I've read through some of the other posts and don't think anyone else has mentioned this yet. I could be wrong.
Has anyone thought about using a standard DC power supply (computer supply, etc.) and then using solid-state voltage regulators to produce the required voltages for different devices?
Let's say you have an ATX power supply that puts out +5, +12, and -12. You might even be able to do it with a big brick that just puts out 12VDC, depending on what regulators you buy. Then you use solid-state DC-DC regulators to produce the specific voltages you need for each device. For the more common voltages you could use off-the-shelf fixed regulators (Radioshack even sells a few), and for the less common ones you can use variable regulators and set the voltage with some resistors and capacitors. An example part is LM2734 from National, although I'm sure there are lots of others.
The obvious downside from this is that the max voltage you'd be able to deliver is slightly over 12V (some regulators can produce Vout higher than Vin, but not by much), and the current is limited by the regulators. Also, they need heatsinking, and waste a fair amount of energy I think. But given that DC power supplies are fairly cheap, this might not be a bad thing.
Of course you'd have to get the ATX supply, if that's what you'd choose to use, to produce power in the first place. Sometimes as other people have noticed, they won't work without a load. I use one on my workbench, and in order to get it to produce power at all I have an ancient 5.25" hard drive that I keep plugged in to one of the outputs as a load. I'm sure I could fake the load with resistors or something if I wanted to, but why bother.
Anyway, just a thought for the enterprising hobbyist or experimenter. I'm pretty sure you could use this to power all your small DC devices like peripherals. For large current-draw devices like computers or displays, this might not be an appropriate solution.
I think there's a fourth category. From personal experience, I know that some particularly, er, animated professors dislike being recorded because they don't want to censor themselves.
There were a few professors that used to end up with quotes on the Daily Jolt website with regularity...I can imagine that a few juicy audio quotes, widely shared, might put a dent in a young professor's career. And this would be unfortunate because some of these people, despite (or perhaps because of) the occasional profanity or humorous exaggeration, are among the best lecturers I've taken class with.
I had a physics professor that used to talk with reverence about a professor that he had in school, who used to not allow students to take notes in class. He felt that it was a 'mental endeavor' and that not having students obsessed with taking notes resulted in a more interactive class experience.
While I'm not advocating this position, I think it's an interesting anecdote. I've certainly found in my personal experience that some of my most memorable classes I have very few paper notes from, because I was either busy participating or just too enraptured with the lecture to write anything down. And while I doubt that the addition of any sort of recording apparatus is ever going to make a boring lecture that much more interesting (heck, I take notes in order to stay awake sometimes), I do wish i had recorded some of the more memorable lectures from my college career.
Nothing really, that I can see. I have an iPod (but no recording attachment, which I guess the Duke kiddies also got for free), but long before I had that, I had an Olympus DS-330 recorder.
It uses some proprietary recording format, but the software will convert it to AIFF as you download it to the computer. Although they now have ones with a lot more recording time, I see no reason to replace it. I suppose if I didn't already own one, I'd probably just get the $30 recording add-on for my iPod instead of a standalone device, although the Olympus is significantly smaller and lighter.
If Duke had distributed these to everyone, I suppose we'd be having the same issues. My point is just that the iPod isn't the first device since the analog mini-cassette recorder to give students the capability to record and share a lecture. And as I said in a post further up, in an environment where laptops were ubiquitous, there are programs that do audio recording and note-taking at the same time, and produce a single document that you can share with others.
If all the students are just taping the course, and the professor is just a tape of a prerecorded lecture...
Couldn't everyone just save a lot of time and just email everyone the stupid lecture? Or for that matter, why bother leaving home at all? Just take it online.
While maybe this might appeal to some/.ers, it's not for me. Personally I'd have been pretty offended if I signed up to take a class and walked in to find a tape deck playing a recorded lecture. I'd drop that class pretty quickly, if I had the option. Now with that said, I have and do record lectures (not with an iPod, but with an Olympus digital recorder) and use them for later review, but I've never recorded anything that I wasn't actually there for. At my college, recording devices haven't become widespread enough for people to start trading recordings (I've only even been asked for a copy of mine once), and the few other people that do use recorders do so mostly in addition to paper notes.
Frankly I think that the new "notebook" document type in MS Office, which combines an audio recording with typed notes and knows where in the recording the notes were taken, is potentially more useful for students than an iPod recorder, because it combines regular notetaking skills with the ability to hear what triggered those notes. And I say this as someone who's not normally a fan of MS products -- it's fairly slick. If you're in an environment where lots of people have laptops and bring them to class, this might have a greater impact in the long run than a bunch of iPods. The impact where I am has been limited, because people don't bring laptops to class very often.
Phone companies work on the assumption -- and it's a good one -- that you want service 24/7/365. Even wireless companies (and I'm no huge fan, trust me) put some thought into taking transmitters on and offline so as not to blackout huge service areas.
I'm not sure I'd want to replace that with a patchwork of small transmitters, each running for their own reasons and on their own schedules, with my phone service basically as an afterthought. This is why I think WiFi is an interesting addon (as you described) to cell phones, but not a replacement for one outside of carefully controlled environments. (Like using wifi to replace cordless phones in a big warehouse store or corporate/educational campus.) I'm not criticizing you at all, but there are some other people who seem to like to propose distributed systems as a replacement for everything that's centralized, and I don't think that's a good idea. From a technological perspective I don't have much to complain about the POTS system -- it works when I need it, it's unobtrusive when I don't. I wish it wasn't so expensive...so that's where the market for VoIP comes in.
Can't you just order blank media from the US and have it shipped to Canada? I assume you'd probably be breaking some law at some point in the process... but it's probably what I'd do if I lived there. How about buying it on eBay? I can't imagine knowing that every CD-R you buy results in some small amount of money going to the copyright whores.
Very interesting thought. I'd mod it up if I had any points and weren't participating in the discussion.
Doesn't combining all of the compute-bound processes together onto one processor by definition almost create a bottleneck? Wouldn't it be better to balance your compute- and input-bound processes, so that when the input-bound processes are waiting for input, both processors are working on the compute-bound ones?
I'm not saying you're wrong, I'm just wondering whether this approach wouldn't create more wasted processor cycles than it prevented. However, even if it did that, if the product was a more responsive GUI and interface off of the same hardware, then I might be all for it, wasted cycles or no. (But this may be because I have a personal obsession with responsiveness, and despise any system that has any sort of perceptible lag.)
Well the reason people are "hung up on" the perceived need to dedicate one core to mundane tasks like virus protection and anti-spyware is because just that was a major focus of the PCMag article. And the underlying assumption of the article was that all Windows users apparently need and/or want virus scanners and spyware scanners running constantly on their systems, to the point where it interferes with normal user processes.
That focus is a product of the article, not the Linux-heads. (And I mean no offense to Linux-heads, either; I'm writing this using Konqueror.)
According to the article linked to higher in the thread, the upcoming IBM chip will have two cores, each with a separate cache. It also opined that to the OS, it will appear as if the machine has two processors. (Actually the example was discussing two multicored procs appearing as four processors, but you get the idea.)
Maybe somebody can explain this to me... how from a usage standpoint is a dual-core machine any different than a dual-processor one? Obviously from a design standpoint it's much nicer to have one chip with two cores rather than two separate chips, especially in terms of cost, but does a dual-core processor ACT any differently to the software than two separate processors would?
And given that dual-proc machines have been around for a while now, why is there so much being said about possible uses of dual-core machines? Wouldn't the "uses" for the second core be exactly the same as the "uses" for a second processor in a dual-proc setup? I'm not sure I understand why we're getting all the excitement, except that it might make the benefits of dual-processing more available to the masses in time, and allow current dual-processor designs to essentially become quad boxes.
User: "So, uh, why did you decide to make a word processor that uses 80 megs of RAM and bogs down anything less than a 2 GHz machine?" Programmer: "Why? Why? Muahahha.... BECAUSE I CAN."
Using more resources than necessary to complete a task doesn't demonstrate any sort of talent.
I've always wondered how long it would take processors before they consumed so much power that it was no longer practical to supply them through the traces on the motherboard anymore, and they'd need separate wire leads running from the socket directly back to the power supply. Now I don't pretend to be a mobo designer but it seems that might not be a terrible idea, especially if it let you make the board itself smaller (thinner traces) and/or lighter.
...you need a second processor core just to run the anti-virus and anti-spyware programs.
Good god. More seriously, just seeing people put ideas like that out makes me cringe, not because it's not necessary but because it seems to me that thinking like that will only lead companies like Microsoft to dedicate the second core to nothing but fixing problems that shouldn't be there in the first place. I suppose it's inevitable, though. Programming, especially of the bad, lazy or bloated variety, always seems to expand to fill and tax whatever hardware is available to it.
Apple makes a significant amount of money off of licensing the Dock Connector interface (that big wide locking connector on the bottom of the iPod) to peripheral manufacturers.
Not all peripherals use this, obviously, but some of them do. Especially the bigger and more expensive ones, e.g. the Bose "boombox" attachment.
Anyway, I think it's a sound strategy. It avoids putting them into the low-price commodity market but still lets them make a little money off of it through licensing. I'm not a huge fan of the proprietary Dock Connector, but they don't seem to be being particularly niggardly about letting people license it; I heard in one case they "sold" the rights to the connector in return for selling a particular product through the Apple Store for a set amount of time. So it's not necessarily only available to big companies (in fact most of the cooler accessories seem to being made by small firms, especially the cases and such although they don't use the Connector).
I may have to take that back. I did read somewhere that it was a custom IC, but when trying to find that link again, I came across fairly reputable article that says the accelerometer is a stock part connected to the I2C bus, which also has the fan controllers and temperature sensors.
Philips developed the I2C bus in the early 1980s. I2C is a multi-master control bus using which various ICs in a system can communicate with each other. It uses only two control lines, and has a software-defined protocol. I2C compatible devices typically have on-chip communication interfaces for direct communication with each other over the bus.
There's been a lot of talk about the accelerometer (called the Apple Motion Sensor or ams in the technical documentation) because there are some fairly cool hacks which use it as an input device of sorts. The site noted above seems to be the original source of one of these.
If what you say is true, then SOYO, Tyan, every other mobo manufacturer, and certainly Dell (essentially an assembler) aren't "hardware companies" either.
Apple buys the processor from IBM, but they design and make the motherboards (not in the states, though), the cases, and even have a significant number of custom ICs (e.g. accelerometer chip in PowerBooks), that go into their computers. That seems to me like a hardware company.
Apple certainly is a hardware company, personally I think hardware is one of the things they do best. Just because they don't own a fab plant and make the processor doesn't make them "not a hardware company." Furthermore, your comment seems to imply that Apple used to be a hardware company (by which you seem to mean a processor manufacturer) before they started buying them from IBM. This is also untrue -- before they bought them from IBM, they bought them from Motorola. Apple, to the best of my knowledge, has never made processors. Despite this, they have always been a hardware company.
No, far more secure than any artifact-based security system (ID cards, identity "papers," etc.) would just be to have a reasonably intelligent, alert, well-paid, armed security guard at the entrance to your building. Preferably behind bullet-proof glass, with a remote control to unlock the door.
He would know you from your physical appearance, which lots of hollywood movies and latex masks to the contrary, is rather hard to fake (and only a few people in the population would even have the basic facial structure to fake it). He would know that you're supposed to be in the building, and he'd also be smart enough--hopefully--to recognize if you were under duress or if something else odd was going on that a machine wouldn't pick up on. E.g., if a thug was standing right behind you and trying to make you unlock the door so that they could gain entrance.
There's a reason why inside every U.S. embassy I've ever been into, the "security system" on the door is a Marine guard standing behind a few inches of bullet-proof glass. If you want real security, put the best 'neural net' yet devised at the center of it: a human being. (Okay, some people would argue that Marines aren't quite human anymore, but you get the point.)
Now obviously human beings have strengths and weaknesses. The strengths are obvious, the weakness are that they're relatively expensive compared to machines, and aren't a one-time investment. They also tend to like to work only about 8 hours per day, so you'll need at least four of them to keep things going continuously. So while it's not a solution appropriate to every problem, in terms of access control IMHO it's the best thing going. But there is a reason why some of the most security conscious places in the world are guarded not solely by machines, but also by humans -- you need the person at the controls to respond to new situations that are outside the scope of your computers' programming.
Actually no, I hadn't. But I looked it up, and while it does not invalidate my earlier point, it is a rather interesting phenomenon.
salt fountain: A hypothesized perpetual fountain where a long, narrow heat-conducting pipe inserted vertically through a region of ocean where warm, salty water overlies colder, fresher (and therefore denser) water. Water pumped upwards through the pipe would reach the same temperature as the surroundings at the same level (by conduction of heat through the wall of the pipe), while it remained fresher and therefore lighter. A fountain started thusly (in either direction) will continue to flow so long as there is a vertical gradient of salinity to supply potential energy. The idea was first advanced by Stommel etal. (1956) and is discussed in Turner (1973). from http://stommel.tamu.edu/~baum/paleo/ocean/node35.h tml
You'd have to get the water from the bottom of the ocean flowing through the pipe, first of all, and then the velocity would end up being dependent on the heat transfer through the pipe, the salinity, and the geometry. I'm not sure how strong it'd be, but you might be able to use this to keep the flow going, although not to start it.
Slashdot Mirror
Has anyone thought about using a standard DC power supply (computer supply, etc.) and then using solid-state voltage regulators to produce the required voltages for different devices?
Let's say you have an ATX power supply that puts out +5, +12, and -12. You might even be able to do it with a big brick that just puts out 12VDC, depending on what regulators you buy. Then you use solid-state DC-DC regulators to produce the specific voltages you need for each device. For the more common voltages you could use off-the-shelf fixed regulators (Radioshack even sells a few), and for the less common ones you can use variable regulators and set the voltage with some resistors and capacitors. An example part is LM2734 from National, although I'm sure there are lots of others.
The obvious downside from this is that the max voltage you'd be able to deliver is slightly over 12V (some regulators can produce Vout higher than Vin, but not by much), and the current is limited by the regulators. Also, they need heatsinking, and waste a fair amount of energy I think. But given that DC power supplies are fairly cheap, this might not be a bad thing.
Of course you'd have to get the ATX supply, if that's what you'd choose to use, to produce power in the first place. Sometimes as other people have noticed, they won't work without a load. I use one on my workbench, and in order to get it to produce power at all I have an ancient 5.25" hard drive that I keep plugged in to one of the outputs as a load. I'm sure I could fake the load with resistors or something if I wanted to, but why bother.
Anyway, just a thought for the enterprising hobbyist or experimenter. I'm pretty sure you could use this to power all your small DC devices like peripherals. For large current-draw devices like computers or displays, this might not be an appropriate solution.
There were a few professors that used to end up with quotes on the Daily Jolt website with regularity...I can imagine that a few juicy audio quotes, widely shared, might put a dent in a young professor's career. And this would be unfortunate because some of these people, despite (or perhaps because of) the occasional profanity or humorous exaggeration, are among the best lecturers I've taken class with.
While I'm not advocating this position, I think it's an interesting anecdote. I've certainly found in my personal experience that some of my most memorable classes I have very few paper notes from, because I was either busy participating or just too enraptured with the lecture to write anything down. And while I doubt that the addition of any sort of recording apparatus is ever going to make a boring lecture that much more interesting (heck, I take notes in order to stay awake sometimes), I do wish i had recorded some of the more memorable lectures from my college career.
It uses some proprietary recording format, but the software will convert it to AIFF as you download it to the computer. Although they now have ones with a lot more recording time, I see no reason to replace it. I suppose if I didn't already own one, I'd probably just get the $30 recording add-on for my iPod instead of a standalone device, although the Olympus is significantly smaller and lighter.
If Duke had distributed these to everyone, I suppose we'd be having the same issues. My point is just that the iPod isn't the first device since the analog mini-cassette recorder to give students the capability to record and share a lecture. And as I said in a post further up, in an environment where laptops were ubiquitous, there are programs that do audio recording and note-taking at the same time, and produce a single document that you can share with others.
Couldn't everyone just save a lot of time and just email everyone the stupid lecture? Or for that matter, why bother leaving home at all? Just take it online.
While maybe this might appeal to some
Frankly I think that the new "notebook" document type in MS Office, which combines an audio recording with typed notes and knows where in the recording the notes were taken, is potentially more useful for students than an iPod recorder, because it combines regular notetaking skills with the ability to hear what triggered those notes. And I say this as someone who's not normally a fan of MS products -- it's fairly slick. If you're in an environment where lots of people have laptops and bring them to class, this might have a greater impact in the long run than a bunch of iPods. The impact where I am has been limited, because people don't bring laptops to class very often.
Phone companies work on the assumption -- and it's a good one -- that you want service 24/7/365. Even wireless companies (and I'm no huge fan, trust me) put some thought into taking transmitters on and offline so as not to blackout huge service areas.
I'm not sure I'd want to replace that with a patchwork of small transmitters, each running for their own reasons and on their own schedules, with my phone service basically as an afterthought.
This is why I think WiFi is an interesting addon (as you described) to cell phones, but not a replacement for one outside of carefully controlled environments. (Like using wifi to replace cordless phones in a big warehouse store or corporate/educational campus.)
I'm not criticizing you at all, but there are some other people who seem to like to propose distributed systems as a replacement for everything that's centralized, and I don't think that's a good idea. From a technological perspective I don't have much to complain about the POTS system -- it works when I need it, it's unobtrusive when I don't. I wish it wasn't so expensive...so that's where the market for VoIP comes in.
Can't you just order blank media from the US and have it shipped to Canada? I assume you'd probably be breaking some law at some point in the process ... but it's probably what I'd do if I lived there. How about buying it on eBay? I can't imagine knowing that every CD-R you buy results in some small amount of money going to the copyright whores.
Right. And the bugs -- they're features! Duh.
Nothing except the risk of death, you mean.
If a hacker told me he had a bad crash last week, I'd be sympathetic. If an aircraft mechanic told me that, I'd probably be
Doesn't combining all of the compute-bound processes together onto one processor by definition almost create a bottleneck? Wouldn't it be better to balance your compute- and input-bound processes, so that when the input-bound processes are waiting for input, both processors are working on the compute-bound ones?
I'm not saying you're wrong, I'm just wondering whether this approach wouldn't create more wasted processor cycles than it prevented. However, even if it did that, if the product was a more responsive GUI and interface off of the same hardware, then I might be all for it, wasted cycles or no. (But this may be because I have a personal obsession with responsiveness, and despise any system that has any sort of perceptible lag.)
That focus is a product of the article, not the Linux-heads. (And I mean no offense to Linux-heads, either; I'm writing this using Konqueror.)
According to the article linked to higher in the thread, the upcoming IBM chip will have two cores, each with a separate cache. It also opined that to the OS, it will appear as if the machine has two processors. (Actually the example was discussing two multicored procs appearing as four processors, but you get the idea.)
And given that dual-proc machines have been around for a while now, why is there so much being said about possible uses of dual-core machines? Wouldn't the "uses" for the second core be exactly the same as the "uses" for a second processor in a dual-proc setup? I'm not sure I understand why we're getting all the excitement, except that it might make the benefits of dual-processing more available to the masses in time, and allow current dual-processor designs to essentially become quad boxes.
Programmer: "Why? Why? Muahahha.... BECAUSE I CAN."
Using more resources than necessary to complete a task doesn't demonstrate any sort of talent.
I've always wondered how long it would take processors before they consumed so much power that it was no longer practical to supply them through the traces on the motherboard anymore, and they'd need separate wire leads running from the socket directly back to the power supply. Now I don't pretend to be a mobo designer but it seems that might not be a terrible idea, especially if it let you make the board itself smaller (thinner traces) and/or lighter.
Good god. More seriously, just seeing people put ideas like that out makes me cringe, not because it's not necessary but because it seems to me that thinking like that will only lead companies like Microsoft to dedicate the second core to nothing but fixing problems that shouldn't be there in the first place. I suppose it's inevitable, though. Programming, especially of the bad, lazy or bloated variety, always seems to expand to fill and tax whatever hardware is available to it.
The pizza would probably become some sort of singularity, and instantly annihilate the rest of universe. But hey, at least there'd be pizza.
Not all peripherals use this, obviously, but some of them do. Especially the bigger and more expensive ones, e.g. the Bose "boombox" attachment.
Anyway, I think it's a sound strategy. It avoids putting them into the low-price commodity market but still lets them make a little money off of it through licensing. I'm not a huge fan of the proprietary Dock Connector, but they don't seem to be being particularly niggardly about letting people license it; I heard in one case they "sold" the rights to the connector in return for selling a particular product through the Apple Store for a set amount of time. So it's not necessarily only available to big companies (in fact most of the cooler accessories seem to being made by small firms, especially the cases and such although they don't use the Connector).
Hey I don't care how it works, if it can make the pizza show up five minutes before I order it, I'm sold.
from http://www.kernelthread.com/software/ams/:
There's been a lot of talk about the accelerometer (called the Apple Motion Sensor or ams in the technical documentation) because there are some fairly cool hacks which use it as an input device of sorts. The site noted above seems to be the original source of one of these.
Apple buys the processor from IBM, but they design and make the motherboards (not in the states, though), the cases, and even have a significant number of custom ICs (e.g. accelerometer chip in PowerBooks), that go into their computers. That seems to me like a hardware company.
Apple certainly is a hardware company, personally I think hardware is one of the things they do best. Just because they don't own a fab plant and make the processor doesn't make them "not a hardware company." Furthermore, your comment seems to imply that Apple used to be a hardware company (by which you seem to mean a processor manufacturer) before they started buying them from IBM. This is also untrue -- before they bought them from IBM, they bought them from Motorola. Apple, to the best of my knowledge, has never made processors. Despite this, they have always been a hardware company.
He would know you from your physical appearance, which lots of hollywood movies and latex masks to the contrary, is rather hard to fake (and only a few people in the population would even have the basic facial structure to fake it). He would know that you're supposed to be in the building, and he'd also be smart enough--hopefully--to recognize if you were under duress or if something else odd was going on that a machine wouldn't pick up on. E.g., if a thug was standing right behind you and trying to make you unlock the door so that they could gain entrance.
There's a reason why inside every U.S. embassy I've ever been into, the "security system" on the door is a Marine guard standing behind a few inches of bullet-proof glass. If you want real security, put the best 'neural net' yet devised at the center of it: a human being. (Okay, some people would argue that Marines aren't quite human anymore, but you get the point.)
Now obviously human beings have strengths and weaknesses. The strengths are obvious, the weakness are that they're relatively expensive compared to machines, and aren't a one-time investment. They also tend to like to work only about 8 hours per day, so you'll need at least four of them to keep things going continuously. So while it's not a solution appropriate to every problem, in terms of access control IMHO it's the best thing going. But there is a reason why some of the most security conscious places in the world are guarded not solely by machines, but also by humans -- you need the person at the controls to respond to new situations that are outside the scope of your computers' programming.
You'd have to get the water from the bottom of the ocean flowing through the pipe, first of all, and then the velocity would end up being dependent on the heat transfer through the pipe, the salinity, and the geometry. I'm not sure how strong it'd be, but you might be able to use this to keep the flow going, although not to start it.
Good to know -- thanks.
Maybe you can use pico instead of a GUI editor? I'm pretty sure that sudo pico works.