... and here I thought Apple made most of their money with their hardware sales and a pittance on each track, giving the majority to the producer.
He-he. You are SO wrong.
I have been updating myself with the terminology and (mainly Russian) copyright law due to AllOfMp3 case, so I easily see you misunderstanding.
Ending of the phrase should read: "giving the majority to the copyright holder."
Not to people holding "authorship" (== artists), nor to "producers" (== recording companies) - but to the people who took it all in their hands: "copyright holders". Most of the money are NOT going back to people who write music and songs. Nor to people who are involved in organizing concerts and arranging recording sessions for CD. They would go back to few who usually remain in dark, but control all money flows.
P.S. Just think how screwed your U.S. entertainment industry is if your language already made "artist" != "producer". The people who have "authorship" (and are on LP/CD covers) more or less automatically precluded from copyrights by your system.
Since environments are passed down from parent process to child, a preload-based rootkit would need to replace/sbin/init with a wrapper that sets LD_PRELOAD prior to running the real init, to make sure that every process on the system preloads the malicious library.
Environment isn't inherited by default. It is free service of simplified call exec().
Also, init has no environment, first, and, second, login is what creates a new copy of environment for every new user session. So you would need to replace login with root-kit. (login is called by *getty (which handles initial setup of terminal) when you enter username. *getty is instantiated by init. Chain is: init -> *getty -> login -> -sh.)
IOW, it would work, but bit differently from what you have outlined;)
well-written preload rootkit would intercept execve()
I wonder why the possibility never crossed my mind.
like the "int 21h"
It is actually "int 0x80". libc always choses best way to call kernel, but Linux (and I believe *BSD) do support entry thru "int 0x80" (on IA-32 of course).
Additionally, libc has prototypes and macros for syscalls. If application would use the macro to generate a call to syscall, then you wouldn't be able to intercept it - because it would be inlined right in the calling code.
Still, it could be fairly effective if done right.
True.
LD_PRELOAD wouldn't work for suid-root applications, so running fresh "login" should kind of sanitize environment. But I think that now only few applications are suid-root, so shaking off the thing might be a real pain.
P.S. And of course do not forget, that LD_PRELOAD works only for dynamically linked apps. In older times, every system had for bootstrap purposes a copy of statically linked shell, which would be immune to LD_PRELOAD crack vector.
-Time will tell. But almost all customers we have been talking to thinks it would be good if OOXML became an ISO standard.
Reading between lines. Who are the "all customers" of Microsoft who think it would be "good"?? (*)
Does M$ still sell anything directly? I doubt it.
That leaves us with only option: the "customers" are partners/channel partners of M$ who do real deals selling hardware and OEMs. Shortly - partners.
Now, the phrase "almost all [snip]partners[/snap] we have been talking to thinks it would be good if OOXML became an ISO standard" as opposed to original quote sounds logical: for M$'s partners OOXML being a standard means more sales.
(*) I was alarmed by the phrase, because when refusing to implement some features, M$ always refers to the mythical customers who did not ask. You know, all internet - blogosphere and magazines included - retell story that feature is demanded by many many users. But then M$ PR droids come out and say: "we see no customer demand." "We see" part is pretty clear - they "see" only what they want to see. But the "customer" part was always puzzling.
under linux you could use the ld_preload variable to intercept any library call.
Not much effective, since most of secure software is written in special scripting languages (e.g. perl -T) and uses analogue of '/usr/bin/env -' to clean environment before doing anything. And has bunch of secure wrappers before doing anything serious.
And do not forget, that with ld_preload you can overload libc calls - but it all fails if application calls syscalls directly. And some applications do call syscalls just to check whether environment is sane and libc isn't screwed. (What happens sometimes even w/o cracking attempts.)
And well, user can always do manually in terminal 'unset LD_PRELOAD'... How would you protect root-kit against that? Environment is just lump of RAM with bunch of strings in it - application (consequently user) can do with it whatever it/s/he likes to.
It's possible. For particular known in advance kernel version. In other words, thanks to multitude of Linux configurations, such attack vector isn't practically feasible. Rootkits try to patch syscall table but it is not always trivial from user-space. And again - not reliable. Now with so short update cycle (about 3-6 month) I haven't seen Linux root-kits in a wild for very very long time. Before in 2.0/2.2 times there were root-kits as well as popular security systems against them.
On other side, Linux file system API does support so called namespaces (or what windows calls mount points). IOW it is possible to remove something so it would be invisible to user and his/her applications. But then it is feature for user - not against user - so s/he can easily see that something was manipulated and undo the manipulations.
Gigabit Ethernet adapters are specially designed to not to burden CPU nor buses. Lots of queueing/processing is done by the NICs themselves. And most of the burden goes to peripheral bus - not CPU itself.
IOW, you presumption is wrong. On average, GE NIC might load system less than FE (Fast Ethernet) one. If you would be trying to pump gigabit traffic - then your bus would start choking first. And you would notice that pretty quick: everything (not only sound) would start crawling.
All Linux systems I have worked last two-three years had GE NICs. My actual PC has two of them. And no performance degradations were ever spotted.
But Windows Mixer is a well known sucker. Even has earned one hardware workaround. People just got sick begging M$ to fix mixer and did all the work by themselves.
when Firefox is consuming 65-70% of main memory and slower than #%#$
I hope you also understand that when you run applications on system which cannot probably support them, expecting perfect performance is just silly. And expecting some interactivity when system is swapping - is just stupid of you.
What's more, nobody is going to compromise design of kernel for some edge case relevant to obscure minority which cannot configure their hardware decently (or get some spares on flea market - right where you did get the computer!).
Though, of course, if you have idea in your mind how to fix your problem you are always welcome to lkml with patches;)
That was discussed couple of years ago and there were no solution found. I mean FAT32 is no solution - more of a problem. Albeit being read by most if not all OSs.
Many people in past had recommended for OS specific stuff to use ZIP archives (since they are also universally available). Additionally to preserve verbatim information from *nix/MacOS volumes you can create disk image (laying on FAT32 volume). All decent OSs allow you to mount such disk images. Formats are different so it is not portable solution to preserve not portable OS-specific information about files.
Just to reiterate FAT32 is more or less only such solution.
P.S. I have looked also into ext2 support. In MacOS 10.3.x there were no official drivers (nor such drivers materialized in 10.4). Second party solution (I found only one) crashed my MacOS during installation and didn't worked in the end. For Windows there are multiple working ext2 solutions. Though not nice, yet allowing you to extract your files from ext2 volume. Not fitting for usual everyday work - but passable.
To me the real question, would be any meat in the show? or just another jedi/force show off?
I personally have seen enough of it. New story set in Star Wars world would be definitely more interesting then reiteration of all the stuff we all read/seen/watched before many times.
So it's just a question: would Lucas manage to find a good writers for the show? If story line would be good, I frankly can tolerate crappy graphics. But if it would be great graphics and thousand times repeated again "let the force be with you, luke"... heck, I had enough of it already. The story is over, I want something new.
I'm getting tired of the back and forth between AMD and Nvidia.
+10
That's why I have bought a Wii. It's graphics... sucks. But games are good - because gameplay is good.
Wii - is my response to all the crap drivers crap (or cock size competition) both ATI and nVidia started many years ago. Momentarily both ATI and nVidia are winning (judging by their PR) and apparently it's customers who lost the race.
[...] Nintendo' Brain Age: 'It's certainly a very different kind of game from Halo or even Miyamoto's own Zelda series [...]
And if anybody cared to check playing population, they would have found say 5-10% overlap of said demographics.
To start playing Brain Age, you need no previous knowledge. You can take it any time for five minute ride. And since it's portable DS game it can be always with you.
Halo is shooter. To many normal people games about going around and killing others are... repugnant. Simple as that. And Halo also requires some training. And you would go through the training only if you are in to the genre. If you are now sci-fi shooter fan - Halo is just not for you.
Zelda is fantasy action. Which again (as Halo to sci-fi shooter) would appeal to people who are into fantasy action.
What I'm trying to get to, is that Brain Age is game anybody can play - as long as you want your brain exercised. And some time killed. For Halo/Zelda - you really need to be a youth with flood of hormones to waste and brains looking to be filled with new information and new experiences. (And since men are never grow ^_^ such games remain with us for long time.) Feel the difference.
[..] boring [...] mundane [...]
Would you call soccer boring? 22 guys run over field hitting ball. Sometimes they even hit something with the ball. But rarely.
Mundane? Yes. Boring? Depends.
P.S. Notice that adult games like Brain Age are made not for a challenge - but just to pass time pleasantly. The games for kids - Halo/Zelda - are made as challenges young people love to take on to stand out of crowd. They are games, but game journalists being mostly (ex-)hard-core gamers (who else would want to cover games??) are obviously biased, since they used to games as challenges, not to games as way to pass time.
I personally, find more more action in Brain Age than I find in Mario.
Intel has designed and begun marketing it's own low-cost laptop targeted at education in developing countries.
Shouldn't that read "targeted against OLPC in developing countries"?
Just like with Intel v. Motorola (== i386), Intel v. AMD (== x64) and Intel v. Transmeta (== Centrino), Intel has to be hit hard in testicles to start doing anything - especially something targeted at consumers.
I cannot put this into polite form. In original form it sounds like "because NWN2 sucks big time".
NWN2 now at 1.05 revision and it still crashes often enough for me to stop playing it completely.
Worst part of crashes, for 1.05 game crashes after lengthy cut scenes: you can't avoid them, you can't save before them (because you do not know they are coming) and you can't skip them. In my third attempt of NWN2 I have waisted about 3 or 4 hours of game play due to two crashes. Third time I said "f*ck it", uninstalled NWN2, went to shop and bought NWN1 Platinum. It feels so-o-o-o much better. No crashes - and it literally flies on my amd4200+x2/7800gt. And depth of expansion packs - SoU and HotU - make NWN2 single player campaign look like Diabolo (== senseless non-stop shooting) in comparison.
ARP & ICMP are not layers - they are part of layer implementations.
ARP is part of IP over Ethernet implementation. (And of course there are bunch of ARPs - for every physical layer IP runs over.)
ICMP is part of IP and also used to deliver remote TCP/UDP errors.
Please, don't ever offer TCP/IP networking stack as an example of proper layering.
But it is example of "proper layering": layering which (compared to ISO OSI) have stood to test of time. It also really works.
Your confusion comes from the fact that layers with time have tendency to split into even more layers - to accommodate new requirements. (That's why rigid 7 layer OSI model only worked on paper - never in practice.) Many such splits are getting their own names - largely to promote and help interoperability. Network people like to have a name for everything.
Though, network layering has little to do with layering inside of kernel OS. Linux has strict layering since its hackers want Linux to remain simple. SunOS/Solaris is well know for breaking all kind of stuff. e.g. Solaris has three levels of multitasking: threads, light weight processes and normal processes. Linux has one. End result? Under Solaris context switch is bitch. Under Linux - it doesn't even register during profiling. Same applied to I/O subsystem: strict balance has to be stroked between disk drivers, file systems and user applications. Breaking layering to optimize for single application is easy (as Solaris 10 did to show off that it runs Apache2 faster) but to provide consistent performance across board is one hell of complicated target. And performance-wise all-so-elite Solaris 10 with its uber-kawl ZFS doesn't come even close to what any non-elite Linux user has on his/her desktop.
And can anybody explain me why new file system?? And why all the hype? (I understand the hype PR-wise: Sun has barely any outstanding product right now and without hype everybody would forget about Sun pretty quickly.) We under Linux have half dozen of them. Yet as practice have shown that end users care NOT what file system they run. Most people do not know what file system is - they just want place to save their file on. Those who care, can and would choose based of features/support/stability/etc and all the hype for them is just distraction.
From RTFA:
ZFS solves a huge number of long-standing problems for users, system administrators, and hardware makers by rejecting decades of entrenched "wisdom" about filesystems and volume management.
As "long-standing" user, administrator and also developer, I'd like to now the list of the "long-standing" problems. For I haven't had single problem with e.g. ext2 (simplest but feature full fs) for all the time I have used it. Reading Wikipedia's ZFS page I see nothing what ZFS "solves". Looks like another snowball-like agglutination of features - oh so typical to proprietary vendors. I wonder how much jobs would create wider adoption of ZFS.
Man pages which are displayed by BSD's unique 'more'. Thanks you very much.
If you are really into command line, then check out Gentoo. Right after installing you would get feeling that Gentoo devels are really using command line - since it is so well made and polished.
Targeted at Linux users you say? "If you've been using Linux for a few years and have chosen the command line lifestyle over a gui lifestyle then OpenBSD should interest you." WTF?
BSD is dead. As long as they have the antique command line tools.
Think whatever you want, but I cannot live w/o GNU command line. bash alone isn't sufficient - text-tools, file-tools are also important.
e.g. BSD's moronic find requires directory name - while GNU one picks current directory by default. All GNU tools support --help and --version - try to find common help displaying option in BSD variants. Not that BSD tools helps output is any useful anyway. Also BSD's ps suck big time. The stupid insistence on using 'more' instead of 'less' isn't helping either.
Also, it might surprise you, 'vi' is no more. Everybody had forgotten what it is - for good - and are using 'vim' instead. But the fact remain: BSD has no sane decent text editor preinstalled. Because POSIX 'vi' cannot be called 'sane' nor 'decent'.
GNU tools by themselves are already pretty old - and kind of outdated. BSD tools are just rotten dead. And with them to me whole BSD: kernel is good, shell around it - is dead.
Constructive note. BSD should align themselves with Debian or Gentoo. Yeah, I know BSD was first. But first doesn't mean better - and those who came after BSD had learned on its mistakes - while BSD remained in its slumber. BSD can learn from Debian it's excellent software management facilities (/etc/alternatives is just god send). From Gentoo BSD can learn how to build software easily in a user-friendly fashion: integrating cross compilation, ccache, distcc all for good fast build on any system.
The article goes into into technical side of merge highlighting technical differences between the two file formats. And from my reading it seems like UOF is superior to ODF in many aspects.
With a legal license, you have the RIGHT to have a copy of it.
Geeee. Another poor soul fallen to RIAA/MPAA propaganda. Having a copy is (and always was) OK. (They would need to prove that you have acquired it illegally - and knowing of illegality - but this is extremely complicated. And of course, you might have acquired it legally e.g. in course of private sharing (you friend made a private copy for you) (*).)
It is distribution for which you need to have a permission of copyright holder/licensor.
They aren't trying to confirm that you have a copy - they try to confirm that you share a copy with others thus distributing file without proper license. That's a breach of copyright law.
IOW, getting file off P2P is pretty OK - you were not bound by license agreement nor were you aware of copyrights attached to the file. (Protection of copyright is duty of copyright holder, not yours.) Putting that file back on to P2P network is of course illegal since that's distribution.
(*) In US they might try to ask who is your friend and try to convince him for breach of DMCA - for removing copy protection measures. But they cannot force you tell them who your friend is. You are not obliged to tell them.
Since you are selling not original - but copy - no way it would classify as "first sale". IOW, private copies are reserved for private use - sale/rent/etc aren't private uses.
Still that doesn't make suddenly targeting any easier in Gears of War. The standard controller is good - but not for shooters.
To get most out of such games - you still need a "point and shoot" type of controller. And in fact such controllers were developed quite some time ago: I am pretty surprised that "next gen" consoles do not offer anything like that while providing FPS games at launch.
If not MSFT itself, then probably some courageous game producer/publisher might bundle such "gun" controller with game (same way as Guitar Hero does "guitar" controller).
First Person Shooters. Standard game controllers SUCK. Big time.
Wii is my first console. After some Zelda bow shooting action, I have tried Xbox360' Prey & Gears of War. Controls are... "terrible" - that is only word I can come with. Gears of War are not that bad - A+motion is great - but targeting is awful. Prey has totally stupid control scheme - totally unplayable.
I was initially complaining that WiiMote targeting is much harder compared to mouse. (Some might say realistic - the same problems (hand shaking, breathing, trigger release, etc) arise when shooting with real gun.) But after trying Xbox360 controller - I have to say that WiiMote is ages ahead of what Xbox360 (and PS3 I gather too) can offer to shooters.
I would love to see MSFT releasing special FPS controller - especially since top games (e.g. Gears of War) are precisely FPS and all can benefit enormously from such pointing interface.
Though WiiMote deserve some critique too: it has too few buttons - A & B. I do not think I'm alone who was frustrated that Zelda has no dedicated key to wield a sword. You have to swing controller instead. That's terribly slow and very unpractical in cases when you already in battle - e.g. sudden attack - the cases when you need your sword most.
Slashdot Mirror
He-he. You are SO wrong.
I have been updating myself with the terminology and (mainly Russian) copyright law due to AllOfMp3 case, so I easily see you misunderstanding.
Ending of the phrase should read: "giving the majority to the copyright holder."
Not to people holding "authorship" (== artists), nor to "producers" (== recording companies) - but to the people who took it all in their hands: "copyright holders". Most of the money are NOT going back to people who write music and songs. Nor to people who are involved in organizing concerts and arranging recording sessions for CD. They would go back to few who usually remain in dark, but control all money flows.
P.S. Just think how screwed your U.S. entertainment industry is if your language already made "artist" != "producer". The people who have "authorship" (and are on LP/CD covers) more or less automatically precluded from copyrights by your system.
Environment isn't inherited by default. It is free service of simplified call exec().
Also, init has no environment, first, and, second, login is what creates a new copy of environment for every new user session. So you would need to replace login with root-kit. (login is called by *getty (which handles initial setup of terminal) when you enter username. *getty is instantiated by init. Chain is: init -> *getty -> login -> -sh.)
IOW, it would work, but bit differently from what you have outlined ;)
I wonder why the possibility never crossed my mind.
It is actually "int 0x80". libc always choses best way to call kernel, but Linux (and I believe *BSD) do support entry thru "int 0x80" (on IA-32 of course).
Additionally, libc has prototypes and macros for syscalls. If application would use the macro to generate a call to syscall, then you wouldn't be able to intercept it - because it would be inlined right in the calling code.
True.
LD_PRELOAD wouldn't work for suid-root applications, so running fresh "login" should kind of sanitize environment. But I think that now only few applications are suid-root, so shaking off the thing might be a real pain.
P.S. And of course do not forget, that LD_PRELOAD works only for dynamically linked apps. In older times, every system had for bootstrap purposes a copy of statically linked shell, which would be immune to LD_PRELOAD crack vector.
Reading between lines. Who are the "all customers" of Microsoft who think it would be "good"?? (*)
Does M$ still sell anything directly? I doubt it.
That leaves us with only option: the "customers" are partners/channel partners of M$ who do real deals selling hardware and OEMs. Shortly - partners.
Now, the phrase "almost all [snip]partners[/snap] we have been talking to thinks it would be good if OOXML became an ISO standard" as opposed to original quote sounds logical: for M$'s partners OOXML being a standard means more sales.
(*) I was alarmed by the phrase, because when refusing to implement some features, M$ always refers to the mythical customers who did not ask. You know, all internet - blogosphere and magazines included - retell story that feature is demanded by many many users. But then M$ PR droids come out and say: "we see no customer demand." "We see" part is pretty clear - they "see" only what they want to see. But the "customer" part was always puzzling.
Not much effective, since most of secure software is written in special scripting languages (e.g. perl -T) and uses analogue of '/usr/bin/env -' to clean environment before doing anything. And has bunch of secure wrappers before doing anything serious.
And do not forget, that with ld_preload you can overload libc calls - but it all fails if application calls syscalls directly. And some applications do call syscalls just to check whether environment is sane and libc isn't screwed. (What happens sometimes even w/o cracking attempts.)
And well, user can always do manually in terminal 'unset LD_PRELOAD'... How would you protect root-kit against that? Environment is just lump of RAM with bunch of strings in it - application (consequently user) can do with it whatever it/s/he likes to.
It's possible. For particular known in advance kernel version. In other words, thanks to multitude of Linux configurations, such attack vector isn't practically feasible. Rootkits try to patch syscall table but it is not always trivial from user-space. And again - not reliable. Now with so short update cycle (about 3-6 month) I haven't seen Linux root-kits in a wild for very very long time. Before in 2.0/2.2 times there were root-kits as well as popular security systems against them.
On other side, Linux file system API does support so called namespaces (or what windows calls mount points). IOW it is possible to remove something so it would be invisible to user and his/her applications. But then it is feature for user - not against user - so s/he can easily see that something was manipulated and undo the manipulations.
Gigabit Ethernet adapters are specially designed to not to burden CPU nor buses. Lots of queueing/processing is done by the NICs themselves. And most of the burden goes to peripheral bus - not CPU itself.
IOW, you presumption is wrong. On average, GE NIC might load system less than FE (Fast Ethernet) one. If you would be trying to pump gigabit traffic - then your bus would start choking first. And you would notice that pretty quick: everything (not only sound) would start crawling.
All Linux systems I have worked last two-three years had GE NICs. My actual PC has two of them. And no performance degradations were ever spotted.
But Windows Mixer is a well known sucker. Even has earned one hardware workaround. People just got sick begging M$ to fix mixer and did all the work by themselves.
Well, this is very very very old Russian hacker tradition: introduce flaw in the crack/exploit to prevent it from being (ab)used by idiots.
I hope you also understand that when you run applications on system which cannot probably support them, expecting perfect performance is just silly. And expecting some interactivity when system is swapping - is just stupid of you.
What's more, nobody is going to compromise design of kernel for some edge case relevant to obscure minority which cannot configure their hardware decently (or get some spares on flea market - right where you did get the computer!).
Though, of course, if you have idea in your mind how to fix your problem you are always welcome to lkml with patches ;)
That was discussed couple of years ago and there were no solution found. I mean FAT32 is no solution - more of a problem. Albeit being read by most if not all OSs.
Many people in past had recommended for OS specific stuff to use ZIP archives (since they are also universally available). Additionally to preserve verbatim information from *nix/MacOS volumes you can create disk image (laying on FAT32 volume). All decent OSs allow you to mount such disk images. Formats are different so it is not portable solution to preserve not portable OS-specific information about files.
Just to reiterate FAT32 is more or less only such solution.
P.S. I have looked also into ext2 support. In MacOS 10.3.x there were no official drivers (nor such drivers materialized in 10.4). Second party solution (I found only one) crashed my MacOS during installation and didn't worked in the end. For Windows there are multiple working ext2 solutions. Though not nice, yet allowing you to extract your files from ext2 volume. Not fitting for usual everyday work - but passable.
It looks not bad. For a TV show.
To me the real question, would be any meat in the show? or just another jedi/force show off?
I personally have seen enough of it. New story set in Star Wars world would be definitely more interesting then reiteration of all the stuff we all read/seen/watched before many times.
So it's just a question: would Lucas manage to find a good writers for the show? If story line would be good, I frankly can tolerate crappy graphics. But if it would be great graphics and thousand times repeated again "let the force be with you, luke" ... heck, I had enough of it already. The story is over, I want something new.
+10
That's why I have bought a Wii. It's graphics ... sucks. But games are good - because gameplay is good.
Wii - is my response to all the crap drivers crap (or cock size competition) both ATI and nVidia started many years ago. Momentarily both ATI and nVidia are winning (judging by their PR) and apparently it's customers who lost the race.
And if anybody cared to check playing population, they would have found say 5-10% overlap of said demographics.
To start playing Brain Age, you need no previous knowledge. You can take it any time for five minute ride. And since it's portable DS game it can be always with you.
Halo is shooter. To many normal people games about going around and killing others are ... repugnant. Simple as that. And Halo also requires some training. And you would go through the training only if you are in to the genre. If you are now sci-fi shooter fan - Halo is just not for you.
Zelda is fantasy action. Which again (as Halo to sci-fi shooter) would appeal to people who are into fantasy action.
What I'm trying to get to, is that Brain Age is game anybody can play - as long as you want your brain exercised. And some time killed. For Halo/Zelda - you really need to be a youth with flood of hormones to waste and brains looking to be filled with new information and new experiences. (And since men are never grow ^_^ such games remain with us for long time.) Feel the difference.
Would you call soccer boring? 22 guys run over field hitting ball. Sometimes they even hit something with the ball. But rarely.
Mundane? Yes. Boring? Depends.
P.S. Notice that adult games like Brain Age are made not for a challenge - but just to pass time pleasantly. The games for kids - Halo/Zelda - are made as challenges young people love to take on to stand out of crowd. They are games, but game journalists being mostly (ex-)hard-core gamers (who else would want to cover games??) are obviously biased, since they used to games as challenges, not to games as way to pass time.
I personally, find more more action in Brain Age than I find in Mario.
Shouldn't that read "targeted against OLPC in developing countries"?
Just like with Intel v. Motorola (== i386), Intel v. AMD (== x64) and Intel v. Transmeta (== Centrino), Intel has to be hit hard in testicles to start doing anything - especially something targeted at consumers.
I cannot put this into polite form. In original form it sounds like "because NWN2 sucks big time".
NWN2 now at 1.05 revision and it still crashes often enough for me to stop playing it completely.
Worst part of crashes, for 1.05 game crashes after lengthy cut scenes: you can't avoid them, you can't save before them (because you do not know they are coming) and you can't skip them. In my third attempt of NWN2 I have waisted about 3 or 4 hours of game play due to two crashes. Third time I said "f*ck it", uninstalled NWN2, went to shop and bought NWN1 Platinum. It feels so-o-o-o much better. No crashes - and it literally flies on my amd4200+x2/7800gt. And depth of expansion packs - SoU and HotU - make NWN2 single player campaign look like Diabolo (== senseless non-stop shooting) in comparison.
ARP & ICMP are not layers - they are part of layer implementations.
ARP is part of IP over Ethernet implementation. (And of course there are bunch of ARPs - for every physical layer IP runs over.)
ICMP is part of IP and also used to deliver remote TCP/UDP errors.
But it is example of "proper layering": layering which (compared to ISO OSI) have stood to test of time. It also really works.
Your confusion comes from the fact that layers with time have tendency to split into even more layers - to accommodate new requirements. (That's why rigid 7 layer OSI model only worked on paper - never in practice.) Many such splits are getting their own names - largely to promote and help interoperability. Network people like to have a name for everything.
Though, network layering has little to do with layering inside of kernel OS. Linux has strict layering since its hackers want Linux to remain simple. SunOS/Solaris is well know for breaking all kind of stuff. e.g. Solaris has three levels of multitasking: threads, light weight processes and normal processes. Linux has one. End result? Under Solaris context switch is bitch. Under Linux - it doesn't even register during profiling. Same applied to I/O subsystem: strict balance has to be stroked between disk drivers, file systems and user applications. Breaking layering to optimize for single application is easy (as Solaris 10 did to show off that it runs Apache2 faster) but to provide consistent performance across board is one hell of complicated target. And performance-wise all-so-elite Solaris 10 with its uber-kawl ZFS doesn't come even close to what any non-elite Linux user has on his/her desktop.
And can anybody explain me why new file system?? And why all the hype? (I understand the hype PR-wise: Sun has barely any outstanding product right now and without hype everybody would forget about Sun pretty quickly.) We under Linux have half dozen of them. Yet as practice have shown that end users care NOT what file system they run. Most people do not know what file system is - they just want place to save their file on. Those who care, can and would choose based of features/support/stability/etc and all the hype for them is just distraction.
From RTFA:
As "long-standing" user, administrator and also developer, I'd like to now the list of the "long-standing" problems. For I haven't had single problem with e.g. ext2 (simplest but feature full fs) for all the time I have used it. Reading Wikipedia's ZFS page I see nothing what ZFS "solves". Looks like another snowball-like agglutination of features - oh so typical to proprietary vendors. I wonder how much jobs would create wider adoption of ZFS.
Man pages which are displayed by BSD's unique 'more'. Thanks you very much.
If you are really into command line, then check out Gentoo. Right after installing you would get feeling that Gentoo devels are really using command line - since it is so well made and polished.
C'mon, we do not want to start the old song again. You know, most first Linux drivers were ported over from BSD.
Development of both BSD & Linux isn't commercialized - so word "lifted" is unfit here. It is more about "exchange of ideas" ;)
Targeted at Linux users you say? "If you've been using Linux for a few years and have chosen the command line lifestyle over a gui lifestyle then OpenBSD should interest you." WTF?
BSD is dead. As long as they have the antique command line tools.
Think whatever you want, but I cannot live w/o GNU command line. bash alone isn't sufficient - text-tools, file-tools are also important.
e.g. BSD's moronic find requires directory name - while GNU one picks current directory by default. All GNU tools support --help and --version - try to find common help displaying option in BSD variants. Not that BSD tools helps output is any useful anyway. Also BSD's ps suck big time. The stupid insistence on using 'more' instead of 'less' isn't helping either.
Also, it might surprise you, 'vi' is no more. Everybody had forgotten what it is - for good - and are using 'vim' instead. But the fact remain: BSD has no sane decent text editor preinstalled. Because POSIX 'vi' cannot be called 'sane' nor 'decent'.
GNU tools by themselves are already pretty old - and kind of outdated. BSD tools are just rotten dead. And with them to me whole BSD: kernel is good, shell around it - is dead.
Constructive note. BSD should align themselves with Debian or Gentoo. Yeah, I know BSD was first. But first doesn't mean better - and those who came after BSD had learned on its mistakes - while BSD remained in its slumber. BSD can learn from Debian it's excellent software management facilities (/etc/alternatives is just god send). From Gentoo BSD can learn how to build software easily in a user-friendly fashion: integrating cross compilation, ccache, distcc all for good fast build on any system.
Andy Updegrove gave interesting link to article: Sun's McNealy Proposes Merging ODF with Chinese Counterpart .
The article goes into into technical side of merge highlighting technical differences between the two file formats. And from my reading it seems like UOF is superior to ODF in many aspects.
Geeee. Another poor soul fallen to RIAA/MPAA propaganda. Having a copy is (and always was) OK. (They would need to prove that you have acquired it illegally - and knowing of illegality - but this is extremely complicated. And of course, you might have acquired it legally e.g. in course of private sharing (you friend made a private copy for you) (*).)
It is distribution for which you need to have a permission of copyright holder/licensor.
They aren't trying to confirm that you have a copy - they try to confirm that you share a copy with others thus distributing file without proper license. That's a breach of copyright law.
IOW, getting file off P2P is pretty OK - you were not bound by license agreement nor were you aware of copyrights attached to the file. (Protection of copyright is duty of copyright holder, not yours.) Putting that file back on to P2P network is of course illegal since that's distribution.
(*) In US they might try to ask who is your friend and try to convince him for breach of DMCA - for removing copy protection measures. But they cannot force you tell them who your friend is. You are not obliged to tell them.
That's very nice. I own copyright for the comment - so I can practice pretext!!!
RIAA you are cool! Keep good job turning U.S. into true oligarchy!! Next please outlaw those free speech suckers!!! [/sarcasm]
Since you are selling not original - but copy - no way it would classify as "first sale". IOW, private copies are reserved for private use - sale/rent/etc aren't private uses.
P.S. IANAL
Still that doesn't make suddenly targeting any easier in Gears of War. The standard controller is good - but not for shooters.
To get most out of such games - you still need a "point and shoot" type of controller. And in fact such controllers were developed quite some time ago: I am pretty surprised that "next gen" consoles do not offer anything like that while providing FPS games at launch.
If not MSFT itself, then probably some courageous game producer/publisher might bundle such "gun" controller with game (same way as Guitar Hero does "guitar" controller).
First Person Shooters. Standard game controllers SUCK. Big time.
Wii is my first console. After some Zelda bow shooting action, I have tried Xbox360' Prey & Gears of War. Controls are ... "terrible" - that is only word I can come with. Gears of War are not that bad - A+motion is great - but targeting is awful. Prey has totally stupid control scheme - totally unplayable.
I was initially complaining that WiiMote targeting is much harder compared to mouse. (Some might say realistic - the same problems (hand shaking, breathing, trigger release, etc) arise when shooting with real gun.) But after trying Xbox360 controller - I have to say that WiiMote is ages ahead of what Xbox360 (and PS3 I gather too) can offer to shooters.
I would love to see MSFT releasing special FPS controller - especially since top games (e.g. Gears of War) are precisely FPS and all can benefit enormously from such pointing interface.
Though WiiMote deserve some critique too: it has too few buttons - A & B. I do not think I'm alone who was frustrated that Zelda has no dedicated key to wield a sword. You have to swing controller instead. That's terribly slow and very unpractical in cases when you already in battle - e.g. sudden attack - the cases when you need your sword most.