I don't know about Dell and IBM, but Apple never stopped selling 12-inch laptops.... The 12-inch iBook has been in continuous production since May, 2001.
This is a common misconception. The reason you need a Mac version of the firmware is usually because the card vendor built detection routines into the driver to prevent it from working with the cheaper PC version of the card, not because the firmware itself is anything special (though it may add extra capabilities).
Open Firmware drivers (usually in the form of a Mac-specific firmware revision) are generally only needed for:
drive interface cards (ATA, SCSI) if you are booting from them.
video cards if you want Open Firmware to be able to use them at boot time (command-option-o-f).
ethernet cards if you want to netboot your machine.
anything else you want to use as a boot device unless Open Firmware has a built-in driver for it.
Basically if you want the device to be able to be controlled by the boot firmware (which isn't necessary to use the device once the OS has actually loaded), you need an OF driver. Otherwise, you don't.
See, that's the thing. This is only a problem for artists because they generally have zero knowledge of the legal systemâ"less than even an average personâ"and because they're scared highâ"school-aged children who don't want to be sued.
The reality of the matter is that letters of intent are binding if and only if all material terms of the agreement were worked out prior to the signing of the letter of intent. There isn't a state in the U.S. where "let's make a deal" written on a napkin would hold up in court, and the record companies know this.
The handful that engage in such unscrupulous actions do so under the assumption that the majority of people who they tak to aren't smart enough to realize that such a letter of intent isn't worth the price of the napkin it's written on, and are so desperate to sign with someone that they'll take a bad deal just to get signed.
Quite frankly, it's just social darwinismâ"the more intelligent preying upon the unbelievably stupid. The band could simply ignore the letter of intent, since it isn't binding because material terms were not agreed upon prior to its signing. The record company knows they'd be laughed out of court, so if they threatened to sue the band, the problems could be solved by hiring a lawyer to send the record company a simple cease & desist letter.
And before you say, "Oh, but it -is- legally binding", go look up the law yourself.
I was actually wondering how they deal with the fact that a heck of a lot of some companies' wireless hardware ends up shipping with someone else's name on it. Is that counted as being sold by company A, company B, or both? Could be fun if the answer were both, as it could explain a good part of the 120% increase.:-)
Yeah, and the new optical mouse is too wide and flat on top for my hand. It puts pressure on the two pads of my palm, and basically sits right on top of my medial nerve (see also CTS) and pinches it, resulting in significant discomfort. I've gone back to an ADB Mouse II (the curvy one) and haven't had wrist problems since.
Since I've already talked to at least a dozen people who bought legit copies of their MP3 collection via the iTunes Music Store, and only about two who didn't, I'm gonna have to go with "yes", but only if the price is more reasonable than it was when they chose to steal the car originally.
P2P is not the problem. The problem is the price-value ratio. Period.
Whoa. Deja vü..... That would be true if they actually serviced the lines. Where I used to live, it was nearly impossible to get a line repaired. Three years and lots of phone calls later, they finally fixed the problem after it rained and we were able to hear our next door neighbor's conversations clearly.
Give me a break. If the same phone companies provided the same service to rural customers that they do for urban customers, it would be more expensive. Of course, if pigs could fly, we'd all carry umbrellas.
You are correct that it is possible for an application to crash the OS even in a well-written protected-memory OS. It's called a bug, and every now and then it happens. You're also absolutely right that there is little the OS can do to protect an application from itself.:-)
That having been said, here's a list of some of the more egregious errors in your post:
1. Linux does not use a microkernel (unless it's MkLinux).
2. Regarding the Win32/NT kernel separation, the BSD personality layer behaves similarly (assuming you described NT reasonably well), communicating with Mach (the OS kernel, as you put it) using client/server communication (Mach messaging).
3. No, I don't have to admire anything about the NT kernel. I run in terror at the very thought. Until 2000, it was barely usable, far less so than even the public beta of Mac OS X.
We now return you to your regularly scheduled programming.
25 Mbps is as fast as you -can- acquire it. That's the absolute maximum data rate for consumer DV, whether through firewire or any other means.
And yes, I -do- occasionally watch DV footage over an HTTP connection over 100Base-T ethernet at work. To say that it is fast enough is like saying Pavarotti is a singer.:-)
True, but I think if you analyze the heat output carefully, you'll find that the power supply is doing most of that, and that the CPU itself is putting out less than a low-output incandescent light bulb.:-)
If by "makes", you mean "fabricates", then you are correct. However:
I'm pretty sure there is Apple IP in every PowerPC processor ever built.
Apple was one of the two companies that founded Advanced RISC Machines (ARMâ"there's a variant of the Apple/Acorn architecture in your iPod, your Compaq iPaq, and probably the next Palm you buy).
Just thought that should be cleared up.... As for whether Apple optimizes the processor design to fit its software, according to the announcements at the time, the G3 was specifically designed to optimize classic Mac OS performance. So, yeah.
I'll agree with you on most of your points, but I have to take issue with your suggestion that Airport Extreme uses CompactPCI.
CompactPCI is a standard designed to replace VME in telecomm environments. A CompactPCI card is roughly 14 inches square if memory serves, which large connectors along one edge that are almost as wide as an iBook is thick.
You're right that PCMCIA isn't fast enough for 802.11g, but nobody (PC or Mac) has used pure PCMCIA for many, many years. Instead, they use CardBus, which is basically a removable PCI standard, not entirely unlike CompactPCI, but at least an order of magnitude smaller.
The Airport Extreme cards, AFAIK, are either CardBus or PCI, though I'm not sure which.
That's pretty amazing, since DV is only 25 Mbits/second (unless you're doing DV50, in which case it's 50. You must be sucking down an awful lot of DV streams at once....
Please, please, tell me this was just a store that sold Mac hardware, and that it wan't an actual Apple retail store. If it was an actual Apple retail store, please tell me which one so I can pass the complaint on to the right people....:-|
I've had a lot of luck polishing the CDs with toothpaste. I use a finger to vigorously rub the surfaceâ"not a toothbrush, as that could make the scarring worseâ"and rub mainly perpendicular to the groove. By reducing the sharp angles, you reduce the amount of light that gets deflected at weird angles and make it more likely that a player will be able to correct the remaining errors with the usual CRC mechanism. Your mileage may vary.
Which is 100% unenforceable. By law, an NDA becomes null and void if the material it covers becomes public, regardless of how it becomes public or when it became public... unless you made it public while under the NDA. If you have distributed the code previously, the NDA cannot cover it, regardless of language to the contrary.
In effect, that means that the NDA is probably not worth the paper it is printed on, assuming that we're all correct that the code in qustion was just stuff AT&T stole from Berkeley back in the 80s.... That having been said, I wouldn't want to test that theory by signing the NDA and then publishing their source code....:-)
Since this is about education, and being a (largely online) writing instructor this quarter, any comment I make on this subject is thus on-topic, I thought I'd give you a good laugh while pointing out that you're mostly but not entirely completely not right.:-)
You should change paragraphs when you change speakers. Quoting someone else is probably not a change of speaker, so it is probably not necessary to split it into a separate paragraph, though for clarity, it is often a good idea to do so.
For example:
"The pork is green," Tom said. "That's all right by me, though."
Note: same speaker, same paragraph.
"The pork is green," Tom said.
"That's all right by me, though," Mark replied.
Note: different speaker, different paragraph.
Stevenson et al says, "the pork from suidae seuss (sp?) may be green if the meat is cooked improperly." However, we do not feel that this is correct because we have never witnessed pork turning green as a result of culinary imcompetence.
Note: even though the speaker is quoting what was said in someone else's book, the same implied person (in this case, the writer) is making both parts of the statement, and thus it is not necessary to break it into a paragraph.
That having been said, using a quote without attribution is plagiarism. It probably should have read:
You said, "...I think some of the classes are laid out rediculously." Of course, with spelling like that, you're a good person to comment on education.
On a connection that is saturated upstream (like most DSL connections where you're running a server), you would gain the ability to saturate the downstream pipe. You'd only have to get a single response packet out at the end of the connection to request any missing pieces. The overhead saved is very significant.
As to the issue of flow control, that's precisely the reason why TCP/IP is terrible for most peope with a slow link. All those exponential back-offs mean that the second there's a dropped packet (PPP or modem re-handshaking, saturated link with too much latency or packet dropping, traffic shaping routers in schools, etc.), your throughput goes to hell in a handbasket.
Basically, TCP/IP was never designed with DSL in mind, much less standard PPP. At even 1-3% packet loss (and I would see consistent 1% packet loss for hours on end with my prior DSL provider), your performance is worse than useless. At 20% packet loss, you might as well not have a connection at all, since you'll barely even be able to view web pages over it, much less do anything useful. Heaven help you if you plug a cable into your Ethernet hub incorrectly and have a "dual-port" shared between a device and the uplink and get 50% packet loss....
In an ideal (IMHO) scheme, the flow control would involve the two endpoints knowing their connection speed to the next hop out (since the endpoint link is the slowest link about 99.99% of the time), and the sending end agreeing not to push more data than that per second.
With a minimal amount of effort, that could scale to multiple connections. Only getting 1/3rd of the packets? Tell the sender to slow to 1/3rd its current speed. In every case, try to go right up to the edge, but not over. Since the control connection would be over TCP/IP, you'd be assured delivery of the careful backoff, even if it was delayed slightly.
The key is to ignore any inconsistent packet loss. For example, if lowering the rate doesn't fix the packet loss, or at least improve it, raise the rate back up, since no amount of lowering is going to make any difference anyway.
SCO has a -lot- to lose. If it is determined that they knowingly filed charges under false pretense, they can be charged with perjury and various other ethics violations, and their legal team can and almost certainly will be disbarred. It is very much in their best interest for this -not- to go to court, particularly if they ever want to work in any job other than burger-flipper.
Further, they can, as individuals, be held personally liable for their actions in the form of civil suits by IBM and the various Linux distributors (and maybe even Linux users and developers) for libel, slander, harassment, and so many other charges that I don't know where to begin. With enough public pressure, there's a good chance they might even be brought up on criminal charges.
Make no mistake, SCO as a company may have little to lose, but a company is made up of individuals who are ultimately responsible for their own actions, and while they cannot be held liable for SCO's debts if SCO gets sued into oblivion, they are still very much accountable in the eyes of the law for any illegal actions that led to that demise, and knowingly filing false charges is very much illegal.
Forgive the rant: -after- SCO loses in the ugliest court battle the tech industry has ever seen, I want to see every single one of their executives and lawyers fined and, where possible, sentenced to jail time in a federal pound-me-in-the-ass prison. It isn't enough just to make an example out of the company. You must also make an example of anyone who knowingly contributed to this mockery of our judicial system. -Then- you will have slowed the tide of outrageous IP claims.
You know, it might -be- possible to screw back. I don't know the details of the AT&T/BSD lawsuit settlement were, but basically, AT&T stole BSD code and used it improperly. BSD agreed to let them use it, and AT&T agreed not to sue over reverse violations that UCB committted.
If the folks at Berkeley were smart, they put a clause that allows them to retaliate if AT&T (or someone who bought rights from AT&T) ever sued anyone for using BSD code. If SCO points out "stolen" code in Linux and it turns out that even one single line of that code came from BSD, and if there was such a clause in the original agreement, UC Berkeley could then nail SCO's @$$ so the wall.
Long story short, if there weren't enough reasons already for SCO to back down, imagine the possibility of Berkeley forcing SCO UNIX to suddenly ship without a networking stack, and declaring every prior copy of SCO UNIX to be illegal.
Slashdot Mirror
Open Firmware drivers (usually in the form of a Mac-specific firmware revision) are generally only needed for:
- drive interface cards (ATA, SCSI) if you are booting from them.
- video cards if you want Open Firmware to be able to use them at boot time (command-option-o-f).
- ethernet cards if you want to netboot your machine.
- anything else you want to use as a boot device unless Open Firmware has a built-in driver for it.
Basically if you want the device to be able to be controlled by the boot firmware (which isn't necessary to use the device once the OS has actually loaded), you need an OF driver. Otherwise, you don't.The reality of the matter is that letters of intent are binding if and only if all material terms of the agreement were worked out prior to the signing of the letter of intent. There isn't a state in the U.S. where "let's make a deal" written on a napkin would hold up in court, and the record companies know this.
The handful that engage in such unscrupulous actions do so under the assumption that the majority of people who they tak to aren't smart enough to realize that such a letter of intent isn't worth the price of the napkin it's written on, and are so desperate to sign with someone that they'll take a bad deal just to get signed.
Quite frankly, it's just social darwinismâ"the more intelligent preying upon the unbelievably stupid. The band could simply ignore the letter of intent, since it isn't binding because material terms were not agreed upon prior to its signing. The record company knows they'd be laughed out of court, so if they threatened to sue the band, the problems could be solved by hiring a lawyer to send the record company a simple cease & desist letter.
And before you say, "Oh, but it -is- legally binding", go look up the law yourself.
1. Buy FrameMaker (apparently abandoned, only works in a legacy emulation mode) or
2. Buy an eMac.
I think you can see where the problem lies.
Oh well. :-p
P2P is not the problem. The problem is the price-value ratio. Period.
It should consist of the palm of a hand with the thumb, first, ring, and pinkie fingers curled against it.
Give me a break. If the same phone companies provided the same service to rural customers that they do for urban customers, it would be more expensive. Of course, if pigs could fly, we'd all carry umbrellas.
That having been said, here's a list of some of the more egregious errors in your post:
1. Linux does not use a microkernel (unless it's MkLinux).
2. Regarding the Win32/NT kernel separation, the BSD personality layer behaves similarly (assuming you described NT reasonably well), communicating with Mach (the OS kernel, as you put it) using client/server communication (Mach messaging).
3. No, I don't have to admire anything about the NT kernel. I run in terror at the very thought. Until 2000, it was barely usable, far less so than even the public beta of Mac OS X.
We now return you to your regularly scheduled programming.
And yes, I -do- occasionally watch DV footage over an HTTP connection over 100Base-T ethernet at work. To say that it is fast enough is like saying Pavarotti is a singer. :-)
- I'm pretty sure there is Apple IP in every PowerPC processor ever built.
- Apple was one of the two companies that founded Advanced RISC Machines (ARMâ"there's a variant of the Apple/Acorn architecture in your iPod, your Compaq iPaq, and probably the next Palm you buy).
Just thought that should be cleared up.... As for whether Apple optimizes the processor design to fit its software, according to the announcements at the time, the G3 was specifically designed to optimize classic Mac OS performance. So, yeah.CompactPCI is a standard designed to replace VME in telecomm environments. A CompactPCI card is roughly 14 inches square if memory serves, which large connectors along one edge that are almost as wide as an iBook is thick.
You're right that PCMCIA isn't fast enough for 802.11g, but nobody (PC or Mac) has used pure PCMCIA for many, many years. Instead, they use CardBus, which is basically a removable PCI standard, not entirely unlike CompactPCI, but at least an order of magnitude smaller.
The Airport Extreme cards, AFAIK, are either CardBus or PCI, though I'm not sure which.
In effect, that means that the NDA is probably not worth the paper it is printed on, assuming that we're all correct that the code in qustion was just stuff AT&T stole from Berkeley back in the 80s.... That having been said, I wouldn't want to test that theory by signing the NDA and then publishing their source code.... :-)
You should change paragraphs when you change speakers. Quoting someone else is probably not a change of speaker, so it is probably not necessary to split it into a separate paragraph, though for clarity, it is often a good idea to do so.
For example:
"The pork is green," Tom said. "That's all right by me, though."
Note: same speaker, same paragraph.
"The pork is green," Tom said.
"That's all right by me, though," Mark replied.
Note: different speaker, different paragraph.
Stevenson et al says, "the pork from suidae seuss (sp?) may be green if the meat is cooked improperly." However, we do not feel that this is correct because we have never witnessed pork turning green as a result of culinary imcompetence.
Note: even though the speaker is quoting what was said in someone else's book, the same implied person (in this case, the writer) is making both parts of the statement, and thus it is not necessary to break it into a paragraph. That having been said, using a quote without attribution is plagiarism. It probably should have read:
You said, "...I think some of the classes are laid out rediculously." Of course, with spelling like that, you're a good person to comment on education.
Good call on the comma after "yeah", though.
As to the issue of flow control, that's precisely the reason why TCP/IP is terrible for most peope with a slow link. All those exponential back-offs mean that the second there's a dropped packet (PPP or modem re-handshaking, saturated link with too much latency or packet dropping, traffic shaping routers in schools, etc.), your throughput goes to hell in a handbasket.
Basically, TCP/IP was never designed with DSL in mind, much less standard PPP. At even 1-3% packet loss (and I would see consistent 1% packet loss for hours on end with my prior DSL provider), your performance is worse than useless. At 20% packet loss, you might as well not have a connection at all, since you'll barely even be able to view web pages over it, much less do anything useful. Heaven help you if you plug a cable into your Ethernet hub incorrectly and have a "dual-port" shared between a device and the uplink and get 50% packet loss....
In an ideal (IMHO) scheme, the flow control would involve the two endpoints knowing their connection speed to the next hop out (since the endpoint link is the slowest link about 99.99% of the time), and the sending end agreeing not to push more data than that per second.
With a minimal amount of effort, that could scale to multiple connections. Only getting 1/3rd of the packets? Tell the sender to slow to 1/3rd its current speed. In every case, try to go right up to the edge, but not over. Since the control connection would be over TCP/IP, you'd be assured delivery of the careful backoff, even if it was delayed slightly.
The key is to ignore any inconsistent packet loss. For example, if lowering the rate doesn't fix the packet loss, or at least improve it, raise the rate back up, since no amount of lowering is going to make any difference anyway.
Further, they can, as individuals, be held personally liable for their actions in the form of civil suits by IBM and the various Linux distributors (and maybe even Linux users and developers) for libel, slander, harassment, and so many other charges that I don't know where to begin. With enough public pressure, there's a good chance they might even be brought up on criminal charges.
Make no mistake, SCO as a company may have little to lose, but a company is made up of individuals who are ultimately responsible for their own actions, and while they cannot be held liable for SCO's debts if SCO gets sued into oblivion, they are still very much accountable in the eyes of the law for any illegal actions that led to that demise, and knowingly filing false charges is very much illegal.
Forgive the rant: -after- SCO loses in the ugliest court battle the tech industry has ever seen, I want to see every single one of their executives and lawyers fined and, where possible, sentenced to jail time in a federal pound-me-in-the-ass prison. It isn't enough just to make an example out of the company. You must also make an example of anyone who knowingly contributed to this mockery of our judicial system. -Then- you will have slowed the tide of outrageous IP claims.
If the folks at Berkeley were smart, they put a clause that allows them to retaliate if AT&T (or someone who bought rights from AT&T) ever sued anyone for using BSD code. If SCO points out "stolen" code in Linux and it turns out that even one single line of that code came from BSD, and if there was such a clause in the original agreement, UC Berkeley could then nail SCO's @$$ so the wall.
Long story short, if there weren't enough reasons already for SCO to back down, imagine the possibility of Berkeley forcing SCO UNIX to suddenly ship without a networking stack, and declaring every prior copy of SCO UNIX to be illegal.
<jar-jar>SCO's-a-gonna-die?</jar-jar>