We just discussed this in my databases class. The problem is, you can't necessarily roll back a transaction which involves real-world state changes. One of my friends had attempted to get a fairly large amount of money (near his daily limit) from an ATM for a trip he was going on. The ATM crashed, and he didn't get his money, but it was deducted from his account. Furthermore, he couldn't make another withdrawal because of the limit.
Does the admin priviledge include install a bash script in the startup folder?
No. You could install one in your own startup items folder, but then it would be executed by admin again, not by root. I wouldn't be surprised if root can take over the system, but I doubt it's that easy.
To the GP poster: there is actually a root account, but you can't login as root. You can, however, sudo -s, as demonstrated in the following example (I don't run as admin normally):
The reviewer paid $3,000 for the dual G5! A nice top of the line x86 is around $1,200. I built my own x86 for about $700 (not including monitor) and it just feels so much faster then a single G5. Also, to get better performance out of a dual G5 with the latest Mac OS, you need a bunch of memory that further drives up the cost. The reviewer put 4GB, the wimpy 512MB that comes with a $3,000 system is not acceptable.
The reviewer was half-right. The memory that comes with new Macs is simply not enough, and their video cards aren't top-of-the-line either; this is the most consistently irritating thing about Macs that I've seen.
However, his claim about speed simply isn't true. I'm running an eMac 700MHz G4 right now, with 768MB RAM and an external hard drive replacing the internal one. It is by no means a fast computer. It has a 32MB video card. I found it quite slow under MacOS 10.1.
However, my computer doesn't seem slow. Most interface operations don't have noticeable latency even with many windows open; Expose is perfectly smooth with up to 10-15 windows and still looks nice at 25. Collapsing a transparent window into the dock makes no delays. The aggressive caching makes operations with inactive applications fast. The web pages that he claimed took 4 seconds to load on a dual G5 took only 3 seconds on my system, which makes me question his numbers.
There are a few things in the system that are slow. Launching an xterm is slow for some reason. All my other X applications launch almost instantly (except the GIMP of course), but xterm is slow. Displaying a mailbox takes a few hundred microseconds per message, which adds up in boxes with several thousand messages. And of course, raytracing and compile jobs are slow, but I either nice them and go surf the web or walk away and do some reading whether I'm on a Mac or a PC. The compile jobs can also make other programs run slower, even when niced, by flushing pages from the cache.
I've never seen the G5s in the computer lab (nor any of the X86s) take an appreciable amount of time to do anything compute- or memory-bound other than serious batch-jobs such as compiling or raytracing. For this reason, I've come to the belief that with properly designed software, almost any desktop computer built in the past 5 years is plenty fast for everyday tasks.
Hehe. His sig is the quote on the last set of technologies in Alpha Centauri, Transcendent Thought. It's poetic but it quickly becomes annoying at the end of a game when you hear Voice/Planet say it like 10 times.
How about the now century-plus old campaign to convince Americans there are only two choices?
Easy. There are only two choices. The system is locked. You can only express one preference, so there's no point in voting for someone if you're pretty sure they're not going to get elected as long as you have a preference between two (or more) candidates who are doing well in the polls. For various reasons (yay statistics), you generally expect to see only a few candidates doing well enough in the polls that people will vote for them, or that (almost as important) corps will send them money to campaign with.
This is not just a social problem, it's a mathematical one. Basic game theory tells you that in terms of who's going to get elected this November, there's just no point in voting for Nader unless you truly do not give a damn whether Bush or Kerry wins.
Therefore, the solution must be a mathematical one instead of a social one; that is, we should change the election system. We should do this anyway, because the Electoral College is pretty much obsolete.
Condorcet Voting, for instance, does a much better job of allowing multiple parties to compete, for the reason that putting your favorite party (which you believe to be unlikely to win) at the top won't really hurt your favorite realistic candidate's chances. (There is a theorem that this can't be perfectly true in any voting system, but it's pretty close.)
Unfortunately, there is no analog that works this way for election funding. You only have so much money to give. Perhaps some sort of campaign finance reform could bring the system closer to sane, but enforcing mathematical principles on funding doesn't really work very well....
Although I'm sure this will troll... Okay, in the general, how does being able to recompile your kernel help? You lose support once you do that, i know this from experience. Second, having the complete source to the OS only helps a handful of people who can actually truly call themselves systems programmers. For most everything else, which is pretty much the entirety of the computing world, Linux is a mere equal, if not an inferior in some cases.
Well, I can't speak for Linux, but on FreeBSD, here's what it gets you:
First, you can disable modules and kernel features that you won't use. I want Linux compatibility, but others may not, and it takes substantial resources. I don't have any SCSI or Fibre Channel or tape drives, so I can disable those parts of the kernel. This also boosts security and reliability if those parts are buggy.
Second, you can choose debugging options. If you're running -CURRENT, you can choose to live on the edge, without debugging, or make it slow but useful and turn debugging on.
Third, you can set system parameters for your own system. In particular, you can set cache sizes and memory limits for various parts of the kernel. My cheapass server only has 80k cache, so some parameters should be set smaller for performance; in particular -O2 would be in order instead of -O3.
Lastly, you can stay on top of current kernels without the dev team compiling a kernel for every platform at every commit.
While OpenFirmware is cool and all -- as are many of the Apple's fatures, you're still stuck with a propritary CPU (maybe not completely propritary, but seriously, who else uses those?)
IBM. A lot of their Linux blade servers use the PPC 970, aka G5.
I've been running FreeBSD 5-CURRENT. It's been more or less stable, not as much as 4.10, but hopefully it will be getting more so with the establishment of a 5-STABLE branch.
The plusses: 5.x is faster, especially on an SMP or hyperthreading machine. It also supports goodies like ACLs and snapshots.
How about starting with the United States, Israel, England, France, India...
So basically you have Russia, France, India, Pakistan, China, Israel, the UK, and us. With nukes. We've got about 2600 more than anybody else... whatever.
The real assumption here is that radio waves are an efficient means to communicate, and that a race which is capable of communicating over long distances (i.e. one that is capable of being found by us under any circumstances) will know about it. Furthermore, SETI assumes that if they are interested in contacting other races (again a prerequisite for finding them), that they will use radio waves (perhaps in addition to something else) to broadcast this fact.
The only way to broadcast via radio the existence of something that goes beyond simple, natural phenomenon, is to send a signal that other races might see (i.e. on a frequency that they are likely to be listening for) and is so eminently unnatural that it had to be created by an intelligent force. The prime numbers are an example of this, but they are not the only one; any pattern that does not occur in nature (preferably one that can be identified easily) could be used.
But what SETI is looking for is the carrier signal. The first evidence of a signal should be radio waves tuned to a specific frequency; wideband communications do not travel very far. They had to fudge on the frequencies for lack of processor power; small multiples and obvious shifts of the decay frequency of hydrogen is the main set they're listening on, but this is sort of arbitrary. It's an important enough frequency that it has a chance of being used, though.
Right. Let's say we're broadcasting to a target audience of aliens at our level of civilization, say, between mid-20th-century and 10,000 years in the future (any more advanced, and they're not going to even bother... more important things to do).
Maybe there's a limit to technological advance? It doesn't have to be an "after this point, we've discovered everything there is to discover" limit, but maybe after some point, you have reached the limits of useful things that you can do. Eg, you can prove that your computers/spacecraft/bioenhancements can't get more than 1% faster/more powerful, so there's not much point to improving them.
On the other hand, there is a definite possibility that they will destroy themselves. This is the Fermi principle: any intelligent, technological culture will either anihilate itself or ascend to superintelligence.
So we're looking for aliens in a 10,000 year window. The universe is 15 billion years old, give or take. It sounds unlikely to me.
Even if this were the case, the universe is a big place. If one solar system in a million has an Earth-like planet, and one Earth-like planet in a thousand has life, and one in a thousand of those has life with communications technology, and one in a million of those is still around and non-godlike when we radio them... then we would make contact with a thousand other worlds, were time not an obstacle.
You might quibble about these numbers, but there's not much to be said: nobody knows anything about them, as we have only a few data points for the first one, one for the second two, and none for the last one. And that one data point doesn't count for much (anthropic principle).
Most people think that distance is the limiting factor between alien civilizations. I don't - I think it's time.
This is sort of true. Even if we did make contact with another race, if it was several billion light-years away (as is most of the universe), the possibilities of two-way contact would be... pretty much nothing unless one side has some kind of warp drive.
DSA and GPG aren't hashes. They're signature schemes (well, GPG implements PGP, an encoding of either RSA or ElGamal signatures), and they use other hashes in their checksums, like SHA1 and MD5. Of course, you can make GPG do RIPEMD160, which IIRC has not been broken yet either.
What is this supposed to prove? Java is strongly typed, but not dynamically typed. The only reason your code does not require a cast is because you are using dynamic typing and "figuring out" the type at runtime. Big deal.
No. OCaml is strongly, statically typed and its type-safety is figured out at compile time. Of course, if you're running in the interpreter, compile time is directly before runtime, but if I had tried to compile the code (either to bytecode or to machine code) instead of run it in the interpreter, it would have given me the same error (with a line number).
ArrayList l = new ArrayList(); l.append("A String"); String s = (String)l.item(0);
Is an upcast. and I dare you to find a list implementation in any type-strong language that doesn't require an upcast in this situation. You need it to be able to store objects of an anonymous type on a list.
Yes, but typestrong languages don't like lists of objects of anonymous types.
For example, in OCaml:
# let l = []
let l = "A string":: l
let s = List.hd l ;; val l : string list = ["A String"] val s : string = "A String"
Note that I didn't specify any types there (I could have, but I chose not to), and the compiler figured it out. Of course, I didn't use an ArrayList, or even any real objects. Let's try that again with objects, supposing I have a class ['a] arrayList.
# let l = new arrayList
let _ = l#append "A String"
let s = l#item 0 ;; val l : string testlist = <obj> val s : string = "A String"
What if I'd goofed and tried to use it as an integer?
# let l = new arrayList
let _ = l#append "A String"
let s = 5 + l#item 0 ;; This expression has type string but is used here with type int
(And it underlines the 5+l#item 0 in the console.)
I'm using CRM114. It may not get 99.98% of my spam, but it gets at least 95% (and I haven't trained it all summer), and hasn't misclassifed a good message in a long time. I've done one add-on, though, a mod that helps it find certain forms of dictionary salad.
When used properly, One Time Pad is impossible to break.
Not quite true - a one time pad cipher where the key is as long, or longer than the message is impossible to break.
That would be using it improperly.
The phone line between the Kremlin and the White house is enciphered this way.
According to the Crypto Museum, during the cold war, there was a teletype (not a phone line) between the Kremlin and the White House, and it was enciphered, but not with a one-time pad.
"Today's PC has on average 64 megabytes of cache and 20 to 60 gigabyte hard drives. "
64 MEGABYTES? Who has 64 megabytes of cache?
Cache here doesnt't mean the processor's instruction or data caches, but rather main memory. In many server applications and some desktop ones, main memory is primarily used as a cache for the disk.
Slashdot Mirror
We just discussed this in my databases class. The problem is, you can't necessarily roll back a transaction which involves real-world state changes. One of my friends had attempted to get a fairly large amount of money (near his daily limit) from an ATM for a trip he was going on. The ATM crashed, and he didn't get his money, but it was deducted from his account. Furthermore, he couldn't make another withdrawal because of the limit.
So ATM databases aren't perfect.
That's no moon...it's a space station!
For everything else, there's massacres.
I live in NYC too and I agree with the OP; iPods are a dime a dozen around here.
Wow. I've got to get into New York. I can't find an iPod for under $100 anywhere around here...
No. You could install one in your own startup items folder, but then it would be executed by admin again, not by root. I wouldn't be surprised if root can take over the system, but I doubt it's that easy.
To the GP poster: there is actually a root account, but you can't login as root. You can, however, sudo -s, as demonstrated in the following example (I don't run as admin normally):
The reviewer paid $3,000 for the dual G5! A nice top of the line x86 is around $1,200. I built my own x86 for about $700 (not including monitor) and it just feels so much faster then a single G5. Also, to get better performance out of a dual G5 with the latest Mac OS, you need a bunch of memory that further drives up the cost. The reviewer put 4GB, the wimpy 512MB that comes with a $3,000 system is not acceptable.
The reviewer was half-right. The memory that comes with new Macs is simply not enough, and their video cards aren't top-of-the-line either; this is the most consistently irritating thing about Macs that I've seen.
However, his claim about speed simply isn't true. I'm running an eMac 700MHz G4 right now, with 768MB RAM and an external hard drive replacing the internal one. It is by no means a fast computer. It has a 32MB video card. I found it quite slow under MacOS 10.1.
However, my computer doesn't seem slow. Most interface operations don't have noticeable latency even with many windows open; Expose is perfectly smooth with up to 10-15 windows and still looks nice at 25. Collapsing a transparent window into the dock makes no delays. The aggressive caching makes operations with inactive applications fast. The web pages that he claimed took 4 seconds to load on a dual G5 took only 3 seconds on my system, which makes me question his numbers.
There are a few things in the system that are slow. Launching an xterm is slow for some reason. All my other X applications launch almost instantly (except the GIMP of course), but xterm is slow. Displaying a mailbox takes a few hundred microseconds per message, which adds up in boxes with several thousand messages. And of course, raytracing and compile jobs are slow, but I either nice them and go surf the web or walk away and do some reading whether I'm on a Mac or a PC. The compile jobs can also make other programs run slower, even when niced, by flushing pages from the cache.
I've never seen the G5s in the computer lab (nor any of the X86s) take an appreciable amount of time to do anything compute- or memory-bound other than serious batch-jobs such as compiling or raytracing. For this reason, I've come to the belief that with properly designed software, almost any desktop computer built in the past 5 years is plenty fast for everyday tasks.
The latest iMac is upgradable. Parts you can upgrade yourself include the RAM, HDD, optical drive, PSU, CPU/mobo, LCD, wireless card, and modem.
Grr, my can of coke is half empty. >:I
Still tasty though.
Hehe. His sig is the quote on the last set of technologies in Alpha Centauri, Transcendent Thought. It's poetic but it quickly becomes annoying at the end of a game when you hear Voice/Planet say it like 10 times.
How about the now century-plus old campaign to convince Americans there are only two choices?
Easy. There are only two choices. The system is locked. You can only express one preference, so there's no point in voting for someone if you're pretty sure they're not going to get elected as long as you have a preference between two (or more) candidates who are doing well in the polls. For various reasons (yay statistics), you generally expect to see only a few candidates doing well enough in the polls that people will vote for them, or that (almost as important) corps will send them money to campaign with.
This is not just a social problem, it's a mathematical one. Basic game theory tells you that in terms of who's going to get elected this November, there's just no point in voting for Nader unless you truly do not give a damn whether Bush or Kerry wins.
Therefore, the solution must be a mathematical one instead of a social one; that is, we should change the election system. We should do this anyway, because the Electoral College is pretty much obsolete.
Condorcet Voting, for instance, does a much better job of allowing multiple parties to compete, for the reason that putting your favorite party (which you believe to be unlikely to win) at the top won't really hurt your favorite realistic candidate's chances. (There is a theorem that this can't be perfectly true in any voting system, but it's pretty close.)
Unfortunately, there is no analog that works this way for election funding. You only have so much money to give. Perhaps some sort of campaign finance reform could bring the system closer to sane, but enforcing mathematical principles on funding doesn't really work very well....
You hit the nail right on the head. This is why I use a Mac for my main box (most of the time) and FreeBSD on a cheap homemade box for my fileserver.
Although I'm sure this will troll... Okay, in the general, how does being able to recompile your kernel help? You lose support once you do that, i know this from experience. Second, having the complete source to the OS only helps a handful of people who can actually truly call themselves systems programmers. For most everything else, which is pretty much the entirety of the computing world, Linux is a mere equal, if not an inferior in some cases.
Well, I can't speak for Linux, but on FreeBSD, here's what it gets you:
First, you can disable modules and kernel features that you won't use. I want Linux compatibility, but others may not, and it takes substantial resources. I don't have any SCSI or Fibre Channel or tape drives, so I can disable those parts of the kernel. This also boosts security and reliability if those parts are buggy.
Second, you can choose debugging options. If you're running -CURRENT, you can choose to live on the edge, without debugging, or make it slow but useful and turn debugging on.
Third, you can set system parameters for your own system. In particular, you can set cache sizes and memory limits for various parts of the kernel. My cheapass server only has 80k cache, so some parameters should be set smaller for performance; in particular -O2 would be in order instead of -O3.
Lastly, you can stay on top of current kernels without the dev team compiling a kernel for every platform at every commit.
Most phones sold in the US have positioning capability, be it GPS or cell-tower, for use in the E911 system.
While OpenFirmware is cool and all -- as are many of the Apple's fatures, you're still stuck with a propritary CPU (maybe not completely propritary, but seriously, who else uses those?)
IBM. A lot of their Linux blade servers use the PPC 970, aka G5.
I've been running FreeBSD 5-CURRENT. It's been more or less stable, not as much as 4.10, but hopefully it will be getting more so with the establishment of a 5-STABLE branch.
The plusses: 5.x is faster, especially on an SMP or hyperthreading machine. It also supports goodies like ACLs and snapshots.
Try the Handbook for Linux compatibility mode.
--Mike
How about starting with the United States, Israel, England, France, India...
So basically you have Russia, France, India, Pakistan, China, Israel, the UK, and us. With nukes. We've got about 2600 more than anybody else... whatever.
That's a brilliant troll. I'll byte anyway.
The real assumption here is that radio waves are an efficient means to communicate, and that a race which is capable of communicating over long distances (i.e. one that is capable of being found by us under any circumstances) will know about it. Furthermore, SETI assumes that if they are interested in contacting other races (again a prerequisite for finding them), that they will use radio waves (perhaps in addition to something else) to broadcast this fact.
The only way to broadcast via radio the existence of something that goes beyond simple, natural phenomenon, is to send a signal that other races might see (i.e. on a frequency that they are likely to be listening for) and is so eminently unnatural that it had to be created by an intelligent force. The prime numbers are an example of this, but they are not the only one; any pattern that does not occur in nature (preferably one that can be identified easily) could be used.
But what SETI is looking for is the carrier signal. The first evidence of a signal should be radio waves tuned to a specific frequency; wideband communications do not travel very far. They had to fudge on the frequencies for lack of processor power; small multiples and obvious shifts of the decay frequency of hydrogen is the main set they're listening on, but this is sort of arbitrary. It's an important enough frequency that it has a chance of being used, though.
Right. Let's say we're broadcasting to a target audience of aliens at our level of civilization, say, between mid-20th-century and 10,000 years in the future (any more advanced, and they're not going to even bother... more important things to do).
Maybe there's a limit to technological advance? It doesn't have to be an "after this point, we've discovered everything there is to discover" limit, but maybe after some point, you have reached the limits of useful things that you can do. Eg, you can prove that your computers/spacecraft/bioenhancements can't get more than 1% faster/more powerful, so there's not much point to improving them.
On the other hand, there is a definite possibility that they will destroy themselves. This is the Fermi principle: any intelligent, technological culture will either anihilate itself or ascend to superintelligence.
So we're looking for aliens in a 10,000 year window. The universe is 15 billion years old, give or take. It sounds unlikely to me.
Even if this were the case, the universe is a big place. If one solar system in a million has an Earth-like planet, and one Earth-like planet in a thousand has life, and one in a thousand of those has life with communications technology, and one in a million of those is still around and non-godlike when we radio them... then we would make contact with a thousand other worlds, were time not an obstacle.
You might quibble about these numbers, but there's not much to be said: nobody knows anything about them, as we have only a few data points for the first one, one for the second two, and none for the last one. And that one data point doesn't count for much (anthropic principle).
Most people think that distance is the limiting factor between alien civilizations. I don't - I think it's time.
This is sort of true. Even if we did make contact with another race, if it was several billion light-years away (as is most of the universe), the possibilities of two-way contact would be... pretty much nothing unless one side has some kind of warp drive.
DSA and GPG aren't hashes. They're signature schemes (well, GPG implements PGP, an encoding of either RSA or ElGamal signatures), and they use other hashes in their checksums, like SHA1 and MD5. Of course, you can make GPG do RIPEMD160, which IIRC has not been broken yet either.
I think your contest should be for "most screwed up query with innocent intentions". And must contain only true stories. Should be interesting...
Looking for information on how to put diagrams into a TeX document:
LaTeX image
What is this supposed to prove? Java is strongly typed, but not dynamically typed. The only reason your code does not require a cast is because you are using dynamic typing and "figuring out" the type at runtime. Big deal.
No. OCaml is strongly, statically typed and its type-safety is figured out at compile time. Of course, if you're running in the interpreter, compile time is directly before runtime, but if I had tried to compile the code (either to bytecode or to machine code) instead of run it in the interpreter, it would have given me the same error (with a line number).
Yes, but typestrong languages don't like lists of objects of anonymous types.
For example, in OCaml:Note that I didn't specify any types there (I could have, but I chose not to), and the compiler figured it out. Of course, I didn't use an ArrayList, or even any real objects. Let's try that again with objects, supposing I have a class ['a] arrayList.What if I'd goofed and tried to use it as an integer?(And it underlines the 5+l#item 0 in the console.)
I'm using CRM114. It may not get 99.98% of my spam, but it gets at least 95% (and I haven't trained it all summer), and hasn't misclassifed a good message in a long time. I've done one add-on, though, a mod that helps it find certain forms of dictionary salad.
The sh*t, perhaps not, but still darn good.
Not quite true - a one time pad cipher where the key is as long, or longer than the message is impossible to break.
That would be using it improperly.
The phone line between the Kremlin and the White house is enciphered this way.
According to the Crypto Museum, during the cold war, there was a teletype (not a phone line) between the Kremlin and the White House, and it was enciphered, but not with a one-time pad.
Excerpted:
"Today's PC has on average 64 megabytes of cache and 20 to 60 gigabyte hard drives. "
64 MEGABYTES? Who has 64 megabytes of cache?
Cache here doesnt't mean the processor's instruction or data caches, but rather main memory. In many server applications and some desktop ones, main memory is primarily used as a cache for the disk.