"You will destroy the way we live and since we can't imagine any species living in a way other than we do you will destroy yourself. Fools!"
Really, that's pretty much what the old experienced dinosaurs told the young aspiring mammals.
The poster's reasoning goes like this:
I invested a lot of time and money in my education.
So i want to keep software prices high to get a decent return on my investment.
Someone, maybe someone younger, offers his services at lower prices.
I have to convince them to stop competing with me. After all, competition ist bad for both of us.
Face it: no one can escape the market. Competition is a fact of life. Maybe in a country where you are not allowed to open a computer repair shack without a 100kEUR Meisterbrief, you may for a while be able to hang on to the illusion that you live on a privileged island where competition does not exist and you'll never have to adapt to a changing environment, but ultimately, either you get the market or the market gets you. Just ask the dinosaurs.
A tax on sunlight? Something similar has been suggested before. More than 150 years ago Frederic Bastiat wrote his
satire where the light manufaturing industries complain about the unfair foreign competition from the sun and ask the lawmakers to outlaw windows. Excerpts:
"We are suffering from the ruinous competition of a foreign rival who apparently works under conditions so far superior to our own for the production of light that he is flooding the domestic market with it at an incredibly low price; for the moment he appears, our sales cease, all the consumers turn to him, and a branch of French industry whose ramifications are innumerable is all at once reduced to complete stagnation. This rival, which is none other than the sun, is waging war on us so mercilessly that we suspect he is being stirred up against us by perfidious Albion (excellent diplomacy nowadays), particularly because he has for that haughty island a respect that he does not show for us."
"We ask you to be so good as to pass a law requiring the closing of all windows, dormers, skylights, inside and outside shutters, curtains, casements, bull's-eyes, deadlights, and blinds--in short, all openings, holes, chinks, and fissures through which the light of the sun is wont to enter houses, to the detriment of the fair industries with which, we are proud to say, we have endowed the country, a country that cannot, without betraying ingratitude, abandon us today to so unequal a combat."
you are assuming that laws are logical. Let me challenge that assumption: here in Germany we pay sort of a tax on blank media and recorders. Music industry is even trying to broaden the scope of these royalties: they are currently pushing for a copy tax on printers (older link here.).
In addition to that, there is an entity called GEMA which makes sure that radio stations pay for each song they play. Public radio and TV cost consumers a monthly fee, too.
Recently they made a new copyright law. Copying for private use used to be legal, and strictly by the letter of the law still is, but circumventing copy protection mechanisms in order to do something the law explicitly allows you to do is now illegal. In other words: They didn't outlaw crossing the road. They made touching the ground with your feet while crossing the road a crime.
So consumers over here are forced to pay for the same product multiple times. All attempts to set that straight have failed so far. I have a hunch that this kind of legal creativity may become an exportschlager.
Where exectly did i claim that UltraSPARC was not 64 bit? Where did you get that from?
The article i mentioned shows pretty clearly that on a diagram SPECfp vs SPECint for a the common 64 bit processors there are two groups. For brevity, let's call them "fast" and "slow". Intel, AMD, PPC, Alpha, and SPARC64 are in the "fast" group, UltraSPARC and the others are "slow". Let me quote the author, Paul DeMone:
What is particularly striking [..] is the prominent bimodal division of high end processors into "have" and "have not" camps of uniprocessor performance. It is apparent that only the most technically competent RISC processors can keep up with the blistering performance pace set by 32 bit x86 MPUs.
His statement may be a little blunt, but it is true that Sun servers used to outperform intel-based Linux-servers. They no longer do, and the market has noticed that. If Sun wants to sell hardware in the future, they'll have to correct that.
Look at this Slashdot story from Feb 2003. The article referenced there shows that in 64bit land there are haves (Intel, AMD, IBM, SPARC64 and still around: Alpha) and have-nots (everybody else, including UltraSPARC, PA_RISC and MIPS). For Sun, the most logical way to stay competitive in the performance race and get out of the losers' gene pool is to join forces with Fujitsus SPARC64 program. So it looks like the natural thing to do.
As Feynman says and Dell shows, here's always room at the bottom. As the existence of Apple shows, there usually is room at the top, too. In between, that's where the crowd is. To move to an intel platform is not the issue for Apple, and it never was. Not becoming just part of the crowd when doing so is.
Look at SUN. They made the best machines you could buy for internet applications at a time back in the nineties, and charged you a lot of money for it. Today the rest of the world has caught up, we all stack our racks with linux pizzaboxen now, and SUN is in trouble. The company has to decide: is SUN a hardware company? that would mean investing a lot in the development of SPARC, killing the Solaris x86 line and fighting Linux, or move entirely to Intel, giving up software development altogether and become like Mike. Or is SUN a software company? that would mean cancelling further SPARC development and concentrating on Solaris and Java. Eventually, this would kill SPARC.
Strengthening the hardware section in SUN would hurt the software guys, and beefing up the software department could easily hurt the hardware sales. Not a good strategic position. Apple could easily be (or have been) caught in the same situation. To compete with Dell you have to become like Dell. If you don't want to do this, you must find a different market for yourself. Or be just part of the crowd.
This is more pretty-printing than real innovation. They claim to arrange data by relation but the thing still knows active folders, parent folders and subfolders. And the color scheme (subfolders are blue) focuses on the hierarchical structure of the folders and not the relation of the data. So they took one way of organizing and presenting files that works for most people most of the time but has a few big shortcomings, pretty-printed it in a somewhat confusing way and added relational sugar that can only add to the confusion.
In Germany, the toll collect system was supposed to begin September 1st. The implementing consortium (bigshots like DaimlerChrysler and Deutsche Telekom) missed that deadline, so it was pushed back to November 1st, and this new deadline is in doubt, too. The On-Board-Units that every truck has to carry are
not enough: the consortium initially figured they'd need about 50000 units for the entire country
defective: as many as half of the delivered units show defects. Some even self-destruct on starting the truck's engine
expensive: they cost several hundred Euros apiece, and logistics companies who have to buy them are complaining
PR-wise it's a disaster.
Last week, manager-magazin.de ran an interview with Peter Newole, executive at Austria's equivalent, Europpass system, which does the same things TollCollect is supposed to do, only that it's cheaper and actually works. (The interview is in German, sorry.) Basically, what Mr. Newole says is that the two systems are doing similar things in vastly different ways. In both countries, trucks have to carry boxes that can communicate with base stations to register their location. Based on these location profiles, the toll is calculated. But the design of these boxes is completely different:
The Austrian boxes are dumb clients. These things can receive a microwave signal and respond with an ID or, if the box thinks it has been tampered with, the ID and an alarm signal. They can record the IDs of the last dozen base stations passed. They know when their batteries run low. Their user interface consists of beeps: one for a successful pass of a base station, two if it's a prepaid box and the prepaid account has hit zero, many if the batteries have to be changed. Batteries have to be changed every five years and this can be done in a number of places throughout Austria. The boxes are provided to the trucks at no cost and all the driver has to do is glue it to the windscreen.
The German on-board units (OBU) are smart clients. They are supposed to know the license plate of the truck the're installed in, be able to calculate the OBU's location via GPS and transmit this information via the GSM cell phone net to its servers, so these things are GPS receivers and cell phones combined. They are troo big for batteries and have to be hooked up with the truck's electricity circuit, and there is a complicated setup procedure to tell the OBU what truck it's installed in. And the original schedule allowed all of eleven months for development, testing, mass production, deployment and user training of these OBUs.
The design lesson is obvious: The more of something you are going to deploy, the simpler it has to be. Put the logic into the servers and make the clients as dumb as you can.
The old sentence gets even more valid with TCPM. We all are used to the fact that binaries are tied to the computers we're using. Buy a new computer to replace your crashed old one, find out you have to use a newer version of the kernel to support Bozo Gadget 2.78, reinstall your binary application from backup, find out that your new system has glibc 3.14 where the old binaries were linked against libc 1.41, yell an expletive, dig for the source, recompile. Been there, done that.
With TCPM, this will only be stricter, not fundamentally different. Use the Source and you'll be fine.
SCO Germany has to pay a 10000 Euro fine. This decision by the Landgericht Munchen I is based on a preliminary injunction against SCO, granted to the Tarent company and LinuxTag. According to it, SCO may not claim that Linux contains illegally obtained intellectual property from SCO. SCO has allegedly violated this on its homepage, therefore in June Tarent asked for a trial.
According to a Tarent GmbH statement, the court accuses SCO of "negligence" in running its company's homepage, which, even after the preliminary injunction allegedly read: "End users who use the Software Linux may be held liable for violations of SCO's intellectual property."
Tarent attorney Till Jaeger sees the court's decision as confirming that SCO's allegations are "massively business-damaging statements" concerning an "extremely sensitive issue." Unproven claims were used to do fear-induced business on the expense of others.
SCO Germany is currently not available for comment. In early June, asked about the
trial, SCO Germany CEO Hans Bayer emphasised:
"Our intention was to comply." and that the violation of the preliminary incunction had not been intentional.
Maybe they're speculating that the jury will immediately succumb to the magic word 'hash'.
But otherwiese, frankly, i don't see what this could be good for. Hashes (whether MD5 or SHA or some other algorithm) don't prove a thing.
Identity: The identity of the hashes of two MP3s only provey that the MP3s were encoded with identical settings from an identical CD source. If two people, one in NY the other in LA buy the latest Red Hot Chili Peppers album and rip and encode it both on Windows machines using identical versions of RealOne (or any encoder) then the resulting MP3s will have identical hashes. Whether the probability of two different files accidentally having the same hash ist 1 in 2 or 1 in 2^127 is absolutely irrelevant here. The chances of two people using the same software with the same CDDB information to rip the same track from a CD that sold a million copies is a lot higher. Everybody with a half episode of Matlock legal expertise will tear the RIAAs position apart on this ground.
Trackability: Hashes cannot be used to reliably track the path of copies across P2P networks either. Since the hash is more sensitive to minor changes than the ear doing random changes to the ID3 tags or randomly changing a bit or two somewhere in the MP3 will wipe the tracks.
So two files having the same hash doesn't prove they come from a single origin. Two files having different hashes doesn't prove they don't come from a single origin.
This is most likely not a good idea, and not primarily because of the current job market. Chances are high that you will find the same situation elsewhere, too. So moving to a different job (if you can do it at all) will most likely not improve your situation. If you can't manage your customers in your current job, you probably won't manage them anywhere.
There are reasons why customers behave in an insane way sometimes. The most common one is: they think you work for free. Not literally, but in a large organization few people have to pay directly for the work of the IT staff. So to them your work is free as in beer. And if you hand out beer for free, people have an incentive to get drunk. Of course overload will cause headache (as in beer), but the incentive is still there.
The solution is to make them pay for your work. If you can't do this in dollars and cents, make them work. Force them to go through your boss. Make them take the blame if a top banana at your organization can't use powerpoint for a day because you are busy fixing minesweeper installations.
I recently switched from bogofilter to SpamBayes. While it still shows the minor issues a young project always has (incompatibility with the dumbdb in Python 2.2.2 of SuSE 8.2 so you have to use gdbm as the internal db driver etc), i consider it one of the most promising spamfilters around.
It has several frontends. There's the outlook filter, there's a hammiefilter.py program that can be used with procmail as well as on the server side and there's a pop3proxy that does just that and can be used with a HTML interface to retrain the filter in a quick and easy way.
It has three categories: Spam, Ham, and Unsure. This actually helps a lot. I don't redirect spam to/dev/null, so i have to take out the trash manually. This becomes a lot easier if the filter presents you 50+ mails "I'm sure this is spam" and 10 more "Better double-check these".
It has fewer false negatives than bogofilter and works with smaller training sets. In my experience bogofilter is inefficient when trained with less than 400 spam mails. Even with larger sets, bogofilter used to filter only about 80% of the actual spam i receive. SpamBayes on the other hand runs on my machine trained with currently 262 spam and 288 ham, all collected in the past few months. Last week, i received 71 emails, 16 ham and 55 spam. SpamBayes classified all 16 hams as ham, 12 as unsure and 43 as spam. False Positives: 0, false Negatives: 0.
Sure, it's only one data point, and next week will be different, but i think i'll stick with SpamBayes for now.
Every expensive product in human history that faced cheaper (free is just the extreme) competition has at one point resorted to insulting their customers by calling them cheapos. "Freeloader Mentality" is a very hollow word that describes the simple fact that people make many of their economic decisions in a surprisingly economic way: As long as major news sites are free (as in beer), people won't pay for yet another one that charges them. It's that simple and ist's called competition.Get used to it.
...and all Gambling is Investment. That's just the way the ruble rolls.
What i don't get is why anyone---unless they spent the last five years in a coma---would trust the markets to have any predictive power on an issue like this.
Me, i'd bet a thousand that Saddam is either alive or dead, or somewhere in between.
The particular hash they attack stores up to 14 bytes of password data. However, these 14 are split in two chunks of 7, then converted to uppercase then hashed, then concatenated. This is about the stupidest possible way to waste entropy. It means that alphanumeric passwords have effectively at most log_2 (26+10) bits \approx 5.1 bits of entropy per keystroke. Each chunk will have seven times this entropy (seven bytes), and both chunks combined will have an entropy of (7 * lg_2 36) + 1 \approx. 37.1 bits (They state that in their paper).
With a sufficiently high number of key chains that amounts to 1.4 GB of storage and around 13.6 s on a modest 1.5GHz Intel P4 with 512MB RAM. Now, UNIX of course has bigger keys and salt. With salt, you'd need 4096 times the table size and 2^(56-37) = 524288 times the time (this is of course a ballpark estimate, i doubt that the actual calculation scales this easily, especially the memory lookups will suffer heavily from bigger tables). This means with a table size of 5.6 TB you'd be able to crack a DES password (and actually DES itself) in about 82.5 days.
Ballpark estimate or not, considering what it took to build the machine that was used for "Cracking DES" a couple of years ago, this is pretty impressive work.
If that is the source then the original quote is wrong. The footnote clearly states that in an earlier calculation they had included a worst-case assumption that 80% of the machines would have to run VMware for 4 to 5 years. A commitment from SuSE to guarantee lower costs for this made that assumption obsolete. That's what the report actually says.
I am German, but not a lawyer, but i think this is unconstitutional. Paragraph (1) says clearly that everyone has the right to access information from generally available sources. The internet is such a source. Nazi propaganda, while morally repulsive and disgusting, is information. So article 5 applies to the sites in question.
Paragraph (2) then sets a few exceptions to the rule. One of these is to protect children, one says that insults are not protected by article 5, and then the paragraph (2) says something about paragraph (1) being limited by the general laws. This means that a specific law may limit paragraph (1) in specific cases. It does not mean that there is an unwritten codex of what may be said and what not and that a local authority could decide on an ad hc basis what they like or not (that would be censorship) and then prosecute whoever they want. I believe the law the local authority was thinking about is the one that forbids the dissemination of Nazi Propaganda.
The problem with this point of view is that while distributing Mein Kampf in Germany is illegal, possessing it isn't. Neither is transporting it (otherwise the entire Deutsche Post AG would go to jail). So neither the ISP nor the persons accessing the websites in question are doing anything that would be against the law that forbids distribution of Nazi propaganda. And if they don't, the law doesn't apply and paragraph (1) stands.
The press release in which they blame the ISP for yielding to racists is the usual whining. They made a mistake, they don't want to admit it, so they call everybody a Nazi. The only new thing is that i didn't know that Godwin's law applies to muggels.
Slashdot Mirror
"You will destroy the way we live and since we can't imagine any species living in a way other than we do you will destroy yourself. Fools!"
Really, that's pretty much what the old experienced dinosaurs told the young aspiring mammals.
The poster's reasoning goes like this:
- I invested a lot of time and money in my education.
- So i want to keep software prices high to get a decent return on my investment.
- Someone, maybe someone younger, offers his services at lower prices.
- I have to convince them to stop competing with me. After all, competition ist bad for both of us.
Face it: no one can escape the market. Competition is a fact of life. Maybe in a country where you are not allowed to open a computer repair shack without a 100kEUR Meisterbrief, you may for a while be able to hang on to the illusion that you live on a privileged island where competition does not exist and you'll never have to adapt to a changing environment, but ultimately, either you get the market or the market gets you. Just ask the dinosaurs.A tax on sunlight? Something similar has been suggested before. More than 150 years ago Frederic Bastiat wrote his satire where the light manufaturing industries complain about the unfair foreign competition from the sun and ask the lawmakers to outlaw windows. Excerpts:
you are assuming that laws are logical. Let me challenge that assumption: here in Germany we pay sort of a tax on blank media and recorders. Music industry is even trying to broaden the scope of these royalties: they are currently pushing for a copy tax on printers (older link here.).
In addition to that, there is an entity called GEMA which makes sure that radio stations pay for each song they play. Public radio and TV cost consumers a monthly fee, too.
Recently they made a new copyright law. Copying for private use used to be legal, and strictly by the letter of the law still is, but circumventing copy protection mechanisms in order to do something the law explicitly allows you to do is now illegal. In other words: They didn't outlaw crossing the road. They made touching the ground with your feet while crossing the road a crime.
So consumers over here are forced to pay for the same product multiple times. All attempts to set that straight have failed so far. I have a hunch that this kind of legal creativity may become an exportschlager.
Where exectly did i claim that UltraSPARC was not 64 bit? Where did you get that from?
The article i mentioned shows pretty clearly that on a diagram SPECfp vs SPECint for a the common 64 bit processors there are two groups. For brevity, let's call them "fast" and "slow". Intel, AMD, PPC, Alpha, and SPARC64 are in the "fast" group, UltraSPARC and the others are "slow". Let me quote the author, Paul DeMone:
His statement may be a little blunt, but it is true that Sun servers used to outperform intel-based Linux-servers. They no longer do, and the market has noticed that. If Sun wants to sell hardware in the future, they'll have to correct that.
Please RTFABYP next time, will yer?
Look at this Slashdot story from Feb 2003. The article referenced there shows that in 64bit land there are haves (Intel, AMD, IBM, SPARC64 and still around: Alpha) and have-nots (everybody else, including UltraSPARC, PA_RISC and MIPS). For Sun, the most logical way to stay competitive in the performance race and get out of the losers' gene pool is to join forces with Fujitsus SPARC64 program. So it looks like the natural thing to do.
As Feynman says and Dell shows, here's always room at the bottom. As the existence of Apple shows, there usually is room at the top, too. In between, that's where the crowd is. To move to an intel platform is not the issue for Apple, and it never was. Not becoming just part of the crowd when doing so is.
Look at SUN. They made the best machines you could buy for internet applications at a time back in the nineties, and charged you a lot of money for it. Today the rest of the world has caught up, we all stack our racks with linux pizzaboxen now, and SUN is in trouble. The company has to decide: is SUN a hardware company? that would mean investing a lot in the development of SPARC, killing the Solaris x86 line and fighting Linux, or move entirely to Intel, giving up software development altogether and become like Mike. Or is SUN a software company? that would mean cancelling further SPARC development and concentrating on Solaris and Java. Eventually, this would kill SPARC.
Strengthening the hardware section in SUN would hurt the software guys, and beefing up the software department could easily hurt the hardware sales. Not a good strategic position. Apple could easily be (or have been) caught in the same situation. To compete with Dell you have to become like Dell. If you don't want to do this, you must find a different market for yourself. Or be just part of the crowd.
This is more pretty-printing than real innovation. They claim to arrange data by relation but the thing still knows active folders, parent folders and subfolders. And the color scheme (subfolders are blue) focuses on the hierarchical structure of the folders and not the relation of the data. So they took one way of organizing and presenting files that works for most people most of the time but has a few big shortcomings, pretty-printed it in a somewhat confusing way and added relational sugar that can only add to the confusion.
Pretty, but not impressive.
There is a design lesson in here.
In Germany, the toll collect system was supposed to begin September 1st. The implementing consortium (bigshots like DaimlerChrysler and Deutsche Telekom) missed that deadline, so it was pushed back to November 1st, and this new deadline is in doubt, too. The On-Board-Units that every truck has to carry are
- not enough: the consortium initially figured they'd need about 50000 units for the entire country
- defective: as many as half of the delivered units show defects. Some even self-destruct on starting the truck's engine
- expensive: they cost several hundred Euros apiece, and logistics companies who have to buy them are complaining
PR-wise it's a disaster.Last week, manager-magazin.de ran an interview with Peter Newole, executive at Austria's equivalent, Europpass system, which does the same things TollCollect is supposed to do, only that it's cheaper and actually works. (The interview is in German, sorry.) Basically, what Mr. Newole says is that the two systems are doing similar things in vastly different ways. In both countries, trucks have to carry boxes that can communicate with base stations to register their location. Based on these location profiles, the toll is calculated. But the design of these boxes is completely different:
The design lesson is obvious: The more of something you are going to deploy, the simpler it has to be. Put the logic into the servers and make the clients as dumb as you can.
The old sentence gets even more valid with TCPM. We all are used to the fact that binaries are tied to the computers we're using. Buy a new computer to replace your crashed old one, find out you have to use a newer version of the kernel to support Bozo Gadget 2.78, reinstall your binary application from backup, find out that your new system has glibc 3.14 where the old binaries were linked against libc 1.41, yell an expletive, dig for the source, recompile. Been there, done that.
With TCPM, this will only be stricter, not fundamentally different. Use the Source and you'll be fine.
"Ordnungsgeld" will probably end up in the government's pockets.
SCO Germany has to pay a 10000 Euro fine. This decision by the Landgericht Munchen I is based on a preliminary injunction against SCO, granted to the Tarent company and LinuxTag. According to it, SCO may not claim that Linux contains illegally obtained intellectual property from SCO. SCO has allegedly violated this on its homepage, therefore in June Tarent asked for a trial.
According to a Tarent GmbH statement, the court accuses SCO of "negligence" in running its company's homepage, which, even after the preliminary injunction allegedly read: "End users who use the Software Linux may be held liable for violations of SCO's intellectual property."
Tarent attorney Till Jaeger sees the court's decision as confirming that SCO's allegations are "massively business-damaging statements" concerning an "extremely sensitive issue." Unproven claims were used to do fear-induced business on the expense of others. SCO Germany is currently not available for comment. In early June, asked about the trial, SCO Germany CEO Hans Bayer emphasised: "Our intention was to comply." and that the violation of the preliminary incunction had not been intentional.
Maybe they're speculating that the jury will immediately succumb to the magic word 'hash'.
But otherwiese, frankly, i don't see what this could be good for. Hashes (whether MD5 or SHA or some other algorithm) don't prove a thing.
Identity: The identity of the hashes of two MP3s only provey that the MP3s were encoded with identical settings from an identical CD source. If two people, one in NY the other in LA buy the latest Red Hot Chili Peppers album and rip and encode it both on Windows machines using identical versions of RealOne (or any encoder) then the resulting MP3s will have identical hashes. Whether the probability of two different files accidentally having the same hash ist 1 in 2 or 1 in 2^127 is absolutely irrelevant here. The chances of two people using the same software with the same CDDB information to rip the same track from a CD that sold a million copies is a lot higher. Everybody with a half episode of Matlock legal expertise will tear the RIAAs position apart on this ground.
Trackability: Hashes cannot be used to reliably track the path of copies across P2P networks either. Since the hash is more sensitive to minor changes than the ear doing random changes to the ID3 tags or randomly changing a bit or two somewhere in the MP3 will wipe the tracks.
So two files having the same hash doesn't prove they come from a single origin. Two files having different hashes doesn't prove they don't come from a single origin.
Hashes don't prove a thing
This is most likely not a good idea, and not primarily because of the current job market. Chances are high that you will find the same situation elsewhere, too. So moving to a different job (if you can do it at all) will most likely not improve your situation. If you can't manage your customers in your current job, you probably won't manage them anywhere.
There are reasons why customers behave in an insane way sometimes. The most common one is: they think you work for free. Not literally, but in a large organization few people have to pay directly for the work of the IT staff. So to them your work is free as in beer. And if you hand out beer for free, people have an incentive to get drunk. Of course overload will cause headache (as in beer), but the incentive is still there.
The solution is to make them pay for your work. If you can't do this in dollars and cents, make them work. Force them to go through your boss. Make them take the blame if a top banana at your organization can't use powerpoint for a day because you are busy fixing minesweeper installations.
I recently switched from bogofilter to SpamBayes. While it still shows the minor issues a young project always has (incompatibility with the dumbdb in Python 2.2.2 of SuSE 8.2 so you have to use gdbm as the internal db driver etc), i consider it one of the most promising spamfilters around.
Sure, it's only one data point, and next week will be different, but i think i'll stick with SpamBayes for now.
Every expensive product in human history that faced cheaper (free is just the extreme) competition has at one point resorted to insulting their customers by calling them cheapos. "Freeloader Mentality" is a very hollow word that describes the simple fact that people make many of their economic decisions in a surprisingly economic way: As long as major news sites are free (as in beer), people won't pay for yet another one that charges them. It's that simple and ist's called competition.Get used to it.
...and all Gambling is Investment. That's just the way the ruble rolls. What i don't get is why anyone---unless they spent the last five years in a coma---would trust the markets to have any predictive power on an issue like this. Me, i'd bet a thousand that Saddam is either alive or dead, or somewhere in between.
The particular hash they attack stores up to 14 bytes of password data. However, these 14 are split in two chunks of 7, then converted to uppercase then hashed, then concatenated. This is about the stupidest possible way to waste entropy. It means that alphanumeric passwords have effectively at most log_2 (26+10) bits \approx 5.1 bits of entropy per keystroke. Each chunk will have seven times this entropy (seven bytes), and both chunks combined will have an entropy of (7 * lg_2 36) + 1 \approx. 37.1 bits (They state that in their paper).
With a sufficiently high number of key chains that amounts to 1.4 GB of storage and around 13.6 s on a modest 1.5GHz Intel P4 with 512MB RAM. Now, UNIX of course has bigger keys and salt. With salt, you'd need 4096 times the table size and 2^(56-37) = 524288 times the time (this is of course a ballpark estimate, i doubt that the actual calculation scales this easily, especially the memory lookups will suffer heavily from bigger tables). This means with a table size of 5.6 TB you'd be able to crack a DES password (and actually DES itself) in about 82.5 days.
Ballpark estimate or not, considering what it took to build the machine that was used for "Cracking DES" a couple of years ago, this is pretty impressive work.
Wow, these guys just invented the dictionary attack!
If that is the source then the original quote is wrong. The footnote clearly states that in an earlier calculation they had included a worst-case assumption that 80% of the machines would have to run VMware for 4 to 5 years. A commitment from SuSE to guarantee lower costs for this made that assumption obsolete. That's what the report actually says.
linear motors and maglevs are not the same. The latter use the former, that's it. But you can't expect the FT to know this...
I am German, but not a lawyer, but i think this is unconstitutional. Paragraph (1) says clearly that everyone has the right to access information from generally available sources. The internet is such a source. Nazi propaganda, while morally repulsive and disgusting, is information. So article 5 applies to the sites in question.
Paragraph (2) then sets a few exceptions to the rule. One of these is to protect children, one says that insults are not protected by article 5, and then the paragraph (2) says something about paragraph (1) being limited by the general laws. This means that a specific law may limit paragraph (1) in specific cases. It does not mean that there is an unwritten codex of what may be said and what not and that a local authority could decide on an ad hc basis what they like or not (that would be censorship) and then prosecute whoever they want. I believe the law the local authority was thinking about is the one that forbids the dissemination of Nazi Propaganda.
The problem with this point of view is that while distributing Mein Kampf in Germany is illegal, possessing it isn't. Neither is transporting it (otherwise the entire Deutsche Post AG would go to jail). So neither the ISP nor the persons accessing the websites in question are doing anything that would be against the law that forbids distribution of Nazi propaganda. And if they don't, the law doesn't apply and paragraph (1) stands.
The press release in which they blame the ISP for yielding to racists is the usual whining. They made a mistake, they don't want to admit it, so they call everybody a Nazi. The only new thing is that i didn't know that Godwin's law applies to muggels.