Even better, if we have cheap energy, you can use that energy to artificially subject the algae to the sorts of heat and pressure that naturally create oil. No 10 million year wait!
Although I don't know one off the top of my head, I'd assume that there are ways to drive a reaction from a carbon feedstock back to hydrocarbon fuel if energy is assumed to be limitless. As such, there could be a human carbon cycle instituted:
CO2 from atmosphere -> energy intensive process to convert CO2's carbon back into hydrocarbon fuel -> burning fuel to create CO2. This eliminates the problem of adding CO2 to the atmosphere by closing the cycle, the same thing biodiesel advocates want, without the limit of needing to grow plants.
Alternately, any fuel that allows us to retrieve the product from the atmosphere and reform it would work.
Finally: hydrogen. A pure hydrogen cycle would be
H20 + energy -> 2H2 + O2 (electrolysis) 2H2 + O2 -> H20 + energy (combustion)
They are constant in value; however, you are not given that value. This makes them arbitrary constants.
This is important when you start doing calculus; the differentiation of a variable is very different from the differentiation of a constant.
Re:Somone get these ppl some free software!
on
Given Up to Spyware?
·
· Score: 1
HTTPS is run through a proxy as well, so what good would that do me? There are to my knowledge no non-proxied connections allowed outbound from our network.
Most of my courses, the exams were written such that a calculator was completely useless. You were allowed to use one, but when you're integrating:
1-(sin(A*cos (B*sin AB))^C) where A,B, and C are arbitrary constants (i.e. you have to leave them as A,B, and C - you aren't given values), a graphing calculator doesn't help very much.
The easiest way to make a calculator useless is to make all numeric constants into alphabetic constants with undefined values.
Re:Somone get these ppl some free software!
on
Given Up to Spyware?
·
· Score: 1
It was quite some time ago that I last looked at it; when I did, they were just starting to talk about OSCAR but were still using TOC (OSCAR crashed the client).
Re:Somone get these ppl some free software!
on
Given Up to Spyware?
·
· Score: 1
SSH is blocked at the firewall, due to the possiblity of reverse-tunneling.
It was the very first thing I tried.
Re:Somone get these ppl some free software!
on
Given Up to Spyware?
·
· Score: 1
Two things about gaim:
Up until I checked again recently, gaim had been using TOC, not OSCAR. TOC lacks certain nice features, like being able to check people's away messages, server-side buddy lists, etc, etc. Until gaim added OSCAR support, I wasn't willing to use it. Not sure when this was added, but when I played with it a few years ago, it didn't.
Second, work. The official AIM client supports HTTPS tunneling. gaim does not. My employer doesn't care if I use AIM, but they do require I tunnel it through the secure proxy. Why do they require I go through the secure proxy instead of the regular? No idea.
But for the most part, the spyware-infested stuff is *exactly* the kind of software where decent (or at least equivalent) open/free alternatives exist. Too bad people don't always know about them.
A mono input doesn't mean a mono output. That's why God made pan/balance controls. Mono inputs can be mixed in stereo very, very easily, and often are.
More important than the input is the output, and I've never seen a decent PA that has a mono output; PAs are designed to run in stereo, and generally are operated in stereo, even if that capability is only minimally used.
And no, it doesn't have to be all that small; sure, stadium venues don't generally have this issue, but anything up to a 500-600 capacity bar, the amps are going to be some audible portion of the sound - again, the preceding sound DOES NOT need to be louder, it just needs to be audible, and prior. The same effect is used in theatre to localize an amplified singer's voice to the singer. Although the amplification is louder than the singer's voice, the audience 'hears' it as coming from the singer because the singer's voice arrives first and the brain's acoustic processing says "That's the source" even when the amplification arrives a little later. The effect does break down if the amplification arrives too quickly, too late, or is a lot louder than the source, but it works surprisingly well.
You can prove the effect to yourself as follows, with any decent audio program (Reaktor or MAX/MSP would work, others might be - you need to be able to realtime manipulate effects, and path as needed).
Setup a source.
Split the source into two paths. Path A will be fed directly to the left speaker. Path B will be run through a delay, set to something in the 5-10 ms range, and then fed to the right channel.
Set the two speakers in seperated locations, and start playing back the source. Listen and decide where it sounds like the sound is coming from.
The audience get a mono sound, since all the amps, drums, vocals are all mixed through the PA into one channel. I mean if there was stereo at a concert then some of the audience would only here the instruments on the left and vice versa.
Not necessarily correct, for two reasons: the Haas effect and sound people.
Sound people often do pan instruments to one side or the other to get a little bit of seperation; generally they don't go 'guitar A in left channel, B in right', but they may pan 30 degrees each.
The Haas effect is a psychoacoustic effect; the smaller the venue, the more likely it is to come into play. Basically, what it says is that if one sound arrives slightly (within a couple ms) before a louder copy of itself, the brain will localize the sum of both sounds to be coming from the earlier arriving source. In most small venues, the instrument amplifiers will arrive slightly earlier than the PA sound, which is why you actually do get a 'stereo' effect even if the PA is mixed mono.
All that said, surround for music is still a dumb idea.
FPGAs have been around commercially for 20 years or so. If a market hasn't developed in 20 years, its *usually* a safe bet it isn't there.
Time to program an FPGA is fairly short, if you already have a synthesized layout. Time to synthesize a layout is way, way longer than software compile times - an extremely simple design (think ultra-simple 4 bit processor core) might take 15 or 30 minutes to synthesize, something like an open-source SPARC core is probably close to a day worth of time. When we do synth/layout runs on our ASICs (which is similar to doing it for FPGAs), run times of days to get a usable synthesis are not unusual.
You can reprogram the chip as you go, but generally I figure people will want to use at least some of the hardware they build, and if every use requires a new chip, it adds up quickly.
As to getting a board for yourself, if you want to play around with a system, Xilinx sells a Spartan-3 eval board for $100. That's a 200k-gate board, which is enough to do some interesting things, and includes a demo version of their software as well. If you'd rather learn on Altera devices, $150 buys you their MAX-2 dev board and eval edition of their software. If you're in university you might be able to swing their University Board, which seems to include slightly nicer stuff for the same price.
It's not hard to play around with this stuff, it's just expensive to distribute, which is why I continue to believe open hardware *in the same meaning as open software* is unlikely, at least until some way of instancing hardware as cheaply as software appears.
FPGA chips are expensive because they are internally quite complex. The companies that make FPGAs really don't make all that much of their money selling the chips - IIRC, they sell the chips at close to cost/loss. They make their money on the ridiculously expensive toolsets they sell to make use of the chips. Why aren't there general purpose FPGA boards on the consumer market? Not because there are no designs, because there is no desire. The difference between a Linux machine and a Windows machine is the software. The hardware is commodity; if Linux had required big iron to work, the whole thing would be a lot further back than it is. Contrarily, an FPGA appeals to almost no one. There's minimal mass-market appeal to the thing; the whole fun of one is that it can do many things somewhat well, but nothing incredibly well. ASICs, which are much, much cheaper in volume, can do anything an FPGA can do, better... except reconfigure themselves. As a result, mass market apps get ASICs or firmware, which are cheap, and FPGAs remain expensive, because there really isn't a general market for them at this time. Not a lack of designs, a lack of desires.
You want an FPGA board? Go buy one. They exist. I had one for a while, a hand-me-down Xilinx board and some software from work to go with it. You can even buy them for, yes, prices in between a NIC and a video card. Of course, it won't be an incredibly powerful FPGA at those prices.
Why is sharing hardware designs more expensive? Because every new instance of hardware requires a new FPGA board. Software requires minimal plant to make use of compared to hardware. If you want to have one board and only ever have one design at a time, that's cool, but most people would like
I'm not tense. I always write like this. Get used to it.
Christ. With just Munchkin Fu and Bites, we already manage to have 3 hour games (BTW... a fun trick to remember, if you're Yakuza and have a wandering monster card - wait until something nasty hits someone, discard three to take the top monster off the pile (Mr. Nasty), and then wandering monster it back at them. We love to lawyer.)
We haven't even touched epic yet. Frankly, we're afraid.
WLUW was owned by Loyola University, and then I think Chicago Public Radio, but I believe they gave it freedom recently.
WNUR is Northwestern University's.
WDET is a publicly supported freeform station in Detroit.
WCBN is one of the oldest freeform FM radio stations in the country. It also happens to belong to the University of Michigan.
WXOU belongs to Oakland University.
Do you notice a trend towards non-commercial educational stations here? There's a reason my radio never goes above 91.5 (local NPR affiliate) - all the good stuff really is on the left side of the FM dial.
There's also WFMU in New York, which is independent and awesome, but they're something of an exception - they operate in a vicinity where an independent station can actually afford to operate.
When people say 'discrete math', my mind immediately jumps to discrete time sample data systems, maybe because I basically live with them at work. I realize there are other meanings, but I'm thinking of DSP and related fields (sampling theory, communications theory, and control systems all rely heavily on discrete maths in one form or another). It's entirely possible that my mental definition of 'discrete math' is entirely specific to me; it wasn't intended to confuse.
But yeah, I find it hard to believe anyone whose job title will eventually include the words 'science', 'researcher', or 'engineer' won't be benefited by at least basic calculus.
I'd subtract the maybe from Calculus; it's a lot more important to understand calculus than it is to understand networks or databases, at least if you ever want to work for a company that designs *things* as opposed to programs. Engineering companies hire a lot of CS people as well, and if you don't know calc you can't do any of the really fun stuff. Networks and databases are things you can pick up if you need to; if you can get them, do so, but don't sacrifice a decent grounding in calculus for them.
Plus, calc makes a lot of the other things seem easier, particularly discrete math (hey, if you can comprehend infinitesimals, discrete math is *easy*) and stats (understanding why statistics are the way they are is as important as understanding how to use them).
Chicago (WLUW, WNUR, a couple others that aren't coming to mind) and SE Michigan/Detroit (WXOU, WCBN, WDET) both have excellent selections of stations that play a good cross-section of everything. I'm sure NYC does too, though I've never lived there.
The incremental cost of copying software is practically zero. The incremental cost of copying hardware, EVEN WITH PROGRAMMABLE LOGIC, is far from zero. FPGA's, especially those usable for truly interesting tasks, aren't cheap. Considering I work with them in my day job, I'm pretty sure I'm right about this fact.
Oh, and before you give me the "but you have to have a computer to use/work on open source software" - you pretty much have to have a computer to work on open source hardware as well, since all of the fancy ASIC/FPGA synthesis tools run on computers as well. Same goes for the bandwidth transmission costs of software source - how exactly are you planning to share those hardware designs?
The reasoning is true given current circumstances, and is exactly why open source hardware has a much harder time taking off than open source software. It's truth has roughly nothing to do with programmable logic, except that programmable logic is an order of magnitude or two cheaper than, for example, ASICs would be. However, it is still orders of magnitude more expensive than software.
It doesn't hurt to actually think about the things you're writing. Your first post proved you don't read very closely, and this second one proved you don't think very hard either.
By the way:
If this IS THE reason, it's not entirely true (unless you've NO CLUE) what a piece of "programmable logic" is.
parses very poorly. It would have read grammatically as:
If this IS THE reason, it's not entirely true (unless you've NO CLUE what a piece of "programmable logic" is).
Remember that, if you use punctuation to set off a seperate structure, you still need to finish the initial phrase properly.
Our group wound up buying the IDE from Metrowerks and using an in-house RTOS, so no extraction of revenge possible.
But yeah; INTEGRITY seems nice, from my (control systems engineering) perspective as someone who doesn't like dealing with writing software. MULTI, on the other hand, is entirely counter-intuitive to my mind.
GHS will allow paying customers access to the source; for things like DOD projects, where there are certain certifications required to use software, opening the code and *documenting every single line* is a prerequisite.
The advantage to using the sort of OS GHS really sells for embedded computing is that their OSes are provable in many senses - deterministic, etc. Windows and Linux are not. GHS isn't trying to compete with desktop Linux; their product isn't capable of that. They are, however, competing with embedded Linux, and in that arena their software has some very real advantages.
Depending on who was using it, I think I'd rather have some cruise missiles running Linux than (e.g.) INTEGRITY. I've worked with INTEGRITY, it seems like a hell of a nice system (although for our application it was overkill and too expensive).
MULTI, on the other hand, makes me want to kill myself. GHS guy, sorry, but that product literally makes me want to die every time I have to use it.
Slashdot Mirror
Even better, if we have cheap energy, you can use that energy to artificially subject the algae to the sorts of heat and pressure that naturally create oil. No 10 million year wait!
I'm guessing it's some kind of phsycological need to humiliate yourself in public.
Psychological.
Although I don't know one off the top of my head, I'd assume that there are ways to drive a reaction from a carbon feedstock back to hydrocarbon fuel if energy is assumed to be limitless. As such, there could be a human carbon cycle instituted:
CO2 from atmosphere -> energy intensive process to convert CO2's carbon back into hydrocarbon fuel -> burning fuel to create CO2. This eliminates the problem of adding CO2 to the atmosphere by closing the cycle, the same thing biodiesel advocates want, without the limit of needing to grow plants.
Alternately, any fuel that allows us to retrieve the product from the atmosphere and reform it would work.
Finally: hydrogen. A pure hydrogen cycle would be
H20 + energy -> 2H2 + O2 (electrolysis)
2H2 + O2 -> H20 + energy (combustion)
No.
They are constant in value; however, you are not given that value. This makes them arbitrary constants.
This is important when you start doing calculus; the differentiation of a variable is very different from the differentiation of a constant.
HTTPS is run through a proxy as well, so what good would that do me? There are to my knowledge no non-proxied connections allowed outbound from our network.
Most of my courses, the exams were written such that a calculator was completely useless. You were allowed to use one, but when you're integrating:
1-(sin(A*cos (B*sin AB))^C) where A,B, and C are arbitrary constants (i.e. you have to leave them as A,B, and C - you aren't given values), a graphing calculator doesn't help very much.
The easiest way to make a calculator useless is to make all numeric constants into alphabetic constants with undefined values.
It was quite some time ago that I last looked at it; when I did, they were just starting to talk about OSCAR but were still using TOC (OSCAR crashed the client).
SSH is blocked at the firewall, due to the possiblity of reverse-tunneling.
It was the very first thing I tried.
Two things about gaim:
Up until I checked again recently, gaim had been using TOC, not OSCAR. TOC lacks certain nice features, like being able to check people's away messages, server-side buddy lists, etc, etc. Until gaim added OSCAR support, I wasn't willing to use it. Not sure when this was added, but when I played with it a few years ago, it didn't.
Second, work. The official AIM client supports HTTPS tunneling. gaim does not. My employer doesn't care if I use AIM, but they do require I tunnel it through the secure proxy. Why do they require I go through the secure proxy instead of the regular? No idea.
But for the most part, the spyware-infested stuff is *exactly* the kind of software where decent (or at least equivalent) open/free alternatives exist. Too bad people don't always know about them.
A mono input doesn't mean a mono output. That's why God made pan/balance controls. Mono inputs can be mixed in stereo very, very easily, and often are.
More important than the input is the output, and I've never seen a decent PA that has a mono output; PAs are designed to run in stereo, and generally are operated in stereo, even if that capability is only minimally used.
And no, it doesn't have to be all that small; sure, stadium venues don't generally have this issue, but anything up to a 500-600 capacity bar, the amps are going to be some audible portion of the sound - again, the preceding sound DOES NOT need to be louder, it just needs to be audible, and prior. The same effect is used in theatre to localize an amplified singer's voice to the singer. Although the amplification is louder than the singer's voice, the audience 'hears' it as coming from the singer because the singer's voice arrives first and the brain's acoustic processing says "That's the source" even when the amplification arrives a little later. The effect does break down if the amplification arrives too quickly, too late, or is a lot louder than the source, but it works surprisingly well.
You can prove the effect to yourself as follows, with any decent audio program (Reaktor or MAX/MSP would work, others might be - you need to be able to realtime manipulate effects, and path as needed).
Setup a source.
Split the source into two paths. Path A will be fed directly to the left speaker. Path B will be run through a delay, set to something in the 5-10 ms range, and then fed to the right channel.
Set the two speakers in seperated locations, and start playing back the source. Listen and decide where it sounds like the sound is coming from.
The audience get a mono sound, since all the amps, drums, vocals are all mixed through the PA into one channel. I mean if there was stereo at a concert then some of the audience would only here the instruments on the left and vice versa.
Not necessarily correct, for two reasons: the Haas effect and sound people.
Sound people often do pan instruments to one side or the other to get a little bit of seperation; generally they don't go 'guitar A in left channel, B in right', but they may pan 30 degrees each.
The Haas effect is a psychoacoustic effect; the smaller the venue, the more likely it is to come into play. Basically, what it says is that if one sound arrives slightly (within a couple ms) before a louder copy of itself, the brain will localize the sum of both sounds to be coming from the earlier arriving source. In most small venues, the instrument amplifiers will arrive slightly earlier than the PA sound, which is why you actually do get a 'stereo' effect even if the PA is mixed mono.
All that said, surround for music is still a dumb idea.
FPGAs have been around commercially for 20 years or so. If a market hasn't developed in 20 years, its *usually* a safe bet it isn't there.
Time to program an FPGA is fairly short, if you already have a synthesized layout. Time to synthesize a layout is way, way longer than software compile times - an extremely simple design (think ultra-simple 4 bit processor core) might take 15 or 30 minutes to synthesize, something like an open-source SPARC core is probably close to a day worth of time. When we do synth/layout runs on our ASICs (which is similar to doing it for FPGAs), run times of days to get a usable synthesis are not unusual.
You can reprogram the chip as you go, but generally I figure people will want to use at least some of the hardware they build, and if every use requires a new chip, it adds up quickly.
As to getting a board for yourself, if you want to play around with a system, Xilinx sells a Spartan-3 eval board for $100. That's a 200k-gate board, which is enough to do some interesting things, and includes a demo version of their software as well. If you'd rather learn on Altera devices, $150 buys you their MAX-2 dev board and eval edition of their software. If you're in university you might be able to swing their University Board, which seems to include slightly nicer stuff for the same price.
It's not hard to play around with this stuff, it's just expensive to distribute, which is why I continue to believe open hardware *in the same meaning as open software* is unlikely, at least until some way of instancing hardware as cheaply as software appears.
Give the Humboldt kids a few weeks, they'll find a way to make the cameras irrelevant.
FPGA chips are expensive because they are internally quite complex. The companies that make FPGAs really don't make all that much of their money selling the chips - IIRC, they sell the chips at close to cost/loss. They make their money on the ridiculously expensive toolsets they sell to make use of the chips. Why aren't there general purpose FPGA boards on the consumer market? Not because there are no designs, because there is no desire. The difference between a Linux machine and a Windows machine is the software. The hardware is commodity; if Linux had required big iron to work, the whole thing would be a lot further back than it is. Contrarily, an FPGA appeals to almost no one. There's minimal mass-market appeal to the thing; the whole fun of one is that it can do many things somewhat well, but nothing incredibly well. ASICs, which are much, much cheaper in volume, can do anything an FPGA can do, better... except reconfigure themselves. As a result, mass market apps get ASICs or firmware, which are cheap, and FPGAs remain expensive, because there really isn't a general market for them at this time. Not a lack of designs, a lack of desires.
You want an FPGA board? Go buy one. They exist. I had one for a while, a hand-me-down Xilinx board and some software from work to go with it. You can even buy them for, yes, prices in between a NIC and a video card. Of course, it won't be an incredibly powerful FPGA at those prices.
Why is sharing hardware designs more expensive? Because every new instance of hardware requires a new FPGA board. Software requires minimal plant to make use of compared to hardware. If you want to have one board and only ever have one design at a time, that's cool, but most people would like
I'm not tense. I always write like this. Get used to it.
Christ. With just Munchkin Fu and Bites, we already manage to have 3 hour games (BTW... a fun trick to remember, if you're Yakuza and have a wandering monster card - wait until something nasty hits someone, discard three to take the top monster off the pile (Mr. Nasty), and then wandering monster it back at them. We love to lawyer.)
We haven't even touched epic yet. Frankly, we're afraid.
Actually... no. None of them are.
WLUW was owned by Loyola University, and then I think Chicago Public Radio, but I believe they gave it freedom recently.
WNUR is Northwestern University's.
WDET is a publicly supported freeform station in Detroit.
WCBN is one of the oldest freeform FM radio stations in the country. It also happens to belong to the University of Michigan.
WXOU belongs to Oakland University.
Do you notice a trend towards non-commercial educational stations here? There's a reason my radio never goes above 91.5 (local NPR affiliate) - all the good stuff really is on the left side of the FM dial.
There's also WFMU in New York, which is independent and awesome, but they're something of an exception - they operate in a vicinity where an independent station can actually afford to operate.
When people say 'discrete math', my mind immediately jumps to discrete time sample data systems, maybe because I basically live with them at work. I realize there are other meanings, but I'm thinking of DSP and related fields (sampling theory, communications theory, and control systems all rely heavily on discrete maths in one form or another). It's entirely possible that my mental definition of 'discrete math' is entirely specific to me; it wasn't intended to confuse.
But yeah, I find it hard to believe anyone whose job title will eventually include the words 'science', 'researcher', or 'engineer' won't be benefited by at least basic calculus.
I'd subtract the maybe from Calculus; it's a lot more important to understand calculus than it is to understand networks or databases, at least if you ever want to work for a company that designs *things* as opposed to programs. Engineering companies hire a lot of CS people as well, and if you don't know calc you can't do any of the really fun stuff. Networks and databases are things you can pick up if you need to; if you can get them, do so, but don't sacrifice a decent grounding in calculus for them.
Plus, calc makes a lot of the other things seem easier, particularly discrete math (hey, if you can comprehend infinitesimals, discrete math is *easy*) and stats (understanding why statistics are the way they are is as important as understanding how to use them).
Chicago (WLUW, WNUR, a couple others that aren't coming to mind) and SE Michigan/Detroit (WXOU, WCBN, WDET) both have excellent selections of stations that play a good cross-section of everything. I'm sure NYC does too, though I've never lived there.
The incremental cost of copying software is practically zero. The incremental cost of copying hardware, EVEN WITH PROGRAMMABLE LOGIC, is far from zero. FPGA's, especially those usable for truly interesting tasks, aren't cheap. Considering I work with them in my day job, I'm pretty sure I'm right about this fact.
Oh, and before you give me the "but you have to have a computer to use/work on open source software" - you pretty much have to have a computer to work on open source hardware as well, since all of the fancy ASIC/FPGA synthesis tools run on computers as well. Same goes for the bandwidth transmission costs of software source - how exactly are you planning to share those hardware designs?
The reasoning is true given current circumstances, and is exactly why open source hardware has a much harder time taking off than open source software. It's truth has roughly nothing to do with programmable logic, except that programmable logic is an order of magnitude or two cheaper than, for example, ASICs would be. However, it is still orders of magnitude more expensive than software.
It doesn't hurt to actually think about the things you're writing. Your first post proved you don't read very closely, and this second one proved you don't think very hard either.
By the way:
If this IS THE reason, it's not entirely true (unless you've NO CLUE) what a piece of "programmable logic" is.
parses very poorly. It would have read grammatically as:
If this IS THE reason, it's not entirely true (unless you've NO CLUE what a piece of "programmable logic" is).
Remember that, if you use punctuation to set off a seperate structure, you still need to finish the initial phrase properly.
Wait, they've released Munchkin 2 and 3, and Star Munchkin 2, and Munchkin Fu 2? I already had Munchkin Blender, but damn, this is tempting. Tempting.
(Actually, considering SJ's history of releasing interminable numbers of sourcebooks, I'd pick on GURPS first.)
Our group wound up buying the IDE from Metrowerks and using an in-house RTOS, so no extraction of revenge possible.
But yeah; INTEGRITY seems nice, from my (control systems engineering) perspective as someone who doesn't like dealing with writing software. MULTI, on the other hand, is entirely counter-intuitive to my mind.
GHS will allow paying customers access to the source; for things like DOD projects, where there are certain certifications required to use software, opening the code and *documenting every single line* is a prerequisite.
The advantage to using the sort of OS GHS really sells for embedded computing is that their OSes are provable in many senses - deterministic, etc. Windows and Linux are not. GHS isn't trying to compete with desktop Linux; their product isn't capable of that. They are, however, competing with embedded Linux, and in that arena their software has some very real advantages.
Depending on who was using it, I think I'd rather have some cruise missiles running Linux than (e.g.) INTEGRITY. I've worked with INTEGRITY, it seems like a hell of a nice system (although for our application it was overkill and too expensive).
MULTI, on the other hand, makes me want to kill myself. GHS guy, sorry, but that product literally makes me want to die every time I have to use it.
Don't forget Star Munchkin, Munchkin Fu, and Munchkin Bites.
(Actually, Munchkin is a lot of fun.)