Actually, my list of 10 would be: IIS, IE, Outlook, Apache, random UNIX services you don't mean to have, sendmail, bind, sshd, plaintext passwords, ftpd. There's only 3 Windows ones, because you can probably break into more Windows boxes than you want to with just those 3. None of the other Windows vulnerabilities matter much, because you can't use them to break into UNIX systems.
Processors do this to the extent that it's possible at runtime; that's what out-of-order execution is, basically. The problem is that it only makes your single threaded program into 2 or 3 threads; beyond that, you need to look at bigger chunks of the program than the processor ever sees at once.
Beyond that, you really need to be able to look at the program as a whole in order to do anything that clever, so you're talking language, compiler, or library features, and you generally have to involve the programmer somewhat, although you don't necessarily have to do it as explicit threads. (E.g., there's a C variant with a keyword that says it's okay to evaluate all of the arguments to a function at the same time)
If you've got a congressional election this fall (most people do), you might want to contact each of the candidates. If you can make whoever wins think they might not have if they hadn't approved of this act, that has a much stronger effect than if you merely contacted them once they were in office.
It doesn't take very long to compile these days, and it takes as long as it ever did to read the documentation, so either way, it'll be ready for you to use before you're ready to use it.
I think the real advantage is being able to get the same package regardless of your library versions. So long as they're source compatible (and library providers generally try to make sure of that), the same source package will install on a huge number of systems which would require different binaries (if it links against 10 shared libraries, and there are 2 binary-but-not-source-incompatible versions of each, that's 1024 different binaries you'd need).
Windows is the industry standard for a good user interface by MicroSoft fiat, not by actually being usable. So I can understand developers wanting to break away from it (although they haven't historically put enough thought into the interface they end up using).
On the other hand, regardless of the reason, Windows is the industry standard for a good user interface. But what matters really is the function; it would be better, in my opinion, to have the ability to use a UI that doesn't look much like Windows, but has the same functionality. Icons which are different drawings of the same things have the same uses, there are different colors and patterns, and so forth, but following the same instructions will give you the same result whenever possible.
Of course, the next step is to be able to switch each part individually away from the Windows feel, because then people can learn each day to do one more thing efficiently.
Either that, or use a totally different user interface based on a principled design and promise the users that, if they use your software instead of MS software, they'll never have to learn another UI again after this one.
In any case, there's no advantage to superficial similarity to Windows these days; it's not like you have to hide the fact that you're using Linux from your boss, and it will only confuse people who have to use two operating systems that work differently if there aren't visual cues as to which one something is.
Actually, business travellers to Boston will not see any benefit from the Big Dig (except for not having an ugly highway in the way). Getting between the airport, hotels, and convention centers doesn't use the highway at all; you'd either take the subway or a taxi on other roads. While Boston traffic is kind of a mess, the traffic that the Big Dig is supposed to relieve is primarily people driving past downtown Boston getting squeezed onto a narrow highway winding among the office buildings, going between I-90 to the west, I-93 to the north and south, and the airport to the east.
Actually, the economy of scale in printing a book run isn't that great, and it's falling as print-on-demand gets better, driven by people going to copy centers to print a dozen of the thing they're holding.
Doing large print runs is inefficient in storage and because you have to move the books to the place they're sold. That's starting to matter more than the actual printing.
The main thing is that Linux, HURD, and the BSDs have different fundamental designs. Nobody knows which design is the best, or even if there is a single best design (maybe each of them is better for a different sort of load). It would be a problem if everybody used a design that turned out not to be very good (imagine if, in order to make Linux run well on the hardware people will have in two years, it had to be turned entirely into HURD).
Furthermore, there's no reason to think that Linux development would go better with more people. Linux is on the border of being a large project with all of the management problems that causes. Linux development is actually helped most, probably, but the existence of separate projects which try different approaches, so that Linux developers can then look at the results and consider different techniques.
Actually, I gave my mother an account on my machine, which she uses with ttssh (or mindterm, now that I've set that up). It was much easier to explain Pine and a little bit of shell than to explain anything graphical (If you forget what the keys are, it tells you on the screen in small words; no mysterious pictures that could be anything). She has a complete explanation written on a small piece of paper next to the computer, too.
Grandma: This is great! My hands shake more than the size of anything I have to click on if I use the mouse, but I've been able to type since before your parents were born.
Just tell them that JBoss actually implements the specs correctly and doesn't take forever to compile your application. We switched to JBoss and the time it took to make the switch was probably recovered in the next two weeks in the shorter build cycle.
The main thing you have to deal with in going from WebLogic to JBoss is that WL supports a number of intermediate and non-strict versions of EJB, and JBoss is a pretty strict implementation of the final versions. Also, there are a number of mistakes that JBoss doesn't give useful errors about, which makes switching annoying.
The weird thing is that this whole thing really sounds like MS is trying to catch up to Linux. It sounds like Ballmer is saying, "We currently suck at everything that matters, but eventually we'll be as good as Linux." But MS has previously said they're ahead, and if they've fallen behind, there's not reason to think they won't just fall farther behind. I'd expect MS to want to focus on places they're ahead or about to get ahead. These statements would make sense if they were about to release a clustering solution better than a single old Linux box, or if they were about to give customers something more useful than the source, or some system for forming a support community easily, or something. I have to assume zdnet selected quotes to make it sound this way, or maybe Ballmer hasn't been getting enough sleep, but this article really butchers MS.
The examples are all trivial to replace with the non-shortcut version except, perhaps, "2". It's just a shorthand. Of course, for "formal writing", you shouldn't turn in something with shorthand, but that doesn't mean you shouldn't write it that way. It's not really different from underlining titles (titles should be italicized; underlining is a handwriting/typewriter shorthand for italics) or writing in cursive.
It's not like they're using inconsistant spelling and abbreviating things all over the place like, say, Shakespear.
There actually have been a number of episodes which refer to previous episodes; mostly of the sort where they do something in one episode, and then encounter people who have heard about it and want them to do something else because of it.
I think they've done a good job of maintaining the Trek "everything is fixed at the end of the episode", while having people notice that this group has an impressive record of getting into situations and resolving them.
I'm not entirely convinced that Daniels's side is actually good; it seems like all of their information has come from Daniels, and he might just be feeding them propaganda. Sullick is clearly against them, but might have reasons for it. There's a lot of potential for a major plot twist at some point.
There's something different going on with the time travel on Enterprise. First, we have people "from the future", who could be just well-informed people from the present. Then, we have the captain going "to the past", but he doesn't seem to have changed anything in the present by doing that; he finds out a bunch of stuff about the present there. Finally, we have the captain going "to the future", but it's indistinguishable from any other destroyed city. When the captain went to the past and spent some time there, he returned to the present later than he left.
Most likely, there isn't time travel going on, just shape-changing aliens recording events and playing them back in a holodeck-like thing. They could just have a clever sort of teleporter, a holodeck, and be using time travel as their excuse for having a lot of information that they don't want to admit to collecting.
If you don't know, the installer had better be suggesting something appropriate, or you're not using a good distribution.
There's no reason to switch from (ext3?) to XFS. But it's quite possible that the next time you install, if you're formatting a new drive, it will suggest XFS. Of course, converting an existing disk is enough of a pain that you probably don't want to do it.
Halloween is the deadline for development. After that, there will be a while tracking down bugs, and then probably a release in January or thereabouts. Linus is planning on turning things over to someone who's a better release manager (Marcello, I think), which means that the release is likely not to drag on, and likely to actually be stable when it happens.
I suspect that a number of distributions will include 2.6 pretty quickly this time, because it'll be handled by someone who is good at stability. Also, the distribution makers are actually pretty close to the 'official' process, and they're really in the best position to judge stability on a wide variety of systems. By the time 2.6 is declared stable, most of the distribution makers will be comfortable with it, both in the official version and with their patches.
If you consider that the Pharohs basically considered life preparation for the afterlife, building the tombs was for them about what building a military base is for us. It's extremely important that everything get done right, and the risk of sabotage is huge. You don't want anyone there who doesn't want to be there. You don't give security clearances to slaves.
That's how you break the encryption on a document. People will generally choose the strength of the encryption (that is, how long it will take to break it by the best known method) so that, by the time you've done this, they don't care any more.
Breaking an encryption method is a different thing. This is done by analyzing the method in order to try to find a better method for breaking encrypted documents than the best method previously known. If this is sucessful, it means that all of the encryption that has been done with that method so far is weaker than had been thought. Of course, the initial method is just trying all possible keys, and it may actually be somewhat foolish to think that there is any cryptosystem which doesn't have a better method; that would mean that all of the key contributes to strength, rather than any of it being weak but necessary to get the algorithm to work. And, from the perspective of what you should use, even the strongest attack so far (if it even works) on AES is harder than brute force on triple DES.
If they can pull this off, it will be an impressive success for the open source model. For a set of contractors to go from getting the problem description to a complete implementation in 3 1/2 months, due to the existence of a good set of tools, would really show the strength of the model.
Getting a custom installation generally takes far longer than that. If this project works, it will start to look reasonable for companies who are planning to get a proprietary solution to get an open source one at the same time to use until the proprietary one is ready.
Re:hammer mp: consequences for kernel arch?
on
AMD Delays Hammer
·
· Score: 2
Running times have never been particularly reproducable, and they're only going to get less so. Pretty soon, we'll have processors whose clock speed is dependant continuously on temperature; they run as fast as they can do accurately without damaging the chip. Of course, that means that your computer is faster during the winter.
As far as existing chips go, hyperthreading also messes a lot with running times, because when you get processor time depends on when another process has cache misses.
I thought the US put all the KKK members on police forces in metropolitan areas. That certainly counts as making them feel more secure...
But seriously, I think it was largely that there stopped being new civil rights for minority races. If the laws have been the same for a while, and nothing really bad has happened to you because of them, you're not really going to be that desperate. Not the actual KKK members, but the potential KKK members, people who would commit terrorism if they thought the people they cared about were about to be driven into poverty.
I think that the US's response to the KKK reduced the damage greatly, and was very important for that reason, but that its momentum was largely broken by economic conditions reaching the post-civil-rights-movement equilibrium, such that the future wasn't frighteningly different from the present, and the present turned out to be survivable. At that point, leading a regular life feels like a good alternative to blowing things up.
People only engage in terrorism if they feel they have no alternative. They don't do it for fun, or even because they think it's morally right. They may do it because they feel a group (perhaps a religious or political group) they belong to has no alternative; they may act as part of that group even though they could personally do something else.
In order to stop terrorism, we must make every potential terrorist feel secure. We must make them all feel that they will not be conquered and ruled by people with no real concern for their needs. We must make them all feel that they will be able to avoid starvation, and that their families and their countrymen will be able to live decent lives. We must make them feel that they can do better by the people they care about by productive work than by blowing other people up. We must make them no longer feel the need for retribution.
We must eliminate all terrorists in order to have security. Terrorism is just too easy for even a few people, regardless of anything we may do to make it harder. Killing all of the terrorists is too much; many more people who wouldn't have been terrorists would then take up the cause. We must, therefore, reform the terrorists. And in order to do that, we must give each and every one of them, and their governments, religious groups, and so forth, a way to exist without desperation.
And Apache will lose market share to whom exactly? If the reason that nobody wants to switch away from Apache 1.3 is that the most important feature of a webserver these days is Apache 1.3 compatibility, nobody's going to switch to anything else.
MicroSoft would be in trouble if nobody switched to new versions of their software from the old version, because they wouldn't make any money. But Apache doesn't care, because they aren't trying to sell new copies. If a corporation were in Apache's position, they would risk running out of money and having their existing software go unsupported. But Apache doesn't need to be supported by new sales, so having people stick with the old version is just fine.
Slashdot Mirror
Actually, my list of 10 would be: IIS, IE, Outlook, Apache, random UNIX services you don't mean to have, sendmail, bind, sshd, plaintext passwords, ftpd. There's only 3 Windows ones, because you can probably break into more Windows boxes than you want to with just those 3. None of the other Windows vulnerabilities matter much, because you can't use them to break into UNIX systems.
Processors do this to the extent that it's possible at runtime; that's what out-of-order execution is, basically. The problem is that it only makes your single threaded program into 2 or 3 threads; beyond that, you need to look at bigger chunks of the program than the processor ever sees at once.
Beyond that, you really need to be able to look at the program as a whole in order to do anything that clever, so you're talking language, compiler, or library features, and you generally have to involve the programmer somewhat, although you don't necessarily have to do it as explicit threads. (E.g., there's a C variant with a keyword that says it's okay to evaluate all of the arguments to a function at the same time)
If you've got a congressional election this fall (most people do), you might want to contact each of the candidates. If you can make whoever wins think they might not have if they hadn't approved of this act, that has a much stronger effect than if you merely contacted them once they were in office.
It doesn't take very long to compile these days, and it takes as long as it ever did to read the documentation, so either way, it'll be ready for you to use before you're ready to use it.
I think the real advantage is being able to get the same package regardless of your library versions. So long as they're source compatible (and library providers generally try to make sure of that), the same source package will install on a huge number of systems which would require different binaries (if it links against 10 shared libraries, and there are 2 binary-but-not-source-incompatible versions of each, that's 1024 different binaries you'd need).
Windows is the industry standard for a good user interface by MicroSoft fiat, not by actually being usable. So I can understand developers wanting to break away from it (although they haven't historically put enough thought into the interface they end up using).
On the other hand, regardless of the reason, Windows is the industry standard for a good user interface. But what matters really is the function; it would be better, in my opinion, to have the ability to use a UI that doesn't look much like Windows, but has the same functionality. Icons which are different drawings of the same things have the same uses, there are different colors and patterns, and so forth, but following the same instructions will give you the same result whenever possible.
Of course, the next step is to be able to switch each part individually away from the Windows feel, because then people can learn each day to do one more thing efficiently.
Either that, or use a totally different user interface based on a principled design and promise the users that, if they use your software instead of MS software, they'll never have to learn another UI again after this one.
In any case, there's no advantage to superficial similarity to Windows these days; it's not like you have to hide the fact that you're using Linux from your boss, and it will only confuse people who have to use two operating systems that work differently if there aren't visual cues as to which one something is.
The weather is always like that; they predict 50-60, and you only get 22. And the people who have to shovel it still swear about it.
Actually, business travellers to Boston will not see any benefit from the Big Dig (except for not having an ugly highway in the way). Getting between the airport, hotels, and convention centers doesn't use the highway at all; you'd either take the subway or a taxi on other roads. While Boston traffic is kind of a mess, the traffic that the Big Dig is supposed to relieve is primarily people driving past downtown Boston getting squeezed onto a narrow highway winding among the office buildings, going between I-90 to the west, I-93 to the north and south, and the airport to the east.
Actually, the economy of scale in printing a book run isn't that great, and it's falling as print-on-demand gets better, driven by people going to copy centers to print a dozen of the thing they're holding.
Doing large print runs is inefficient in storage and because you have to move the books to the place they're sold. That's starting to matter more than the actual printing.
The main thing is that Linux, HURD, and the BSDs have different fundamental designs. Nobody knows which design is the best, or even if there is a single best design (maybe each of them is better for a different sort of load). It would be a problem if everybody used a design that turned out not to be very good (imagine if, in order to make Linux run well on the hardware people will have in two years, it had to be turned entirely into HURD).
Furthermore, there's no reason to think that Linux development would go better with more people. Linux is on the border of being a large project with all of the management problems that causes. Linux development is actually helped most, probably, but the existence of separate projects which try different approaches, so that Linux developers can then look at the results and consider different techniques.
Actually, I gave my mother an account on my machine, which she uses with ttssh (or mindterm, now that I've set that up). It was much easier to explain Pine and a little bit of shell than to explain anything graphical (If you forget what the keys are, it tells you on the screen in small words; no mysterious pictures that could be anything). She has a complete explanation written on a small piece of paper next to the computer, too.
Grandma: This is great! My hands shake more than the size of anything I have to click on if I use the mouse, but I've been able to type since before your parents were born.
Just tell them that JBoss actually implements the specs correctly and doesn't take forever to compile your application. We switched to JBoss and the time it took to make the switch was probably recovered in the next two weeks in the shorter build cycle.
The main thing you have to deal with in going from WebLogic to JBoss is that WL supports a number of intermediate and non-strict versions of EJB, and JBoss is a pretty strict implementation of the final versions. Also, there are a number of mistakes that JBoss doesn't give useful errors about, which makes switching annoying.
The weird thing is that this whole thing really sounds like MS is trying to catch up to Linux. It sounds like Ballmer is saying, "We currently suck at everything that matters, but eventually we'll be as good as Linux." But MS has previously said they're ahead, and if they've fallen behind, there's not reason to think they won't just fall farther behind. I'd expect MS to want to focus on places they're ahead or about to get ahead. These statements would make sense if they were about to release a clustering solution better than a single old Linux box, or if they were about to give customers something more useful than the source, or some system for forming a support community easily, or something. I have to assume zdnet selected quotes to make it sound this way, or maybe Ballmer hasn't been getting enough sleep, but this article really butchers MS.
The examples are all trivial to replace with the non-shortcut version except, perhaps, "2". It's just a shorthand. Of course, for "formal writing", you shouldn't turn in something with shorthand, but that doesn't mean you shouldn't write it that way. It's not really different from underlining titles (titles should be italicized; underlining is a handwriting/typewriter shorthand for italics) or writing in cursive.
It's not like they're using inconsistant spelling and abbreviating things all over the place like, say, Shakespear.
There actually have been a number of episodes which refer to previous episodes; mostly of the sort where they do something in one episode, and then encounter people who have heard about it and want them to do something else because of it.
I think they've done a good job of maintaining the Trek "everything is fixed at the end of the episode", while having people notice that this group has an impressive record of getting into situations and resolving them.
I'm not entirely convinced that Daniels's side is actually good; it seems like all of their information has come from Daniels, and he might just be feeding them propaganda. Sullick is clearly against them, but might have reasons for it. There's a lot of potential for a major plot twist at some point.
There's something different going on with the time travel on Enterprise. First, we have people "from the future", who could be just well-informed people from the present. Then, we have the captain going "to the past", but he doesn't seem to have changed anything in the present by doing that; he finds out a bunch of stuff about the present there. Finally, we have the captain going "to the future", but it's indistinguishable from any other destroyed city. When the captain went to the past and spent some time there, he returned to the present later than he left.
Most likely, there isn't time travel going on, just shape-changing aliens recording events and playing them back in a holodeck-like thing. They could just have a clever sort of teleporter, a holodeck, and be using time travel as their excuse for having a lot of information that they don't want to admit to collecting.
If you don't know, the installer had better be suggesting something appropriate, or you're not using a good distribution.
There's no reason to switch from (ext3?) to XFS. But it's quite possible that the next time you install, if you're formatting a new drive, it will suggest XFS. Of course, converting an existing disk is enough of a pain that you probably don't want to do it.
Halloween is the deadline for development. After that, there will be a while tracking down bugs, and then probably a release in January or thereabouts. Linus is planning on turning things over to someone who's a better release manager (Marcello, I think), which means that the release is likely not to drag on, and likely to actually be stable when it happens.
I suspect that a number of distributions will include 2.6 pretty quickly this time, because it'll be handled by someone who is good at stability. Also, the distribution makers are actually pretty close to the 'official' process, and they're really in the best position to judge stability on a wide variety of systems. By the time 2.6 is declared stable, most of the distribution makers will be comfortable with it, both in the official version and with their patches.
If you consider that the Pharohs basically considered life preparation for the afterlife, building the tombs was for them about what building a military base is for us. It's extremely important that everything get done right, and the risk of sabotage is huge. You don't want anyone there who doesn't want to be there. You don't give security clearances to slaves.
Great, now I'm going to be playing a computer game with characters who have faster computers than I do.
Hmm, maybe that's why people aren't buying high-end chips. They can just have their Sims buy them instead...
That's how you break the encryption on a document. People will generally choose the strength of the encryption (that is, how long it will take to break it by the best known method) so that, by the time you've done this, they don't care any more.
Breaking an encryption method is a different thing. This is done by analyzing the method in order to try to find a better method for breaking encrypted documents than the best method previously known. If this is sucessful, it means that all of the encryption that has been done with that method so far is weaker than had been thought. Of course, the initial method is just trying all possible keys, and it may actually be somewhat foolish to think that there is any cryptosystem which doesn't have a better method; that would mean that all of the key contributes to strength, rather than any of it being weak but necessary to get the algorithm to work. And, from the perspective of what you should use, even the strongest attack so far (if it even works) on AES is harder than brute force on triple DES.
If they can pull this off, it will be an impressive success for the open source model. For a set of contractors to go from getting the problem description to a complete implementation in 3 1/2 months, due to the existence of a good set of tools, would really show the strength of the model.
Getting a custom installation generally takes far longer than that. If this project works, it will start to look reasonable for companies who are planning to get a proprietary solution to get an open source one at the same time to use until the proprietary one is ready.
Running times have never been particularly reproducable, and they're only going to get less so. Pretty soon, we'll have processors whose clock speed is dependant continuously on temperature; they run as fast as they can do accurately without damaging the chip. Of course, that means that your computer is faster during the winter.
As far as existing chips go, hyperthreading also messes a lot with running times, because when you get processor time depends on when another process has cache misses.
I thought the US put all the KKK members on police forces in metropolitan areas. That certainly counts as making them feel more secure...
But seriously, I think it was largely that there stopped being new civil rights for minority races. If the laws have been the same for a while, and nothing really bad has happened to you because of them, you're not really going to be that desperate. Not the actual KKK members, but the potential KKK members, people who would commit terrorism if they thought the people they cared about were about to be driven into poverty.
I think that the US's response to the KKK reduced the damage greatly, and was very important for that reason, but that its momentum was largely broken by economic conditions reaching the post-civil-rights-movement equilibrium, such that the future wasn't frighteningly different from the present, and the present turned out to be survivable. At that point, leading a regular life feels like a good alternative to blowing things up.
Time to channel Desmond Tutu again...
People only engage in terrorism if they feel they have no alternative. They don't do it for fun, or even because they think it's morally right. They may do it because they feel a group (perhaps a religious or political group) they belong to has no alternative; they may act as part of that group even though they could personally do something else.
In order to stop terrorism, we must make every potential terrorist feel secure. We must make them all feel that they will not be conquered and ruled by people with no real concern for their needs. We must make them all feel that they will be able to avoid starvation, and that their families and their countrymen will be able to live decent lives. We must make them feel that they can do better by the people they care about by productive work than by blowing other people up. We must make them no longer feel the need for retribution.
We must eliminate all terrorists in order to have security. Terrorism is just too easy for even a few people, regardless of anything we may do to make it harder. Killing all of the terrorists is too much; many more people who wouldn't have been terrorists would then take up the cause. We must, therefore, reform the terrorists. And in order to do that, we must give each and every one of them, and their governments, religious groups, and so forth, a way to exist without desperation.
And Apache will lose market share to whom exactly? If the reason that nobody wants to switch away from Apache 1.3 is that the most important feature of a webserver these days is Apache 1.3 compatibility, nobody's going to switch to anything else.
MicroSoft would be in trouble if nobody switched to new versions of their software from the old version, because they wouldn't make any money. But Apache doesn't care, because they aren't trying to sell new copies. If a corporation were in Apache's position, they would risk running out of money and having their existing software go unsupported. But Apache doesn't need to be supported by new sales, so having people stick with the old version is just fine.