Something that could make computing fun again like it used to be 20-30 years ago.
My initial thought was "Linux", but I guess we've all been there for years already, and the novelty is starting to wear off. But talking about custom chips, I have recently taken up FPGA design, and it's been great fun. Though in that case you may have to forget about the idea of a "computer" for a while, and think more generally about digital circuits. Forget about programming a given piece of hardware, and instead design a piece of hardware to do it much faster. Or if you like, you can include a Free CPU design and make your computer from scratch, without any soldering.
I must admit I'm not very far in this field yet, so here are some fun links from other people:
I've brought motherboards and power supplies back from the dead with repairs like that.
I wonder if it's just a coincidence, but my repairs have also been mainly motherboards and power supplies. As opposed to displays, for example my 2004 LCD monitor is doing great, and I only bought it to replace a smaller CRT that was alive and well.
My particular gripe is with power bricks; replacing caps is the easy part, after you've pried open the glue seams, and then you have to glue it back together. I wouldn't bother with any random power bricks, but the ones around 100 W that you use with PicoPSU and the like are usually worth it.
It's funny that we are supposed to reduce hazardous waste, by making electronics increasingly single-use and disposable. As other people have pointed out, a CRT full of lead is not leaching into the environment while it's being used. Besides, leaded solder is easier for repairs.
I was thinking the same thing, so I looked into it a bit (no pun intended). Apparently "true" 64-bit processing uses a more modern instruction set on the CPU, so I suppose there are additional performance and security benefits to using it.
Alpha had 64-bit processing in 1992, and that's certainly a more modern instruction set than x86...
Wait, are we talking about x86-64? I thought I had already been running a native Firefox on my x86-64 boxes for years.
I guess it should be managed like a cache, with your brain being the cache, and books being secondary storage
My thoughts exactly. While RAM and disk have grown by huge amounts, our CPU and cache are basically the same as in the caveman days. To get any thinking done, you still need to get that data into the cache, and frequently used data will tend to stay there.
This is also why huge organizations are inefficient, there is too much communication overhead, and it is mainly in small teams where actual work gets done.
I use Linux on a Powerbook with Radeon graphics. For some odd reason, AMD does not provide binary drivers for this platform, but the open driver works great.
On my other machines with x86-64, I use the binary Radeon drivers as it is the only way to get full OpenCL. Even there, it sometimes happens that I need to use an older kernel or disable some kernel features, as the binary driver does not play well with the pace of Linux development.
My GPUs paid for a nice FPGA devkit, after paying for themselves. If Bitcoin continues this way, the devkit may end up paying some more, as it is much more power efficient.
If you are referring something that people have already been running for a few months, you probably mean this. The modular miner thread seems to focus on raw price/performance, ignoring factors like the freedom to program the said FPGA without expensive licenses.
I just hack FPGAs, not trying to emulate old machines though. Who needs software when you can design hardware to do what you want?
Strangely enough, I've never really coded anything at an assembler level, not even much C, so it's not like I'm getting systematically deeper. Then again I have plenty of experience in hands-on electronics. It's great being able to define your data structures at a bitwise level, not worrying about some little endian crap. Also a nice way to get some ideas for parallel programming in software, because things are actually happening in parallel and not just timesliced.
Also, don't forget that FPGA is an acronym, I pronounce it as "fapgay".
I can attest that Verilog is easy and intuitive, since it got me from a total newbie to a project that other people find useful, in a couple of weeks. Of course, some electronics and computing background helped a lot.
My choice of language was partly influenced by the apparent verbosity of VHDL. It's the same reason why I preferred to learn Python instead of Java, for example. However, I also have the impression that VHDL is in some sense lower-level and more strictly defined, so you can have finer control on some things. For example, to do arithmetic in VHDL you need to explicitly load the requisite library. I'm not sure how much this influences the resulting circuit, though.
How exactly can a transaction system that broadcasts all of your transactions to the network "make the world a little better place".
Because it's not Paypal;)
Seriously though, I'm not saying that Bitcoin is perfect. That's kind of the point of geek things, we're trying to break some new ground here. But the specific point of broadcast transaction seems quite essential if you want to keep things distributed, so that other peers will verify your transaction.
The "exchanges" represent a mis-design of Bitcoin. There should have been a way to do an exchange in a distributed way, without the exchange holding customer assets. The NYSE and NASDAQ don't hold customer balances. Brokers do, but you can have your cash swept from a brokerage into a bank daily, or more often if the numbers get big. The Bitcoin exchanges are slow at delivering money - Mt. Gox has a daily transfer limit, and even when they were up, many users reported delays.
There are other exchanges that work more in the way you describe, for example I have personally used bitcoinmarket.eu. Too bad about the MtGox episode, but it has nothing to do with the rest of the Bitcoin-using world. Of course, USD becomes a pyramid scheme the minute somebody hacks into a forex site.
Also, let's not forget that Bitcoin is a geek thing, something that could make the world a little better place, and not just something to make you rich. For example, I learned to program FPGAs because of Bitcoin. I now have a new hobby with tons of interesting prospects, no matter how well Bitcoin fares in the future.
Also, not to nitpick (well, yes to nitpick), but I think that part that says "suddenly every piece of software works..." needs a bit of filling out. Especially at the "suddenly" part.
This made me realize that even paper books are essentially digital. Even after a hundred years and countless readings, an "a" is still an "a", and not some approximation. Some of the letters may wear out entirely, but the redundant coding of natural languages will maintain the integrity of the content, much like the coding in CDs. Nothing like this in vinyl records, it wears out a little every time you play it, and there is no way to get that information back exactly.
ed is also two characters, and they are adjacent on the keyboard. If you are a touch-typist, it takes 50% of the hands and 50% of the fingers necessary to start that inefficient vi editor.
In the typing standard I've learned, E and D are on the same finger. It is much faster to use two fingers for two keys, so clearly vi wins. Not that either would matter, since I already have Emacs running.
And paging through? Spacebar dipshit. Firefox, Acrobat, will page when you press it. Also there are these little keys called "page up" and "page down". Wonder what THEY do?
The problem with the space bar as PgDn is that the equivalent of PgDn is not consistent across applications. Sometimes it is Shift+Space, sometimes B or something else.
Now, actual, dedicated PgUp/PgDn keys, these are something that I miss on many modern laptop keyboards. I guess these started to fade away when scroll wheels on mice became ubiquitous, but I still see many people who use neither. Instead, they click and drag on the scroll bar. For the hardcore users who still use old-fashioned PgUp/Dn, they must be accessed via Fn keys. Even arrow keys are shrinking, since nobody uses them any more. Obviously, everything is easier with the mouse, and soon we won't need the keyboard any more.
Finite wireless spectrum?!? What are you talking about? Let's talk Mhz:
There is Ghz spectrum between say, 2.4 and 3.4 Ghz, which seems limited. So you might break it out into 1 mhz bands, giving you 1,000 usable frequencies. Or break it more finely,into.1 mhz bands giving you 10,000 usable, or.01 giving you 100,000 frequencies, or...
A 0.01 MHz band does not give you much capacity, perhaps something of the order 0.1 Mbps. While bandwidth is not the same thing as data rate, they are proportional.
Spread spectrum technology, first developed by military for secretive radio communications, send information in short bursts in pseudorandom frequencies. This frquency hopping allows for far more efficient use of existing radio frquencies with minimal disruption. Numerous studies show this type of technology could extend the available bandwidth a billionfold or more.
By definition, spread spectrum uses a lot of bandwidth;) The problem with data rate is that when everyone uses spread spectrum, noise floor goes up, and thus the signal/noise ratio gets worse. This, in turn, means a smaller data rate per bandwidth, as explained by Shannon.
Slashdot Mirror
Something that could make computing fun again like it used to be 20-30 years ago.
My initial thought was "Linux", but I guess we've all been there for years already, and the novelty is starting to wear off. But talking about custom chips, I have recently taken up FPGA design, and it's been great fun. Though in that case you may have to forget about the idea of a "computer" for a while, and think more generally about digital circuits. Forget about programming a given piece of hardware, and instead design a piece of hardware to do it much faster. Or if you like, you can include a Free CPU design and make your computer from scratch, without any soldering.
I must admit I'm not very far in this field yet, so here are some fun links from other people:
Exactly. I guess it's news when a big business discovers the world outside x86, where some of us have been living for years.
I've brought motherboards and power supplies back from the dead with repairs like that.
I wonder if it's just a coincidence, but my repairs have also been mainly motherboards and power supplies. As opposed to displays, for example my 2004 LCD monitor is doing great, and I only bought it to replace a smaller CRT that was alive and well.
My particular gripe is with power bricks; replacing caps is the easy part, after you've pried open the glue seams, and then you have to glue it back together. I wouldn't bother with any random power bricks, but the ones around 100 W that you use with PicoPSU and the like are usually worth it.
It's funny that we are supposed to reduce hazardous waste, by making electronics increasingly single-use and disposable. As other people have pointed out, a CRT full of lead is not leaching into the environment while it's being used. Besides, leaded solder is easier for repairs.
You can simply put several signal posts in a row, and read the red/amber/green/blue streak that goes by the cabin.
While you're going at ludicrous speeds, don't forget Doppler.
I was thinking the same thing, so I looked into it a bit (no pun intended). Apparently "true" 64-bit processing uses a more modern instruction set on the CPU, so I suppose there are additional performance and security benefits to using it.
Alpha had 64-bit processing in 1992, and that's certainly a more modern instruction set than x86...
Wait, are we talking about x86-64? I thought I had already been running a native Firefox on my x86-64 boxes for years.
I guess it should be managed like a cache, with your brain being the cache, and books being secondary storage
My thoughts exactly. While RAM and disk have grown by huge amounts, our CPU and cache are basically the same as in the caveman days. To get any thinking done, you still need to get that data into the cache, and frequently used data will tend to stay there.
This is also why huge organizations are inefficient, there is too much communication overhead, and it is mainly in small teams where actual work gets done.
I use Linux on a Powerbook with Radeon graphics. For some odd reason, AMD does not provide binary drivers for this platform, but the open driver works great.
On my other machines with x86-64, I use the binary Radeon drivers as it is the only way to get full OpenCL. Even there, it sometimes happens that I need to use an older kernel or disable some kernel features, as the binary driver does not play well with the pace of Linux development.
My GPUs paid for a nice FPGA devkit, after paying for themselves. If Bitcoin continues this way, the devkit may end up paying some more, as it is much more power efficient.
If you are referring something that people have already been running for a few months, you probably mean this. The modular miner thread seems to focus on raw price/performance, ignoring factors like the freedom to program the said FPGA without expensive licenses.
Who needs software when you can design hardware to do what you want?
Anybody who wants to distribute his work to the public without having to invest capital in manufacturing and shipping physical objects.
Just distribute the design files ;)
I just hack FPGAs, not trying to emulate old machines though. Who needs software when you can design hardware to do what you want?
Strangely enough, I've never really coded anything at an assembler level, not even much C, so it's not like I'm getting systematically deeper. Then again I have plenty of experience in hands-on electronics. It's great being able to define your data structures at a bitwise level, not worrying about some little endian crap. Also a nice way to get some ideas for parallel programming in software, because things are actually happening in parallel and not just timesliced.
Also, don't forget that FPGA is an acronym, I pronounce it as "fapgay".
I can attest that Verilog is easy and intuitive, since it got me from a total newbie to a project that other people find useful, in a couple of weeks. Of course, some electronics and computing background helped a lot.
My choice of language was partly influenced by the apparent verbosity of VHDL. It's the same reason why I preferred to learn Python instead of Java, for example. However, I also have the impression that VHDL is in some sense lower-level and more strictly defined, so you can have finer control on some things. For example, to do arithmetic in VHDL you need to explicitly load the requisite library. I'm not sure how much this influences the resulting circuit, though.
So are they using these instead of HDMI connectors?
How quaint, a time when "android" meant a humanlike robot.
How exactly can a transaction system that broadcasts all of your transactions to the network "make the world a little better place".
Because it's not Paypal ;)
Seriously though, I'm not saying that Bitcoin is perfect. That's kind of the point of geek things, we're trying to break some new ground here. But the specific point of broadcast transaction seems quite essential if you want to keep things distributed, so that other peers will verify your transaction.
The "exchanges" represent a mis-design of Bitcoin. There should have been a way to do an exchange in a distributed way, without the exchange holding customer assets. The NYSE and NASDAQ don't hold customer balances. Brokers do, but you can have your cash swept from a brokerage into a bank daily, or more often if the numbers get big. The Bitcoin exchanges are slow at delivering money - Mt. Gox has a daily transfer limit, and even when they were up, many users reported delays.
There are other exchanges that work more in the way you describe, for example I have personally used bitcoinmarket.eu. Too bad about the MtGox episode, but it has nothing to do with the rest of the Bitcoin-using world. Of course, USD becomes a pyramid scheme the minute somebody hacks into a forex site.
Also, let's not forget that Bitcoin is a geek thing, something that could make the world a little better place, and not just something to make you rich. For example, I learned to program FPGAs because of Bitcoin. I now have a new hobby with tons of interesting prospects, no matter how well Bitcoin fares in the future.
Also, not to nitpick (well, yes to nitpick), but I think that part that says "suddenly every piece of software works..." needs a bit of filling out. Especially at the "suddenly" part.
http://www.guardian.co.uk/uk/2001/nov/15/neilmcintosh
This made me realize that even paper books are essentially digital. Even after a hundred years and countless readings, an "a" is still an "a", and not some approximation. Some of the letters may wear out entirely, but the redundant coding of natural languages will maintain the integrity of the content, much like the coding in CDs. Nothing like this in vinyl records, it wears out a little every time you play it, and there is no way to get that information back exactly.
uterus transplants would be particularly useful in Saudi and other Muslim countries where using surrogate mothers is prohibited by Islamic law.
So it is forbidden to use someone else's womb to bear your child. Except when you take that womb and put it inside your own body.
Thus solving the problem once and for all!
Just like "Peer" is a generic term for an unidentified Internet user in Norway.
Died a virgin, a true geek.
ed is also two characters, and they are adjacent on the keyboard. If you are a touch-typist, it takes 50% of the hands and 50% of the fingers necessary to start that inefficient vi editor.
In the typing standard I've learned, E and D are on the same finger. It is much faster to use two fingers for two keys, so clearly vi wins. Not that either would matter, since I already have Emacs running.
And paging through? Spacebar dipshit. Firefox, Acrobat, will page when you press it. Also there are these little keys called "page up" and "page down". Wonder what THEY do?
The problem with the space bar as PgDn is that the equivalent of PgDn is not consistent across applications. Sometimes it is Shift+Space, sometimes B or something else.
Now, actual, dedicated PgUp/PgDn keys, these are something that I miss on many modern laptop keyboards. I guess these started to fade away when scroll wheels on mice became ubiquitous, but I still see many people who use neither. Instead, they click and drag on the scroll bar. For the hardcore users who still use old-fashioned PgUp/Dn, they must be accessed via Fn keys. Even arrow keys are shrinking, since nobody uses them any more. Obviously, everything is easier with the mouse, and soon we won't need the keyboard any more.
Finite wireless spectrum?!? What are you talking about? Let's talk Mhz:
There is Ghz spectrum between say, 2.4 and 3.4 Ghz, which seems limited. So you might break it out into 1 mhz bands, giving you 1,000 usable frequencies. Or break it more finely,into .1 mhz bands giving you 10,000 usable, or .01 giving you 100,000 frequencies, or...
A 0.01 MHz band does not give you much capacity, perhaps something of the order 0.1 Mbps. While bandwidth is not the same thing as data rate, they are proportional.
http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem
Spread spectrum technology, first developed by military for secretive radio communications, send information in short bursts in pseudorandom frequencies. This frquency hopping allows for far more efficient use of existing radio frquencies with minimal disruption. Numerous studies show this type of technology could extend the available bandwidth a billionfold or more.
By definition, spread spectrum uses a lot of bandwidth ;) The problem with data rate is that when everyone uses spread spectrum, noise floor goes up, and thus the signal/noise ratio gets worse. This, in turn, means a smaller data rate per bandwidth, as explained by Shannon.