Wireless services (for phone, network, etc) and even local or co-gen electric service is popular for another, simpler reason also: People will dig up copper wire and steal it to sell. Wireless kind of prevents that from happening.
No the cubic mapping problem can be seen as an uneven distribution of image resolution -- ie when you are looking directly at a cube face you have less resolution than when you are looking into a cube corner. In practice it really just reduces sharpness.
I think the halflife image has the advantage of having been saved with a lower FOV. The aspect ratio is also a little better which also mitigates the effect somewhat. If you zoom into my kite pano quite a bit you can get the building corners and whatnot to square up
On second thought, I think I might know what you are referring to... It's not a stitching or composition problem, but it boils down to this: When you pan around a QTVR the high default field of view can make things look out of perspective.. IE 90 degree corners look sharper on the edge vs in the center of the picture. If you click the "+" to zoom in you can sort of make this problem go away at the expense of a reduced field of view.
The other minor problem is the way QTVR maps panoramas. QTVR uses a cubic map which essentially draws the panorama onto the inside faces of a cube. Sphere mapping looks marginally better and has finer detail, but remember QTVR was designed back in the days of the 68k macs and thus speed was of greater importance than consistent image resolution.
As an aside most FPS games set the FOV a little higher than it ought to be for a 3:4 display, so maybe these scenes look normal because you are used to seeing them with a high FOV in the games... Or maybe the default FOV for the QTVR is something more appropriate than QTVR's default (like 120 degrees or so)
I don't know what you're talking about. With a proper pano head and some decent stitching effort you can create 'perfect' panoramas without distortion. It helps to have the right equipment and knowledge though. Most panoramas out on the web are unfortunately shot on a regular tripod (which does not pivot around the right point) or worse, handheld, and stiched using some low quality automatic stitching software (Ulead 360, etc)...
Using something that actually *works* like panotools and good wide angle lenses, you can whip up good looking shots with minimal effort.
Here's one of my first 360's made with two shots and a lens with a 183 degree FOV. There are some problems with the horizon due to it being in the extremeties of the picture, but aside from the lack of resolution, it is pretty much seamless except for one small area by the sailboat where I simply could not get panotools to bend things right. If I'd have had more altitude with the kite, it would have been perfect. Next time I shoot one, it will be:) My shot was inspired by Scott Haefner's work on the same subject; however he has far more experience at it than I do (I have only made one shot).. If you really want to see some awesome stuff, check out his site.
What linux distribution installs perl, and not python, by default?
Just answering your question here; not trying to suggest one thing over another, but Debian and most Debian-based distributions do a minimal install of perl (perl-base) with the most basic core install while python is out in the cold. Lots of packages rely on perl to run pre/post install scripts. I am not sure if languages other than sh and perl are actually allowed for this task, so this fact may simply be an artifact of the packaging requirements/guidelines.
Aside from the perl binary and some very basic utility packages, there's not much of perl inside of perl-base, but it is sufficient to run a lot of self-contained perl scripts. Full fledged installations of neither perl nor python are part of the 'default' install though, but if you select certain tasks usually one or both will be installed as part of a dependency.
I will take this opportunity to pimp my own cabinet, however I should also note that asking slashdot is a poor substitute for visiting and exploring the resources at Arcade Controls. The maintainer of the site also has compiled much of the knowledge and experience collected there into a nice book. Do yourself a favor and pick up a copy. If there's something that's not covered, ask in the forums there. It's a far more appropriate place.
Depending on the make/model of car, how the keys are cut/programmed, etc. it's entirely possible that you would have to pay loads of money to get the keys replaced.
On new BMW's for instance, there are several (I think I was told 7) sets of keys cut and programmed when the locksets and ignition were installed into the car. If you lose the keys, you can order your replacement keys while they still exist. After you run out of keys, you have to replace all the locks and the ignition in the car and have the ECU reprogrammed for your new key - This unfortunately costs a heck of a lot of money, but I guess if you are stupid enough to lose your key that many times, you'd probably need to change your locks anyway.
I never said boot speed is the most important metric for anything. I use linux all the time. I rarely care about the boot process, except for instance in situations like my car pc where boot time is absolutely 100% the most critical performance metric of the entire software system.
The most important thing in a boot process is a standard dependancy scheme. Windows has none that works, even in WinXP, and I think not even in Win2003, but not sure. We are in 2005 for christ sake, and MS is still unable to produce a boot process that works, and yes they try, it just still does not work right.
How do you figure? I am not sure I follow your reasoning here since you provide no evidence of anything that you mean by "standard dependency scheme". I don't claim to be an expert, but booting is a fairly straightforward process on any X86 sytem (and most other systems too for that matter)... Bootstrap loads kernel, kernel inits hardware, kernel loads drivers, kernel starts userspace. Userspace starts processes necessary for low level system stuff. Userspace starts user-interactive. I don't think that either Linux or Windows provides a 100% fault tolerant boot process.
It's not all the fault of the software though; I have a couple of servers (Tyan dual athlon systems) that take longer to POST than they take to boot any OS.
simpleinit-msb that I use do all of that fast since several years, and it just works. I could also move X start in an earlier runlevel (I did once), but it just has no sense other than prove that you can start a Linux GUI fast too...
This is the simpleinit that is deprecated by LSBInit - a project that exists only on paper, right? I am familiar with many sysV replacements; however none of them come close to what a modern PC ought to be capable of. The closest project that I know and have used is LinuxBIOS + hand written busybox-based init.. About two seconds from the power button to a directfb UI with full hardware init (except for the IDE subsystem that takes about 5s to come up, but this was handled in the background along with a lot of other kernel module loading for drivers unessential in the first few seconds of system operation such as sound, usb, etc.
Fortunately, at least for windows users, due to the sheer number of times they have to reboot their damn machines, the boot process is fairly streamlined. XP boots on most of my hardware from power on to being logged in and fairly usable in about 30 seconds.
Initially, I thought they had come a long way in this regard, which I suppose may be true when you consider all the junk that is loading up, but what is suprising is that if you boot windows 95 on a modern pc, you can be up to the destkop in about 5 seconds after the BIOS is done.
Love it or hate it, Microsoft deserves at least a little credit for getting the boot process done relatively quickly. I can't wait to see a nice implementation of Redhat's early login efforts. Linuxbios and many other embedded machines have fantastic boot processes that really show what a modern computer can be capable of. My ceiva picture frame (hacked to run linux) boots inside of 5 seconds, and that includes the 1s timeout where it does nothing except allow the user to interrupt the boot.
You do of course remember the time when you had to reboot to change your IP address, don't you? I don't believe MS ever wholeheartedly claimed that you would never have to reboot when installing drivers and whatnot, but you certainly don't have to reboot nearly as much. There's very few modern apps that make you reboot after an install and drivers that do not require rebooting are not uncommon. I don't expect this to change particularly much with Longhorn either.. Certain changes are going to make you reboot. This is because (just like microsoft claims) a lot of companies produce crappy drivers written by lazy programmers that do not want to be bothered about having to write code that can properly init the hardware from any state.
Now, if MS made it a requirement of the windows driver certification that drivers load and unload cleanly without requiring a reboot, then we might make some progress on this front. A clean and better way to replace DLL's and other shared code libs without requiring a reboot to ensure that no other program is locking them open would also help a great deal.
Are you implying that you have to reboot Linux in order to install the video driver? You certainly don't but then again to all the "but you don't have to rebootpeople -- you do have to restart X, which is something of a pain if you don't have a good session manager. To the desktop user, a crashed X is just as destructive as a crashed kernel, and likewise a restart of X is just as interruptive as having to reboot.
I'd imagine that some code to 'ssupend/resume' the state of X might be a pretty neat project to undertake, but I'm not sure anyone has done it yet..
How about you set your mailbox filter to rely on a header rather than a subject tag? If using spamassassin, filter on "X-Spam-Status: yes" rather than whatever markings happen to be in the subject line. The forward (depending on the mail client) ought not to contain this same header.
This is also good practice to use on mailing lists too. Mailman and the like generally include X-Been-There headers. Filtering on this header instead of the subject line has all kinds of benefits such as personal responses to your postings on the list do not get stuffed into the list's mailbox, etc.
This is not particularly news. Some "blogger" discovers something because he never bothered to ask and screams something about the sky is falling.. I'm kind of sick of this "news" reporting. Incidentally, this same issue affects owners of most plasma and LCD tv's with native resolutions below 1920x1080 too.. depending on how you look at it as a problem or not.
Anyway, it's fairly well known that the internal scalers in many devices suck. That is why there is a market for good external scalers. If you are paranoid about watching a lot of 1080i on your 720p projector or LCD TV or Plasma, go buy a scaler. They cost about $1000 but will improve scaled display a lot.
At least if you have an external scaler you will have some options about how you convert 1080i to 720p. The article makes it sound like splitting the fields is a huge sin -- and it is if you discard one field per frame (Half field deinterlacing), but it's perfectly acceptible to scale EACH 540-line field to a seperate 720-line frame and double the framerate. This is called bob deinterlacing and is often the best for converting 1080i video to lower resolutions. If you are watching a 1080i upconvert of a film or something, though, you can have the scaler do full field deinterlacing and inverse telecine for you and see a nice 720p/24fps picture. Scalers also generally have internal audio delays for various types of audio feeds so you won't have to worry about AV sync issues either.
If you have any questions about how your device does this, you should try to find out before you buy it. Most devices don't publish how they do it, though, so your only option may be to derive it -- and that will not be an easy job.
That mkdir and rmdir use more than two letters is a long-standing bug--longer even than the screwy footnote/gap bug in words (which dates to Mac Word 1.0)
Well, Keynote reportedly uses the API; but it's not as if it couldn't do transitions another way - ie without using it. That being said, it's possible that the framework code implementing this is simply duplicated inside of the keynote application and implemented as a framework for the simple reason that several OS components need to use it as well.
There are a lot of applications where UMA or NUMA SMP systems beat the pants off of clusters; obviously these are applications that, while suited to parallel processing, do no good when you spend all of your time synchronizing the data set. Maybe there's not that many of them in the commercial market (ie general purpose applications) but there certainly is room for these systems.
There is another reason to have large n-way systems as well. Strange as it may sound, there is still some very expensive software out there that is licensed 'per server' and that sometimes means everything from a single cpu up to however many are managed by the same kernel. It can be way less expensive in certain cases to license one copy of a piece of software for a single 8-way machine than 8 copies for a cluster of 8 machines, even though the latter might even give better performance.
That is like saying people who plunk down a few hundred thousand dollars for a really nice (or at least expensive) car or boat will not tune it. $1600 is dirt cheap for a 4-way configuration; I expect that the overclockers will give it a lot more attention than they give current quad-cpu offerings.
Well, to be fair, there are 8 buttons per player not counting coin and start. N64 emulation requires lots of buttons, unfortunately, and it was one of my design goals. It looks sort of frankenstein, but it's actually very utilitarian and comfortable to use.
Anyway, I'm planning to use an LED driver board that will illuminate the buttons a certain game uses to uncomplicate things.
In retrospect, the N64 requirement would have been better served by some adaptoids and real N64 controllers. In that setup, I would have picked 7 buttons for p1 and p2 to get a combo layout between fighter and neogeo and only 3 or 4 buttons for p3 and p4.
I am actually planning on a small redesign of the upper deck. I never use the vanguard style buttons on the sides, and I need more function buttons. The top stick will probably get replaced with an analog stick using Ultimarc's new A-Pac and I'll be adding better controls up there then.
Also, I probably should have done the coin buttons by hooking up microswitches to the coin door rejects instead of having another button on the cpanel. Ah well, next time!
The new 9xxx series 3ware cards are much better than the old series and support command queueing and all that jazz that you get from a native SATA implementation. They're not even much more expensive either. The driver has finally made it into 2.6 too, but you gotta patch it into 2.4. You can get really really good performance on them for the cost - I run one with 4 WD Raptor 10K RPM 72GB drives and don't get any of the same kind of slowdowns I used to get on 3ware's older PATA cards, especially during heavy writes.
Thanks for the heads up on Areca cards. I'll keep that filed under 'check out for next time':)
The ol' software raid 1 boot trick depends highly on the behavior of your BIOS under a failed drive condition. This is not the same thing as you get when you unplug a drive. Some BIOS may boot fine; some may not boot from the 2nd hard drive if the first is still attached and failing. It may also depend on how your drive has failed. If the drive electronics are failed and shorting the wrong pins on your IDE controller, then you may not get past the drive detection code in the BIOS at all.
This is simply one advantage to using a real hardware raid card like the 3ware. There are plenty of other reasons too: Does your chipset/hardware support hot swapping? If you use SATA, does it support command queueing? Do your drives? How much cache does it have? Does it have cache? Can it tolerate all types of hardware failure? Does it have *ahem* 16 ports with individual controllers for each drive? It's not like the BIOS/IDE chipset makers write out in their specs how their hardware performs under drive failure conditions so you have the overhead of testing each configuration to make sure it works proeprly before you have to rely on it. It's not so much a performance difference between hardware and software raid (until RAID-5 anyway) but an issue with how the hardware will respond when something goes wrong, which is one of the primary reasons for using anything above RAID-0 anyway.
Yes, running a 3ware card costs more. There are times when that $400 costs a lot less than the time spent configuring and testing an alternative software-only implementation. There are times when it doesn't and spending another $400 doesn't make a lot of sense. I have run both setups. I have machines deployed with both IDE software-only RAID arrays, IDE 3ware arrays, SCSI software RAID5's, SCSI Adaptec RAID's etc.. it's all application specific. There's no reason to call somebody a goon for recommending 3ware hardware. It's really good hardware; maybe you should try it sometime.
I don't know why stating facts has to be questioned. If I told you that you must stop at a stop sign, would you ask for my source? Perhaps you are the orignal poster and are simply flustered that you posted misinformation. I didn't feel the need to duplicate a lot of information here.
If you really want to know the specifics, there are plenty of other posts elsewhere in this massive tree of threads with all the links you require to validate my statement. Try this post for instance. It gives you a nice overview with a paste from the actual law, and it includes links to several reputable sources to confirm it.
Pennies are not legal tender in debts of over $20. So nickels would be fine, dimes are good too, but pennies aren't good over $20 debts.
This incorrect fact is generally passed around with some specific amount attached to it -- most often $2, $5, $10, or $20, but at least in the United States it's not true.
All currency is legal tender in any amount; however, a business can be selective about what they are willing to accept. If a merchant wanted to ONLY accept payment in pennies, it would be legal, though likely unpopular.
You can't legally refuse to accept any real denomination of money in payment of a debt unless you establish it as a policy and post a notice to that effect in advance.
Yes; you can actually. The law is very specific about this, actually. It specifically is designed to prevent just the type of abuse you describe, such as your clever friend who likely spent a lot more time acquiring the pennies than the wrecking yard spent counting them.
That's why you see all those signs about not accepting bills over $50, etc... Without the sign, they can't use the denomination as an excuse not to take the money.
Ever tried to win an argument with a customer? A posted policy on a sign right in front of their face really helps.
At least, that's what the cops concluded when a friend of mine called them from a towing yard after they refused to take $181 in loose pennies as payment to get his car out. After calling it in, the cop basically told the towing guy that since he didn't have a sign, if he didn't take the pennies my friend would own the place after sueing.
The cop was wrong. Remember that police officers generally don't have a broad understanding of the law. This is the job of the courts. Your friend got lucky, because if he had tried to bring a lawusit against the towing yard, he would have certainly lost.
Slashdot Mirror
Wireless services (for phone, network, etc) and even local or co-gen electric service is popular for another, simpler reason also: People will dig up copper wire and steal it to sell. Wireless kind of prevents that from happening.
No the cubic mapping problem can be seen as an uneven distribution of image resolution -- ie when you are looking directly at a cube face you have less resolution than when you are looking into a cube corner. In practice it really just reduces sharpness.
I think the halflife image has the advantage of having been saved with a lower FOV. The aspect ratio is also a little better which also mitigates the effect somewhat. If you zoom into my kite pano quite a bit you can get the building corners and whatnot to square up
On second thought, I think I might know what you are referring to... It's not a stitching or composition problem, but it boils down to this: When you pan around a QTVR the high default field of view can make things look out of perspective.. IE 90 degree corners look sharper on the edge vs in the center of the picture. If you click the "+" to zoom in you can sort of make this problem go away at the expense of a reduced field of view.
The other minor problem is the way QTVR maps panoramas. QTVR uses a cubic map which essentially draws the panorama onto the inside faces of a cube. Sphere mapping looks marginally better and has finer detail, but remember QTVR was designed back in the days of the 68k macs and thus speed was of greater importance than consistent image resolution.
As an aside most FPS games set the FOV a little higher than it ought to be for a 3:4 display, so maybe these scenes look normal because you are used to seeing them with a high FOV in the games... Or maybe the default FOV for the QTVR is something more appropriate than QTVR's default (like 120 degrees or so)
I don't know what you're talking about. With a proper pano head and some decent stitching effort you can create 'perfect' panoramas without distortion. It helps to have the right equipment and knowledge though. Most panoramas out on the web are unfortunately shot on a regular tripod (which does not pivot around the right point) or worse, handheld, and stiched using some low quality automatic stitching software (Ulead 360, etc)...
:) My shot was inspired by Scott Haefner's work on the same subject; however he has far more experience at it than I do (I have only made one shot).. If you really want to see some awesome stuff, check out his site.
Using something that actually *works* like panotools and good wide angle lenses, you can whip up good looking shots with minimal effort.
Here's one of my first 360's made with two shots and a lens with a 183 degree FOV. There are some problems with the horizon due to it being in the extremeties of the picture, but aside from the lack of resolution, it is pretty much seamless except for one small area by the sailboat where I simply could not get panotools to bend things right. If I'd have had more altitude with the kite, it would have been perfect. Next time I shoot one, it will be
What linux distribution installs perl, and not python, by default?
Just answering your question here; not trying to suggest one thing over another, but Debian and most Debian-based distributions do a minimal install of perl (perl-base) with the most basic core install while python is out in the cold. Lots of packages rely on perl to run pre/post install scripts. I am not sure if languages other than sh and perl are actually allowed for this task, so this fact may simply be an artifact of the packaging requirements/guidelines.
Aside from the perl binary and some very basic utility packages, there's not much of perl inside of perl-base, but it is sufficient to run a lot of self-contained perl scripts. Full fledged installations of neither perl nor python are part of the 'default' install though, but if you select certain tasks usually one or both will be installed as part of a dependency.
I will take this opportunity to pimp my own cabinet, however I should also note that asking slashdot is a poor substitute for visiting and exploring the resources at Arcade Controls. The maintainer of the site also has compiled much of the knowledge and experience collected there into a nice book. Do yourself a favor and pick up a copy. If there's something that's not covered, ask in the forums there. It's a far more appropriate place.
Depending on the make/model of car, how the keys are cut/programmed, etc. it's entirely possible that you would have to pay loads of money to get the keys replaced.
On new BMW's for instance, there are several (I think I was told 7) sets of keys cut and programmed when the locksets and ignition were installed into the car. If you lose the keys, you can order your replacement keys while they still exist. After you run out of keys, you have to replace all the locks and the ignition in the car and have the ECU reprogrammed for your new key - This unfortunately costs a heck of a lot of money, but I guess if you are stupid enough to lose your key that many times, you'd probably need to change your locks anyway.
I never said boot speed is the most important metric for anything. I use linux all the time. I rarely care about the boot process, except for instance in situations like my car pc where boot time is absolutely 100% the most critical performance metric of the entire software system.
...
The most important thing in a boot process is a standard dependancy scheme. Windows has none that works, even in WinXP, and I think not even in Win2003, but not sure. We are in 2005 for christ sake, and MS is still unable to produce a boot process that works, and yes they try, it just still does not work right.
How do you figure? I am not sure I follow your reasoning here since you provide no evidence of anything that you mean by "standard dependency scheme". I don't claim to be an expert, but booting is a fairly straightforward process on any X86 sytem (and most other systems too for that matter)... Bootstrap loads kernel, kernel inits hardware, kernel loads drivers, kernel starts userspace. Userspace starts processes necessary for low level system stuff. Userspace starts user-interactive. I don't think that either Linux or Windows provides a 100% fault tolerant boot process.
It's not all the fault of the software though; I have a couple of servers (Tyan dual athlon systems) that take longer to POST than they take to boot any OS.
simpleinit-msb that I use do all of that fast since several years, and it just works. I could also move X start in an earlier runlevel (I did once), but it just has no sense other than prove that you can start a Linux GUI fast too
This is the simpleinit that is deprecated by LSBInit - a project that exists only on paper, right? I am familiar with many sysV replacements; however none of them come close to what a modern PC ought to be capable of. The closest project that I know and have used is LinuxBIOS + hand written busybox-based init.. About two seconds from the power button to a directfb UI with full hardware init (except for the IDE subsystem that takes about 5s to come up, but this was handled in the background along with a lot of other kernel module loading for drivers unessential in the first few seconds of system operation such as sound, usb, etc.
Fortunately, at least for windows users, due to the sheer number of times they have to reboot their damn machines, the boot process is fairly streamlined. XP boots on most of my hardware from power on to being logged in and fairly usable in about 30 seconds.
Initially, I thought they had come a long way in this regard, which I suppose may be true when you consider all the junk that is loading up, but what is suprising is that if you boot windows 95 on a modern pc, you can be up to the destkop in about 5 seconds after the BIOS is done.
Love it or hate it, Microsoft deserves at least a little credit for getting the boot process done relatively quickly. I can't wait to see a nice implementation of Redhat's early login efforts. Linuxbios and many other embedded machines have fantastic boot processes that really show what a modern computer can be capable of. My ceiva picture frame (hacked to run linux) boots inside of 5 seconds, and that includes the 1s timeout where it does nothing except allow the user to interrupt the boot.
You do of course remember the time when you had to reboot to change your IP address, don't you? I don't believe MS ever wholeheartedly claimed that you would never have to reboot when installing drivers and whatnot, but you certainly don't have to reboot nearly as much. There's very few modern apps that make you reboot after an install and drivers that do not require rebooting are not uncommon. I don't expect this to change particularly much with Longhorn either.. Certain changes are going to make you reboot. This is because (just like microsoft claims) a lot of companies produce crappy drivers written by lazy programmers that do not want to be bothered about having to write code that can properly init the hardware from any state.
Now, if MS made it a requirement of the windows driver certification that drivers load and unload cleanly without requiring a reboot, then we might make some progress on this front. A clean and better way to replace DLL's and other shared code libs without requiring a reboot to ensure that no other program is locking them open would also help a great deal.
Are you implying that you have to reboot Linux in order to install the video driver? You certainly don't but then again to all the "but you don't have to rebootpeople -- you do have to restart X, which is something of a pain if you don't have a good session manager. To the desktop user, a crashed X is just as destructive as a crashed kernel, and likewise a restart of X is just as interruptive as having to reboot.
I'd imagine that some code to 'ssupend/resume' the state of X might be a pretty neat project to undertake, but I'm not sure anyone has done it yet..
There is a better solution to your problem.
How about you set your mailbox filter to rely on a header rather than a subject tag? If using spamassassin, filter on "X-Spam-Status: yes" rather than whatever markings happen to be in the subject line. The forward (depending on the mail client) ought not to contain this same header.
This is also good practice to use on mailing lists too. Mailman and the like generally include X-Been-There headers. Filtering on this header instead of the subject line has all kinds of benefits such as personal responses to your postings on the list do not get stuffed into the list's mailbox, etc.
This is not particularly news. Some "blogger" discovers something because he never bothered to ask and screams something about the sky is falling.. I'm kind of sick of this "news" reporting. Incidentally, this same issue affects owners of most plasma and LCD tv's with native resolutions below 1920x1080 too.. depending on how you look at it as a problem or not.
Anyway, it's fairly well known that the internal scalers in many devices suck. That is why there is a market for good external scalers. If you are paranoid about watching a lot of 1080i on your 720p projector or LCD TV or Plasma, go buy a scaler. They cost about $1000 but will improve scaled display a lot.
At least if you have an external scaler you will have some options about how you convert 1080i to 720p. The article makes it sound like splitting the fields is a huge sin -- and it is if you discard one field per frame (Half field deinterlacing), but it's perfectly acceptible to scale EACH 540-line field to a seperate 720-line frame and double the framerate. This is called bob deinterlacing and is often the best for converting 1080i video to lower resolutions. If you are watching a 1080i upconvert of a film or something, though, you can have the scaler do full field deinterlacing and inverse telecine for you and see a nice 720p/24fps picture. Scalers also generally have internal audio delays for various types of audio feeds so you won't have to worry about AV sync issues either.
If you have any questions about how your device does this, you should try to find out before you buy it. Most devices don't publish how they do it, though, so your only option may be to derive it -- and that will not be an easy job.
Except for the fact that many retailers have "accidentally" shipped Tiger preorders early :)
That mkdir and rmdir use more than two letters is a long-standing bug--longer even than the screwy footnote/gap bug in words (which dates to Mac Word 1.0)
A bug which was fixed by DOS!
Well, Keynote reportedly uses the API; but it's not as if it couldn't do transitions another way - ie without using it. That being said, it's possible that the framework code implementing this is simply duplicated inside of the keynote application and implemented as a framework for the simple reason that several OS components need to use it as well.
There are a lot of applications where UMA or NUMA SMP systems beat the pants off of clusters; obviously these are applications that, while suited to parallel processing, do no good when you spend all of your time synchronizing the data set. Maybe there's not that many of them in the commercial market (ie general purpose applications) but there certainly is room for these systems.
There is another reason to have large n-way systems as well. Strange as it may sound, there is still some very expensive software out there that is licensed 'per server' and that sometimes means everything from a single cpu up to however many are managed by the same kernel. It can be way less expensive in certain cases to license one copy of a piece of software for a single 8-way machine than 8 copies for a cluster of 8 machines, even though the latter might even give better performance.
That is like saying people who plunk down a few hundred thousand dollars for a really nice (or at least expensive) car or boat will not tune it. $1600 is dirt cheap for a 4-way configuration; I expect that the overclockers will give it a lot more attention than they give current quad-cpu offerings.
Well, to be fair, there are 8 buttons per player not counting coin and start. N64 emulation requires lots of buttons, unfortunately, and it was one of my design goals. It looks sort of frankenstein, but it's actually very utilitarian and comfortable to use.
Anyway, I'm planning to use an LED driver board that will illuminate the buttons a certain game uses to uncomplicate things.
In retrospect, the N64 requirement would have been better served by some adaptoids and real N64 controllers. In that setup, I would have picked 7 buttons for p1 and p2 to get a combo layout between fighter and neogeo and only 3 or 4 buttons for p3 and p4.
I am actually planning on a small redesign of the upper deck. I never use the vanguard style buttons on the sides, and I need more function buttons. The top stick will probably get replaced with an analog stick using Ultimarc's new A-Pac and I'll be adding better controls up there then.
Also, I probably should have done the coin buttons by hooking up microswitches to the coin door rejects instead of having another button on the cpanel. Ah well, next time!
This, from a a person by the name of 0x461FAB0BD7D2?
The new 9xxx series 3ware cards are much better than the old series and support command queueing and all that jazz that you get from a native SATA implementation. They're not even much more expensive either. The driver has finally made it into 2.6 too, but you gotta patch it into 2.4. You can get really really good performance on them for the cost - I run one with 4 WD Raptor 10K RPM 72GB drives and don't get any of the same kind of slowdowns I used to get on 3ware's older PATA cards, especially during heavy writes.
:)
Thanks for the heads up on Areca cards. I'll keep that filed under 'check out for next time'
The ol' software raid 1 boot trick depends highly on the behavior of your BIOS under a failed drive condition. This is not the same thing as you get when you unplug a drive. Some BIOS may boot fine; some may not boot from the 2nd hard drive if the first is still attached and failing. It may also depend on how your drive has failed. If the drive electronics are failed and shorting the wrong pins on your IDE controller, then you may not get past the drive detection code in the BIOS at all.
This is simply one advantage to using a real hardware raid card like the 3ware. There are plenty of other reasons too: Does your chipset/hardware support hot swapping? If you use SATA, does it support command queueing? Do your drives? How much cache does it have? Does it have cache? Can it tolerate all types of hardware failure? Does it have *ahem* 16 ports with individual controllers for each drive? It's not like the BIOS/IDE chipset makers write out in their specs how their hardware performs under drive failure conditions so you have the overhead of testing each configuration to make sure it works proeprly before you have to rely on it. It's not so much a performance difference between hardware and software raid (until RAID-5 anyway) but an issue with how the hardware will respond when something goes wrong, which is one of the primary reasons for using anything above RAID-0 anyway.
Yes, running a 3ware card costs more. There are times when that $400 costs a lot less than the time spent configuring and testing an alternative software-only implementation. There are times when it doesn't and spending another $400 doesn't make a lot of sense. I have run both setups. I have machines deployed with both IDE software-only RAID arrays, IDE 3ware arrays, SCSI software RAID5's, SCSI Adaptec RAID's etc.. it's all application specific. There's no reason to call somebody a goon for recommending 3ware hardware. It's really good hardware; maybe you should try it sometime.
I don't know why stating facts has to be questioned. If I told you that you must stop at a stop sign, would you ask for my source? Perhaps you are the orignal poster and are simply flustered that you posted misinformation. I didn't feel the need to duplicate a lot of information here.
If you really want to know the specifics, there are plenty of other posts elsewhere in this massive tree of threads with all the links you require to validate my statement. Try this post for instance. It gives you a nice overview with a paste from the actual law, and it includes links to several reputable sources to confirm it.
Pennies are not legal tender in debts of over $20. So nickels would be fine, dimes are good too, but pennies aren't good over $20 debts.
This incorrect fact is generally passed around with some specific amount attached to it -- most often $2, $5, $10, or $20, but at least in the United States it's not true.
All currency is legal tender in any amount; however, a business can be selective about what they are willing to accept. If a merchant wanted to ONLY accept payment in pennies, it would be legal, though likely unpopular.
You can't legally refuse to accept any real denomination of money in payment of a debt unless you establish it as a policy and post a notice to that effect in advance.
Yes; you can actually. The law is very specific about this, actually. It specifically is designed to prevent just the type of abuse you describe, such as your clever friend who likely spent a lot more time acquiring the pennies than the wrecking yard spent counting them.
That's why you see all those signs about not accepting bills over $50, etc... Without the sign, they can't use the denomination as an excuse not to take the money.
Ever tried to win an argument with a customer? A posted policy on a sign right in front of their face really helps.
At least, that's what the cops concluded when a friend of mine called them from a towing yard after they refused to take $181 in loose pennies as payment to get his car out. After calling it in, the cop basically told the towing guy that since he didn't have a sign, if he didn't take the pennies my friend would own the place after sueing.
The cop was wrong. Remember that police officers generally don't have a broad understanding of the law. This is the job of the courts. Your friend got lucky, because if he had tried to bring a lawusit against the towing yard, he would have certainly lost.