It may be a closed hardware architecture, but having the underpinnings of the OS being based upon FreeBSD does not translate into a closed software platform.
Itdoesn't matter if the underpinnings are FreeBSD -- all the candy on top isn't. You won't be able to run apps written for OSX on an operating system other than one Apple provides.
If you wish to write about the closed hardware on the Apple side, I suggest you take a hard analysis of Microsoft's Palladium (sic), ie "Trusted Computing Platform."
As long as you've got multiple companies able to provide the hardware, you're better off than you are with Apple products. You may not like TCP, but when anyone can make TCP hardware, it isn't closed.
Here's another example. In terms of DRM layered music files, which would you prefer? AAC (developed by Dolby) or WMA (Microsoft)?
Lesse, AAC only works on Apple products, and they refuse to license it to 3rd parties.
WMA is liscened to 3rd parties and works on non MS products.
Hmmm, so I've got a choice between something that only works with products made by one company, or products made by multiple companies... gee, which one should I choose. WMA would be the better choice.
"Trust" isn't entering the equation here. It's all about choice. That's the whole point of the Antitrust proceedings -- "choice".
Your solution to fix an apparent lack of choice is to replace it with something that gives you even less choice.
How does web browsing help MS maintain their monopoly? How does IE keep people using Windows over MacOS or Linux? How does it lock people into Windows?
How does WMP help MS maintain their monopoly? How does WMP keep people using Windows over MacOS or Linux? How does it lock people into Windows (recall that MS does liscense their video codecs/algorithms)?
You make it sound like Europeans buying Macs is a bad thing
So you'd rather go from an open hardware platform that runs (mostly) closed software, to a closed hardware platform controlled by one company that runs (mostly) closed software?
Yeah, they SAY they're going to be a miniseries based on the first two books. But after they do the first book, they'll cancel the second half citing poor ratings and high production costs...
Actually, if you define monopoly as narrowly as it was construed for Microsoft (Microsoft was held to have a monopoly in the x86 compatible computer desktop os market), then Apple does indeed have a monopoly (in the PPC computer desktop os market).
If Microsoft were to make WMP a plugable component that would not break when another player was plugged in in it's place it would allow vendors and consumers to choose the best player.
WMP was written just like IE. It's a crappy little shell that uses a subsystem built into windows. The media subsystem is an entirely pluggable system.
Real is quite capable of writing their own codecs and using that subsystem. It isn't hard, as can be demonstrated by the vast number of codecs and non MS media player software. But they don't want to -- they'd rather write a whole player from scratch and force you to install their adware ridden pile of shit to watch videos in their format.
This whole thing is just absurd. Windows has shipped with a media player since the Windows 3.x days. What's next, they're going to insist that MS remove their TCP/IP stack so that Winsock can come back?
Microsoft cares nothing for the consumer so they make WMP such a pain to remove that most people give up trying, effectively removing their right to choose.
Funny, it "removes" just fine via add-remove programs. Oh, you want all of the subsystems that other software depends on to be removed too? Just delete random files from the system folder for the same effect.
Not only was there never a "DX" pentium, but the Pentium was never clocked as low as 25mhz. It debuted at 60mhz.
Additionally, both 25mhz & 33mhz versions of the 486 ran at the same speed as the system bus. The DX2 and DX4 ran at 2x & 4x multiples of the system bus.
I suspect you are comparing performance between the DX version of the processor between the SX and the SL versions.
Apparently most users are just downloading and executing attachments without even thinking
I'm beginning to suspect that most users (ones at work anyway) do it so they have an excuse to slack off while someone fixes their computer for them...
Imagine if e-mail was just plain old ASCII text with no attachment support. *sigh*
All email is plain old ASCII. All of the attachments are encoded in an ASCII form, and decoded by your mail reader. Before mail software supported that operation implicitely, there were tools out there to generate the mime encoding to paste into the email, and there were tools that would generate the files in the encoding... this latest batch proves that users will go to great lengths to run attachments (save zip file to disk, open in zip software, type in password sent in email, extract files to a folder, run file in folder...), so I have no doubt that they would go through the pains to extrace mime content via some set of tools referred to by the email...
You can write extra memory around the stack to detect overruns, however it still depends on software to do the trick (and being software controlled, it can be bypassed).
The study wasn't a test of different systems, it was a survey of compromised sites (where the site was big/important enough for the owning company to instigate an investigation).
These weren't desktop machines. And they weren't server boxes being run out of some guys bedroom off of a cablemodem. These were servers used for "business" purposes.
A decision was made somewhere along the line, for better or worse, that a long should remain a 32bit value. Most likely because that's the "preferred"/optimal value size the AMD chips prefer to operate on.
This wasn't a breaking change. The existing code was already broken. However, the enviornment hid that from you due to the lucky coincidence that the size of a pointer happened to match the preferred size of integer value for the particular processors you worked on.
A breaking change would be a guarantee that sizeof(long)==sizeof(void*), and then removing that guarantee.
Personally, I have no sympathy for people with this problem. If I ever ran into such an assignment on code I worked on I'd slap 'em upside the head...
The only safe assuptions about size that you can make about types in C/C++ is that a char = short = int = long. A void* is guaranteed to be large enough to hold any pointer.
The only safe assuptions about size that you can make about types in C/C++ is that a char = short = int = long. A void* is guaranteed to be large enough to hold any pointer.
There is no direct relationship between a long and a void*; the fact that they're the same on one 32bit system doesn't mean jack. The fact that some header file in your dev enviornment defines them to be the same doesn't mean jack. The language does not provide you with that guarantee.
If you assume a long is equal in size to a void*, your code is making an invalid assumption. Why the hell you'd want to shove a pointer into a signed variable is beyond me in the first place...
If your code depends on certain variables being a certain number of bits wide (at least on an MS platform), use the __int8, __int16, __int32, and __int64 types.
Disney does not have a poison pill. Once the story broke, a lot of articles remarked about how Disney was an easy target for a takeover for that very reason.
Slashdot Mirror
It may be a closed hardware architecture, but having the underpinnings of the OS being based upon FreeBSD does not translate into a closed software platform.
... gee, which one should I choose. WMA would be the better choice.
Itdoesn't matter if the underpinnings are FreeBSD -- all the candy on top isn't. You won't be able to run apps written for OSX on an operating system other than one Apple provides.
If you wish to write about the closed hardware on the Apple side, I suggest you take a hard analysis of Microsoft's Palladium (sic), ie "Trusted Computing Platform."
As long as you've got multiple companies able to provide the hardware, you're better off than you are with Apple products. You may not like TCP, but when anyone can make TCP hardware, it isn't closed.
Here's another example. In terms of DRM layered music files, which would you prefer? AAC (developed by Dolby) or WMA (Microsoft)?
Lesse, AAC only works on Apple products, and they refuse to license it to 3rd parties.
WMA is liscened to 3rd parties and works on non MS products.
Hmmm, so I've got a choice between something that only works with products made by one company, or products made by multiple companies
"Trust" isn't entering the equation here. It's all about choice. That's the whole point of the Antitrust proceedings -- "choice".
Your solution to fix an apparent lack of choice is to replace it with something that gives you even less choice.
How does web browsing help MS maintain their monopoly? How does IE keep people using Windows over MacOS or Linux? How does it lock people into Windows?
How does WMP help MS maintain their monopoly? How does WMP keep people using Windows over MacOS or Linux? How does it lock people into Windows (recall that MS does liscense their video codecs/algorithms)?
You make it sound like Europeans buying Macs is a bad thing
So you'd rather go from an open hardware platform that runs (mostly) closed software, to a closed hardware platform controlled by one company that runs (mostly) closed software?
Ah yes, back to when you had to download and install software for 3 days before you had a usable system...
That was better HOW?
What would be the point (aka: their motivation) for them to do it? They'd never recoupe their costs.
Yeah, they SAY they're going to be a miniseries based on the first two books. But after they do the first book, they'll cancel the second half citing poor ratings and high production costs...
Actually, if you define monopoly as narrowly as it was construed for Microsoft (Microsoft was held to have a monopoly in the x86 compatible computer desktop os market), then Apple does indeed have a monopoly (in the PPC computer desktop os market).
If Microsoft were to make WMP a plugable component that would not break when another player was plugged in in it's place it would allow vendors and consumers to choose the best player.
WMP was written just like IE. It's a crappy little shell that uses a subsystem built into windows. The media subsystem is an entirely pluggable system.
Real is quite capable of writing their own codecs and using that subsystem. It isn't hard, as can be demonstrated by the vast number of codecs and non MS media player software. But they don't want to -- they'd rather write a whole player from scratch and force you to install their adware ridden pile of shit to watch videos in their format.
This whole thing is just absurd. Windows has shipped with a media player since the Windows 3.x days. What's next, they're going to insist that MS remove their TCP/IP stack so that Winsock can come back?
Microsoft cares nothing for the consumer so they make WMP such a pain to remove that most people give up trying, effectively removing their right to choose.
Funny, it "removes" just fine via add-remove programs. Oh, you want all of the subsystems that other software depends on to be removed too? Just delete random files from the system folder for the same effect.
The unintuitively named DX4 actually ran at 3x the system bus, not 4x.
...
You're right, I'd completely forgotten about that
Methinks you meant a 486 DX, not a DX Pentium.
Not only was there never a "DX" pentium, but the Pentium was never clocked as low as 25mhz. It debuted at 60mhz.
Additionally, both 25mhz & 33mhz versions of the 486 ran at the same speed as the system bus. The DX2 and DX4 ran at 2x & 4x multiples of the system bus.
I suspect you are comparing performance between the DX version of the processor between the SX and the SL versions.
...unless the product is installed with a key obtained via one of their volume liscensing programs (product activation isn't required for those keys).
You forgot one:
3. They get to use it as a tax writeoff.
Track Changes is off by default.
Apparently most users are just downloading and executing attachments without even thinking
... this latest batch proves that users will go to great lengths to run attachments (save zip file to disk, open in zip software, type in password sent in email, extract files to a folder, run file in folder...), so I have no doubt that they would go through the pains to extrace mime content via some set of tools referred to by the email...
I'm beginning to suspect that most users (ones at work anyway) do it so they have an excuse to slack off while someone fixes their computer for them...
Imagine if e-mail was just plain old ASCII text with no attachment support. *sigh*
All email is plain old ASCII. All of the attachments are encoded in an ASCII form, and decoded by your mail reader. Before mail software supported that operation implicitely, there were tools out there to generate the mime encoding to paste into the email, and there were tools that would generate the files in the encoding
Gabe should threaten to sue Infinium Labs for damaging their reputation unless they remove the post in question and print a retraction...
(yes, I know that the post was deleted, but it still would have been funny...).
A PPC port had already been done. NT4 would run on PPC architectures.
Except when you're playing games licensed from Atari, in which case they're required to draw a number of bomb icons on the screen...
There is no such thing as a "convicted monopolist." Microsoft was found to be in violation of anti-trust laws. It is not a crime to have a monopoly.
Interesting side note:
Microsoft was held to have a monopoly in the operating system market for x86 compatible computers only.
Using such a narrow definition of monopoly, I would submit that Apple has a monopoly in the operating system market for PPC computers.
You can write extra memory around the stack to detect overruns, however it still depends on software to do the trick (and being software controlled, it can be bypassed).
Linux and BSD run on non x86 processor architectures which have this ability.
The study wasn't a test of different systems, it was a survey of compromised sites (where the site was big/important enough for the owning company to instigate an investigation).
These weren't desktop machines. And they weren't server boxes being run out of some guys bedroom off of a cablemodem. These were servers used for "business" purposes.
A decision was made somewhere along the line, for better or worse, that a long should remain a 32bit value. Most likely because that's the "preferred"/optimal value size the AMD chips prefer to operate on.
This wasn't a breaking change. The existing code was already broken. However, the enviornment hid that from you due to the lucky coincidence that the size of a pointer happened to match the preferred size of integer value for the particular processors you worked on.
A breaking change would be a guarantee that sizeof(long)==sizeof(void*), and then removing that guarantee.
Personally, I have no sympathy for people with this problem. If I ever ran into such an assignment on code I worked on I'd slap 'em upside the head...
The only safe assuptions about size that you can make about types in C/C++ is that a char = short = int = long. A void* is guaranteed to be large enough to hold any pointer.
...
Thank you slashdot for muching the < signs....
Let's try that again
char <= short <= int <= long
The only safe assuptions about size that you can make about types in C/C++ is that a char = short = int = long. A void* is guaranteed to be large enough to hold any pointer.
...
There is no direct relationship between a long and a void*; the fact that they're the same on one 32bit system doesn't mean jack. The fact that some header file in your dev enviornment defines them to be the same doesn't mean jack. The language does not provide you with that guarantee.
If you assume a long is equal in size to a void*, your code is making an invalid assumption. Why the hell you'd want to shove a pointer into a signed variable is beyond me in the first place
If your code depends on certain variables being a certain number of bits wide (at least on an MS platform), use the __int8, __int16, __int32, and __int64 types.
Disney does not have a poison pill. Once the story broke, a lot of articles remarked about how Disney was an easy target for a takeover for that very reason.