You say that Bitcoin is elegant, but solves the wrong problem. I disagree. Read some of the introductory material that explains the flaws of the existing currency models that Bitcoin is trying to address:
You are correct that bootstrapping a currency is the most difficult step. However it appears to have done just that: it went from zero to hundreds of merchants in a little over a year: https://en.bitcoin.it/wiki/Trade
Of course, there is still a long way to go (most of them are individuals or very small shops), but the very fact it has already grown so quickly and so fast is very encouraging. It was possible because of its truly unique and revolutionary features that really no other currency provide, notably the combination of: absence of middlemen, anonymity, cryptographically irreversible money transfers.
I like to tell people that when they first hear about Bitcoin, they will either instantly dismiss it as something that could never work, or they will immediately understand its large potential:)
Bitcoin's existence is essentially technically insuppressible at this moment: it is fully peer-to-peer, there is no central server or company that can be shut down.
The only risk is social: in other words people could simply lose interest in it. But its network strength (hash rate) has demonstrated a remarkable exponential growth recently. It has doubled every 27 days, continuously, in the last 15 months: http://blog.zorinaq.com/?e=49 This implies, of course, popularity. So I would say it is very improbable for Bitcoin to just disappear tomorrow.
It is at the moment still quite profitable to mine on AMD GPUs, assuming your electricity costs are about $0.10/kWh. You will easily recoup at the very least half the price of a high-end video card if you jump in right now.
Also the network continuously self-adjusts itself to distribute coins over a long period of time, as this graph shows:
https://en.bitcoin.it/wiki/File:Total_bitcoins_over_time_graph.png
As you can see, we have only generated about 5 million coins so far, out of 21 million. Additionally, many expect the value of 1 Bitcoin to continue to rise, significantly, on the various exchanges over the next few years.
That's the nice thing about Bitcoin: it is less dependent on trust than all other payment systems.
Most significantly, as a Bitcoin user, you do not have to trust a central authority or 3rd party (eg. Paypal or a bank) that may initiate chargebacks, or freeze your accounts, because there is no middleman in Bitcoin.
Of course you need to trust the Bitcoin design itself. But the system has proven its robustness so far, and will (hopefully) continue to do so. If you are curious you may find a list of attacks or flaws, that the Bitcoin network has successfully repelled so far here: https://en.bitcoin.it/wiki/Incidents Personally I have rarely found an open source community that was so full of smart people. And I say this with 13+ years of using and contributing to a lot of open source projects.
This is untrue. At work I have set up many times Apache serving different HTTPS name-based virtual hosts using the same IP and port. It is possible by creating a certificate with different "subject alternative names" (for example by passing a configuration file to "openssl req -config xxx" defining subjectAltName=DNS:"foo.example.com",DNS:"bar.example.com"). Here is an example of a certifcate valid for both foo.example.com and bar.example.com. (All OSes support subject alt names, even XP!)
$ openssl x509 -noout -text < server.crt Certificate: Data: Version: 3 (0x2) Serial Number: xx:xx:xx:xx:xx:xx:xx:xx Signature Algorithm: sha1WithRSAEncryption Issuer: CN=My Custom CA Validity Not Before: Oct 11 00:00:00 2007 GMT Not After : Oct 12 00:00:00 2020 GMT Subject: CN=foo.example.com Subject Public Key Info: Public Key Algorithm: rsaEncryption RSA Public Key: (3072 bit) Modulus (3072 bit): ... X509v3 extensions: X509v3 Basic Constraints: CA:FALSE X509v3 Subject Key Identifier: xx:xx:xx:xx:xx:xx:xx:xx:xx X509v3 Authority Key Identifier: keyid:xx:xx:xx:xx:xx:xx:xx:xx:xx X509v3 Subject Alternative Name: DNS:foo.example.com, DNS:bar.example.com Signature Algorithm: sha1WithRSAEncryption ...
I have worked with dozens of 1U servers from Dell, HP, Supermicro, whitebox, etc, and I have never seen the RAM DIMMs at a 45 degrees angle. Last time DIMMs were at a 45 degrees was 15+ years ago (EDO RAM I think?)
What's new, compared to other past announcements that "GSM has been broken" is that, 3 days ago, the A5/1 Project just wrote the piece of code to perform lookups in the "Berlin rainbow table set". The table set is 2TB and has been computed some time ago and can be obtained from various origins (the project member who wrote the lookup code --Frank Stevenson-- offered the arrangement of swapping preloaded disks for cash at the Schiphol airport). See my blog for some more info about these recent developments.
Just like these claw crane games that you find in vide arcades or amusement parks: those with a joystick to control a crane to grab stuffed animals or whatnot. Curious, one day I browsed the web to find operator's manuals, and they are programmed to make it look like the crane accidentally drops objects. The operator can enter parameters to define the average price of prizes, the average winning rate, etc so that in the end, just like slot machines, the payout percentage can be controlled very precisely. For more info read Machine configuration and chances of winning.
Knowing this completely takes the fun out of it, doesn't it ?
Disabling javascript is not sufficient. The malicious site could very well redirect to the malicious page after a long period of time, say 10min, with:
Although it is a little less sophisticated, it would work. Personally I have always been using 2 browsers for other reasons (to defend myself against CSRF vulnerabilities) and it turns out that doing so also protects me from 'tabnapping', even though CSRF and tabnapping are 2 completely different attacks. I described my setup here. This is a good example of defense-in-depth: using a security policy that ends up preventing future attacks that were unknown at the time the policy was implemented:-)
Why waste 2 whole spindles on ZFS root pool? Partition each of your 5x1.5TB with a small 20-30GB slice, and a bigger one covering the rest of the disk. Create a 3-way mirror on 3 of the small slices for the root pool, with the 2 remaining small slices as hot spares. Create a 5-way raidz on the 5 large slices for your data pool.
And the 8-core Opteron 6128 2.0GHz starts at $280.
Compare to Intel's cheapest 8-core processor, the Xeon x6550 2.0GHz that sells for $2500. ten times the price! When people tell you AMD provides more value per $, here is the proof!
First link: author is vague and incorrect; OpenSolaris supports most common onboard SATA controllers. I have personally run it on nVidia MCP55 and above, Intel ICH7 and above, AMD SB600 and above, and OpenSolaris usually support all these very common chipsets/onboard SATA controllers.
Second link: the author is using unsupported dev builds of OpenSolaris.
Third link: the post is 2 years old and evidence suggests unreliable hardware.
Fourth link: the author complains about FreeBSD, not OpenSolaris.
Fifth link: the author concluded corruption was caused by unreliable hardware.
Search for "$NAME_OF_TECHNOLOGY unreliable" and google will always return thousands of results.
Personally I have a rather pleasant experience with ZFS. I have been using it for 3+ years at work and at home on 5-6 machines with about 50 drives total. It has been rock solid so far. And it has saved my life a couple times when drives died.
But most users don't care how much theoretical space a "32-bit filesystem" has. They have 1TB drives and want to know practically how many hours of high-def videos they can store on it, how many Bluray movies they can rip. Try doing your computations with a "1TB" drive mixed with power of 2. prefixes.
I have posted the following mulitple times on slashdot: contrary to popular belief, power of 10 prefixes are much more common than power of 2 prefixes in the computer industry. The only few places where the latter are used are to refer to RAM capacities and file sizes, whereas power-of-10 prefixes apply to most other areas and all units (not "only bitrates", as some claim): storage capacity, clock frequency, stream bandwidth, baud, pixel numbers, data throughput, processing power, etc.
An 32 GB USB flash drive is 32 * 10^9 byte (power of 10)
A 16 GB SD card is 16 * 10^9 byte (power of 10)
A 50 GB dual-layer Blu-ray Disc is 50 * 10^9 byte (power of 10)
A 4.7 GB single-layer DVD is 4.7 * 10^9 byte (power of 10)
A 2.5 GHz processor is 2.5 * 10^9 Hz (power of 10)
A PC6400 (as in 6400 MByte/s) memory module is 6400 * 10^6 byte/s (power of 10)
A 25.6 GFLOPS CPU core is 25.6 * 10^9 FLOPS (power of 10)
A 128 kbit/s audio stream is 128 * 10^3 bit/s (power of 10)
An 8 kbaud V.92 modem is 8 * 10^3 baud (power of 10)
A 6 Mpixel digital camera is 6 * 10^6 pixel (power of 10)
A 4000 MB/s HyperTransport link is 4000 * 10^6 byte/s (power of 10)
A 480 Mbit/s USB2 link is 480 * 10^6 bit/s (power of 10)
A 5.0 Gbit/s PCI-E 2.0 lane (after 8b/10b encoding) is 5.0 * 10^9 bit/s (power of 10)
A 500 MB/s PCI-E 2.0 lane (before 8b/10b encoding) is 500 * 10^6 byte/s (power of 10)
A 1 Gbit/s ethernet card is 1 * 10^9 bit/s (power of 10)
A 54 Mbit/s 802.11g network is 54 * 10^6 bit/s (power of 10)
A 6.0 Gbit/s SATA link (after 8b/10b encoding) is 6.0 * 10^9 bit/s (power of 10)
A 600 MB/s SATA link (before 8b/10b encoding) is 600 * 10^6 byte/s (power of 10)
A 6 Mbit/s DSL line is 6 * 10^6 bit/s (power of 10)
Curiosity: a 1.44 MByte floppy disk is 1.44 * 1000 * 1024 byte (mix of power of 10 and 2)
And of course, a 1.5 TByte hard disk drive is 1.5 * 10^12 byte (power of 10)
I realize my comment might be taken too literally by some readers... The point I was trying to make was that theoretical GFLOPS ratings thrown around don't necessarily mean something.
A practical GFLOPS rating that should be reachable by an HD 5970 in the SGEMM benchmark is about 1.2 TFLOPS when scaling the numbers from the FireStream 9270 (300 GFLOPS in SGEMM), because the HD 5970 can execute 3.87x more instructions per cycle than the FireStream 9270 ((3200SPUs*725MHz) / (800SPUs*750MHz)), and its memory bandwidth is more than 3.87x higher... Therefore an HD 5970 should still beat the Larrabee chip that was benchmarked.
Firstly, Larrabee is vaporware right now as it was pointed out by one of my siblings.
Secondly, the 1 TFLOPS figure is incorrect. The real one is higher. Intel claims up to 32 cores @ 2.0GHz, so with the 512-bit LRBni instruction set that means Larrabee could reach 2 TFLOPS. But wait! Don't get your hopes to high, because your article conveniently fails to point out that, as of right now, ATI's HD 5970 rates 4.6 TFLOPS, 2.3x faster than Larrabee! So we have vaporware that is effectively already theoretically outperformed by today's products...
ATI are about to become the leader? They are already the leader in all categories: perf/$, perf/W, absolute perf, and at all price points. See list below. For gaming performance, the GFLOPS rating are a roughly (+/- 30%) good enough approximation to compare ATI vs. Nvidia. For GPGPU performance, the GFLOPS rating is actually unfair to ATI because Nvidia's GT200 microarchitecture causes it to be artificially inflated (they assume a MUL+MAD pair executing 3 floating-point op per cycle, whereas ATI assumes a regular fused MAD executing 2 floatting-point ops per cycle). Meaning that an ATI GPU rated 200 GFLOPS actually executes ALU-bound workloads as fast as an Nvidia GPU rated 300 GFLOPS. ATI's lead is such that it's not even funny anymore. There are rumors of Nvidia killing the high-end (GTX 285, 295) to focus only on the extreme entry-level segment (sub-$100). And GT300 (Fermi) will not enter mass production before the end of Q1 2010. I am concerned by the lack of competition... ATI is free to impose whatever price structure they want.
If you have $500+ to spend: ATI HD 5970 (4640 GFLOPS, 294 Watt, ~$600) vs. Nvidia GTX 295 (1843 GFLOPS, 289 Watt, ~$500).
If you have ~$400 to spend: ATI HD 5870 (2720 GFLOPS, 188 Watt, ~$410) vs. Nvidia GTX 285 (1063 GFLOPS, 204 Watt, ~$400).
If you have ~$300 to spend: ATI HD 5850 (2088 GFLOPS, 151 Watt, ~$310) vs. Nvidia GTX 275 (1011 GFLOPS, 219 Watt, ~$300).
If you have ~$200 to spend: ATI HD 5770 (1360 GFLOPS, 108 Watt, ~$170) vs. Nvidia GTX 260 Core 216 (805 GFLOPS, 182 Watt, ~$200).
If you have ~$150 to spend: ATI HD 5750 (1088 GFLOPS, 86 Watt, ~$155) vs. Nvidia GTX 260 (715 GFLOPS, 182 Watt, ~$170).
If you have ~$100 to spend: ATI HD 4770 (960 GFLOPS, 80 Watt, ~$110) vs. Nvidia GTS 250 (470 GFLOPS, 145 Watt, ~$110).
You are wrong (1000x difference between your numbers -- the summary said 100x).
Also many other posters are wrong when saying that European use dots to separate thousands. Some countries may do it, but precisely because of these different and confusing conventions, French schools teach students to use spaces to separate thousands, eg. 11 500 Euro, 115 000 Euro, 1 150 000 Euros.
Slashdot Mirror
You say that Bitcoin is elegant, but solves the wrong problem. I disagree. Read some of the introductory material that explains the flaws of the existing currency models that Bitcoin is trying to address:
You are correct that bootstrapping a currency is the most difficult step. However it appears to have done just that: it went from zero to hundreds of merchants in a little over a year: https://en.bitcoin.it/wiki/Trade
:)
Of course, there is still a long way to go (most of them are individuals or very small shops), but the very fact it has already grown so quickly and so fast is very encouraging. It was possible because of its truly unique and revolutionary features that really no other currency provide, notably the combination of: absence of middlemen, anonymity, cryptographically irreversible money transfers.
I like to tell people that when they first hear about Bitcoin, they will either instantly dismiss it as something that could never work, or they will immediately understand its large potential
Here is a video, for non-technical persons, to help understand Bitcoin:
http://www.weusecoins.com/
It was just made very recently (today!)
Bitcoin's existence is essentially technically insuppressible at this moment: it is fully peer-to-peer, there is no central server or company that can be shut down.
The only risk is social: in other words people could simply lose interest in it. But its network strength (hash rate) has demonstrated a remarkable exponential growth recently. It has doubled every 27 days, continuously, in the last 15 months: http://blog.zorinaq.com/?e=49 This implies, of course, popularity. So I would say it is very improbable for Bitcoin to just disappear tomorrow.
The goldrush is certainly not over.
It is at the moment still quite profitable to mine on AMD GPUs, assuming your electricity costs are about $0.10/kWh. You will easily recoup at the very least half the price of a high-end video card if you jump in right now. Also the network continuously self-adjusts itself to distribute coins over a long period of time, as this graph shows: https://en.bitcoin.it/wiki/File:Total_bitcoins_over_time_graph.png As you can see, we have only generated about 5 million coins so far, out of 21 million. Additionally, many expect the value of 1 Bitcoin to continue to rise, significantly, on the various exchanges over the next few years.
That's the nice thing about Bitcoin: it is less dependent on trust than all other payment systems.
Most significantly, as a Bitcoin user, you do not have to trust a central authority or 3rd party (eg. Paypal or a bank) that may initiate chargebacks, or freeze your accounts, because there is no middleman in Bitcoin.
Of course you need to trust the Bitcoin design itself. But the system has proven its robustness so far, and will (hopefully) continue to do so. If you are curious you may find a list of attacks or flaws, that the Bitcoin network has successfully repelled so far here: https://en.bitcoin.it/wiki/Incidents Personally I have rarely found an open source community that was so full of smart people. And I say this with 13+ years of using and contributing to a lot of open source projects.
I have worked with dozens of 1U servers from Dell, HP, Supermicro, whitebox, etc, and I have never seen the RAM DIMMs at a 45 degrees angle. Last time DIMMs were at a 45 degrees was 15+ years ago (EDO RAM I think?)
And Linux has bcache.
And Solaris has L2ARC for ZFS (zpool add XXX cache cXdXtX).
Both of these technologies allow the OS to uses a fast SSD as a general cache device for data on slower HDDs.
So, instead you turned all the comments to pedantic responses about relativity of simultaneity.
See what you did?
What's new, compared to other past announcements that "GSM has been broken" is that, 3 days ago, the A5/1 Project just wrote the piece of code to perform lookups in the "Berlin rainbow table set". The table set is 2TB and has been computed some time ago and can be obtained from various origins (the project member who wrote the lookup code --Frank Stevenson-- offered the arrangement of swapping preloaded disks for cash at the Schiphol airport). See my blog for some more info about these recent developments.
Just like these claw crane games that you find in vide arcades or amusement parks: those with a joystick to control a crane to grab stuffed animals or whatnot. Curious, one day I browsed the web to find operator's manuals, and they are programmed to make it look like the crane accidentally drops objects. The operator can enter parameters to define the average price of prizes, the average winning rate, etc so that in the end, just like slot machines, the payout percentage can be controlled very precisely. For more info read Machine configuration and chances of winning.
Knowing this completely takes the fun out of it, doesn't it ?
Disabling javascript is not sufficient. The malicious site could very well redirect to the malicious page after a long period of time, say 10min, with:
Although it is a little less sophisticated, it would work. Personally I have always been using 2 browsers for other reasons (to defend myself against CSRF vulnerabilities) and it turns out that doing so also protects me from 'tabnapping', even though CSRF and tabnapping are 2 completely different attacks. I described my setup here. This is a good example of defense-in-depth: using a security policy that ends up preventing future attacks that were unknown at the time the policy was implemented :-)
There is an oil spill in the Gulf?
Why waste 2 whole spindles on ZFS root pool? Partition each of your 5x1.5TB with a small 20-30GB slice, and a bigger one covering the rest of the disk. Create a 3-way mirror on 3 of the small slices for the root pool, with the 2 remaining small slices as hot spares. Create a 5-way raidz on the 5 large slices for your data pool.
And the 8-core Opteron 6128 2.0GHz starts at $280.
Compare to Intel's cheapest 8-core processor, the Xeon x6550 2.0GHz that sells for $2500. ten times the price! When people tell you AMD provides more value per $, here is the proof!
WOOOOOOOOOSH.
First link: author is vague and incorrect; OpenSolaris supports most common onboard SATA controllers. I have personally run it on nVidia MCP55 and above, Intel ICH7 and above, AMD SB600 and above, and OpenSolaris usually support all these very common chipsets/onboard SATA controllers.
Second link: the author is using unsupported dev builds of OpenSolaris.
Third link: the post is 2 years old and evidence suggests unreliable hardware.
Fourth link: the author complains about FreeBSD, not OpenSolaris.
Fifth link: the author concluded corruption was caused by unreliable hardware.
Search for "$NAME_OF_TECHNOLOGY unreliable" and google will always return thousands of results.
Personally I have a rather pleasant experience with ZFS. I have been using it for 3+ years at work and at home on 5-6 machines with about 50 drives total. It has been rock solid so far. And it has saved my life a couple times when drives died.
I have posted the following mulitple times on slashdot: contrary to popular belief, power of 10 prefixes are much more common than power of 2 prefixes in the computer industry. The only few places where the latter are used are to refer to RAM capacities and file sizes, whereas power-of-10 prefixes apply to most other areas and all units (not "only bitrates", as some claim): storage capacity, clock frequency, stream bandwidth, baud, pixel numbers, data throughput, processing power, etc.
Yeah indeed. An interesting comparison to make too is the 25 years estimated to build the measly 800-mile high-speed train project in California (est. completion by 2035), whereas China is planning what appears to be a roughly 10000-mile project to be completed in 10 years...
Hahaha... Yes I am.
- Vin Diesel.
I realize my comment might be taken too literally by some readers... The point I was trying to make was that theoretical GFLOPS ratings thrown around don't necessarily mean something.
A practical GFLOPS rating that should be reachable by an HD 5970 in the SGEMM benchmark is about 1.2 TFLOPS when scaling the numbers from the FireStream 9270 (300 GFLOPS in SGEMM), because the HD 5970 can execute 3.87x more instructions per cycle than the FireStream 9270 ((3200SPUs*725MHz) / (800SPUs*750MHz)), and its memory bandwidth is more than 3.87x higher... Therefore an HD 5970 should still beat the Larrabee chip that was benchmarked.
Firstly, Larrabee is vaporware right now as it was pointed out by one of my siblings.
Secondly, the 1 TFLOPS figure is incorrect. The real one is higher. Intel claims up to 32 cores @ 2.0GHz, so with the 512-bit LRBni instruction set that means Larrabee could reach 2 TFLOPS. But wait! Don't get your hopes to high, because your article conveniently fails to point out that, as of right now, ATI's HD 5970 rates 4.6 TFLOPS, 2.3x faster than Larrabee! So we have vaporware that is effectively already theoretically outperformed by today's products...
ATI are about to become the leader? They are already the leader in all categories: perf/$, perf/W, absolute perf, and at all price points. See list below. For gaming performance, the GFLOPS rating are a roughly (+/- 30%) good enough approximation to compare ATI vs. Nvidia. For GPGPU performance, the GFLOPS rating is actually unfair to ATI because Nvidia's GT200 microarchitecture causes it to be artificially inflated (they assume a MUL+MAD pair executing 3 floating-point op per cycle, whereas ATI assumes a regular fused MAD executing 2 floatting-point ops per cycle). Meaning that an ATI GPU rated 200 GFLOPS actually executes ALU-bound workloads as fast as an Nvidia GPU rated 300 GFLOPS. ATI's lead is such that it's not even funny anymore. There are rumors of Nvidia killing the high-end (GTX 285, 295) to focus only on the extreme entry-level segment (sub-$100). And GT300 (Fermi) will not enter mass production before the end of Q1 2010. I am concerned by the lack of competition... ATI is free to impose whatever price structure they want.
You are wrong (1000x difference between your numbers -- the summary said 100x).
Also many other posters are wrong when saying that European use dots to separate thousands. Some countries may do it, but precisely because of these different and confusing conventions, French schools teach students to use spaces to separate thousands, eg. 11 500 Euro, 115 000 Euro, 1 150 000 Euros.