected in 3*N stacks of 400 Volts or so (600V IGBTs). Each phase of the motor could then be driven by N phase legs (like a class B amplifier) with their outputs combined in parallel. Current sharing is insured in the short term by high-frequency chokes, and in the long term by using current-mode control in the phase legs. The current-sharing ratio could also be actively controlled to ensure equal discharge of the batteries.
Alternatively, a two-stage approach is used, in which the several battery segments feed separate phases of a polyphase boost converter, supplying a regulated bus voltage to a traditional voltage source three phase inverter. This provides more fine-grained control of the battery charge state and the advantages of a tested design, at the expense of efficiency and a substantially greater component count.
For charging, the battery pack itself would consist of prismatic cells with large contacts on opposite faces. Spaces between the cells would allow charging electrodes to be inserted. These electrodes would consist of brass plates sandwiching a piece of foam to provide contact pressure. Said plates could incorporate coolant channels through which heat could be removed from the cells.
Anyone who doesn't have at least 30-40 tabs open simultaneously just isn't using the internet properly. I have 4 Firefox windows open at all times. One for fanfiction that I occasionally refresh and check for updates, one for porn, one for casual browsing, and one connected through Tor. I use Firefox precisely because it preserves my tabs through updates, reboots, and all manner system crashes. Noscript and downthemall are nice too.
As I said above, more battery space. Or a second HDD, for RAID 1 or more space. I've always thought laptops should have a very small secondary battery, so you could swap batteries without shutting down.
It is thus obvious that this rule should be repealed, and, furthermore, "Take Me Out to the Ball Game" should be replaced in all occurrences with "Yakety Sax".
64 GB SSD for OS and frequently used programs + file server in basement with 20 TB RAID6 of spinny disks. If you choose low-heat components and big slow fans, it could be almost completely silent.
There really isn't much contention for space on the back of a desktop case. The fact of the matter is you can bitbang most any low speed protocol there is with a parallel port and a few resistors.
I agree, however, with the removal of optical drives. Why the hell do I need a DVD writer on my laptop? That thing probably consumes 15% of the internal volume of the case! Shouldn't large, infrequently used mechanical components be banished to the desk? Perhaps we could repurpose that space and have laptops with decent battery life and no testicle hanging from the bottom.
Microwave ovens operate about 30 dB above the most powerful wifi equipment, and that Faraday cage isn't perfect. Furthermore, the legal limit is 1 mW*cm^(-2) 5 cm from the oven at time of manufacture. Worst case, that could be more than 10 W, which would swamp your wifi.
Um, because there's more mixed case strings than single case strings. So hackers have to search through more strings to find your password.
A 10 character random single case alphanumeric string has more entropy than an 8 character mixed case alphanumeric string. In the human mind, the distinction between A and a is less than the distinction between a and b. Therefore, case is an inefficient way to store entropy when a human has to remember the password. The same is true of special characters; people can't remember them well, and they are more difficult to write down unambiguously.
Except they're not equivalent strength if the English text is just slightly longer. English utterances have about 1.5 bits of entropy per character. This is even lower if you restrict yourself to lowercase. Arbitrary strings have 8 bits per character.
I'm talking about random strings, not English. I get about 5.16 bits of entropy per character for single case alphanumeric.
If you use mixed case and special characters, you also have to worry about ambiguities such as I/l/1/|./, "/'':/; and 0/O/o. The fact of the matter is that humans remember and distinguish in print some characters better than others. As such, it is best to restrict the character set used for passwords to optimize entropy/effort.
Just don't use special characters in your passwords. Don't use mixed case either, unless it is mandated (WHY do they do this?). It is much easier to remember or write down a slightly longer single case alphanumeric password than a password with mixed case and special characters of equivalent strength.
Why the hell did you buy that? There's no way it's portable, the screen probably has poor color rendition, and I suspect you could have gotten a desktop with 30% more oomph for the same price. Who even builds a laptop with three hard disks?
No. Reading speed is greatest for line lengths that require no horizontal eye movement. That is, short enough that an entire line can be in focus at once. The two or three column formats used in print are a prime example of this. Multiple columns do not work well on computers however, because scrolling is not monotonic unless an entire column can fit on the screen.
Slashdot Mirror
Mathematics fail.
Why the hell do you presume 50% efficiency? That's absolutely atrocious. We can get 85% from ATX switchmode supplies, and those are so under spec'ed that the current folk wisdom recommends supplies rated at twice what any sensible single-socket pc consumes, not to mention the economies of scale.
Furthermore, your estimate of the amount of power needed to remove the heat is incredibly pessimistic. Think of all the pad mount transformers around large buildings. Those are 800-2000 kVA, and every single one I've seen has been passively cooled. Thermal management in large-scale power electronics systems is not a new problem, and is a well-developed industry in itself. The only place heat dissipation might be a concern is in the batteries themselves. This is, of course, the storage efficiency: the one consideration you decided to neglect!
Considering the charging problem, the obvious approach is to charge in series and discharge in parallel. Assuming a single drive motor, a plurality of prismatic Li-ion cells could be connected in 3*N stacks of 400 Volts or so (600V IGBTs). Each phase of the motor could then be driven by N phase legs (like a class B amplifier) with their outputs combined in parallel. Current sharing is insured in the short term by high-frequency chokes, and in the long term by using current-mode control in the phase legs. The current-sharing ratio could also be actively controlled to ensure equal discharge of the batteries.
Alternatively, a two-stage approach is used, in which the several battery segments feed separate phases of a polyphase boost converter, supplying a regulated bus voltage to a traditional voltage source three phase inverter. This provides more fine-grained control of the battery charge state and the advantages of a tested design, at the expense of efficiency and a substantially greater component count.
For charging, the battery pack itself would consist of prismatic cells with large contacts on opposite faces. Spaces between the cells would allow charging electrodes to be inserted. These electrodes would consist of brass plates sandwiching a piece of foam to provide contact pressure. Said plates could incorporate coolant channels through which heat could be removed from the cells.
Thirty? Pish posh. Right now I have 281 tabs in 581 MiB of RAM.
Perhaps Windows can run in 40 MiB, but can the trojans?
Because it's hard to fit a liquid helium cooling system on an expansion card.
I liked the brown. The new theme is all black plastic and blue LEDs.
Anyone who doesn't have at least 30-40 tabs open simultaneously just isn't using the internet properly. I have 4 Firefox windows open at all times. One for fanfiction that I occasionally refresh and check for updates, one for porn, one for casual browsing, and one connected through Tor. I use Firefox precisely because it preserves my tabs through updates, reboots, and all manner system crashes. Noscript and downthemall are nice too.
I prefer to slime them myself, thank you very much.
As I said above, more battery space. Or a second HDD, for RAID 1 or more space. I've always thought laptops should have a very small secondary battery, so you could swap batteries without shutting down.
It is thus obvious that this rule should be repealed, and, furthermore, "Take Me Out to the Ball Game" should be replaced in all occurrences with "Yakety Sax".
My sister has a functioning PC that boots from a 6 GiB Quantum Fireball. It runs Xubuntu.
64 GB SSD for OS and frequently used programs + file server in basement with 20 TB RAID6 of spinny disks. If you choose low-heat components and big slow fans, it could be almost completely silent.
My kingdom for a modpoint. Alas, I already posted.
There really isn't much contention for space on the back of a desktop case. The fact of the matter is you can bitbang most any low speed protocol there is with a parallel port and a few resistors.
I agree, however, with the removal of optical drives. Why the hell do I need a DVD writer on my laptop? That thing probably consumes 15% of the internal volume of the case! Shouldn't large, infrequently used mechanical components be banished to the desk? Perhaps we could repurpose that space and have laptops with decent battery life and no testicle hanging from the bottom.
What's wrong with BNC connectors?
You're right. Capsaicin would work much better. Oh wait...
Microwave ovens operate about 30 dB above the most powerful wifi equipment, and that Faraday cage isn't perfect. Furthermore, the legal limit is 1 mW*cm^(-2) 5 cm from the oven at time of manufacture. Worst case, that could be more than 10 W, which would swamp your wifi.
Good job Freud.
How is that superior to cosine_radians(3.14159)?
Um, because there's more mixed case strings than single case strings. So hackers have to search through more strings to find your password.
A 10 character random single case alphanumeric string has more entropy than an 8 character mixed case alphanumeric string. In the human mind, the distinction between A and a is less than the distinction between a and b. Therefore, case is an inefficient way to store entropy when a human has to remember the password. The same is true of special characters; people can't remember them well, and they are more difficult to write down unambiguously.
Except they're not equivalent strength if the English text is just slightly longer. English utterances have about 1.5 bits of entropy per character. This is even lower if you restrict yourself to lowercase. Arbitrary strings have 8 bits per character.
I'm talking about random strings, not English. I get about 5.16 bits of entropy per character for single case alphanumeric.
If you use mixed case and special characters, you also have to worry about ambiguities such as I/l/1/| ./, "/'' :/; and 0/O/o. The fact of the matter is that humans remember and distinguish in print some characters better than others. As such, it is best to restrict the character set used for passwords to optimize entropy/effort.
Just don't use special characters in your passwords. Don't use mixed case either, unless it is mandated (WHY do they do this?). It is much easier to remember or write down a slightly longer single case alphanumeric password than a password with mixed case and special characters of equivalent strength.
So Apple finally gets a taste of their own medicine.
You keep using that word. I do not think it means what you think it means.
Why the hell did you buy that? There's no way it's portable, the screen probably has poor color rendition, and I suspect you could have gotten a desktop with 30% more oomph for the same price. Who even builds a laptop with three hard disks?
No. Reading speed is greatest for line lengths that require no horizontal eye movement. That is, short enough that an entire line can be in focus at once. The two or three column formats used in print are a prime example of this. Multiple columns do not work well on computers however, because scrolling is not monotonic unless an entire column can fit on the screen.