Generally, we don't hate companies for being too big, we hate companies when they start acting like asshats. For example, IBM has been well behaved recently, so most people like them.
Besides, Google has been big for a long time, big enough that they've been able to work on some projects that didn't generate any revenue for quite a while. Some of these things are starting to come together, so people are starting to notice, but their size is not a new thing.
Remember, they've been the default search engine for most of the web for years. It's impossible to do that without being big.
"Unlike some of the other posts though, I'm not worried about RAM or hard drive space, I'm worried about L1 and L2 cache space. If a sizable amount of the data cached goes from 32 to 64 bits, the processors could take a big performance hit unless the caches are bigger too."
This is a valid concern, but the hit isn't that big. IIRC, 64-bit code and data tends to be about 10-15% bigger, doubling the cache is good enough to overcome this. Looking at the cache size of major x86 processors, the Prescott has gone from 8k to 16k L1 and 512k to 1024k L2. AMD64 processors are constant (starting to move to 1m L2 on the newest ones), but they had a lot of L1 to begin with.
Have you been watching cache sizes lately? They were constant at around 256k-512 k for close to 10 years (my P2 from 1996 has the same cache as my P4 from 2004). They're just beginning to increase now with 90 nm because cache increase is the most beneficial thing to do with the transistors becoming available.
"Here's hoping programmers keep using 32 bit data structtures when they don't need 64 bits!"
This is not a new thing. How many people now use a 32-bit int when an 8-bit char would do? Pretty much everyone, in my experience. They won't bother optimizing. When they do, it's because it needs a resource that's very scarce (network traffic, or huge numbers of the data structure being used). Also consider is that there's usually a penalty to loading values from a non aligned address. I'm not saying this dominates, but using the smallest type you can is almost never free.
In the case of PowerPC (and SPARC, MIPS, anything else I can think of), the opcodes is still 32 bits. The number of registers stays the same, the number of instructions stays (more or less) the same. There is no reason to extend the size of opcodes.
Re:A few really good Apps could make the differenc
on
Firefox - The Platform
·
· Score: 1
"You didn't click on the link did you? Sure a little scripting of GMail can make it fast, but as far as rich interfaces go, I don't think it would look so good in comparison to a XUL interface. Go look at the demo, and spend some time clicking around and using all the features of it."
I did click in fact, and I don't really see much difference in user experience between gmail and the link you posted. The one you posted looks more like a native app, but to me they seem about as fast as each other,
Re:A few really good Apps could make the differenc
on
Firefox - The Platform
·
· Score: 1
"Now, imagine if GMail started doing that... IE users of GMail get the standard webmail interface, but Firefox users get a full fast XUL interface. Have a look at that demo site again, and do some clicking around... then tell me that that wouldn't be an absolutely killer app for Firefox."
Gmail does do some kind of local scripting. I don't know the specifics, but it's very fast, much faster than it would be if it had to access anything online for every click. Also, there's autocomplete boxes that pop up when you start typing in a From: box that narrow down the list with each letter you type. That and other stuff makes me sure that a lot of what it does happens locally.
For that reason, I kinda doubt XUL would be much faster.
The only time they're worth it is when you're trying to achieve a temperature below the ambient temperature. Otherwise, it's easier to put a heat spreader on the chip (as AMD and Intel already do) so the contact area with the heat sink is bigger.
These peltiers aren't going to go in any general purpose CPUs as we know them.
While I think the grandparent with his head in the clouds deserved to get yelled at for being wrong (as OS X does not rival the high end *nixen), smart aquisitions are part of business and buying up NeXT was a very good idea. Just because it was around before doesn't mean Apple didn't make very good progress in a short period of time.
"Good luck with that. Especially since it's illegal under the DMCA."
That's funny, I thought it was explicitly allowed. Granted that's just Canada... but the US is not the only country out there.
"And since future versions of Intel hardware and Microsoft software will put a hard block on your ability to do this."
pffft
That assumes that every stage of the setup secure, and that details are never leaked. Given that the whole thing is designed by committee on a deadline, and that they're going to be dealing with people that can sniff the bus, I find that unlikely. Indeed, Microsoft claiming they will be able to provide unbreakable DRM is equivilant to Microsoft claiming they can provide perfect security. For example, one of the vulnerabilities that allowed an XBox to play pirated games was in the firmware itself.
And then, at the end of the day, the best case they can hope for is forcing everyone to use the analog hole.
Compared to MacOS X, I would say that Solaris/X11/Gnome does not have advanced graphical capabilites. While Solaris/X11/Gnome does have OpenGL support, it does not (AFAIK) have anything similar to Quartz Extreme.
Is this at all related to that effect where the gravitational influence of the sun is slightly reduced at the point of totality in a solar eclipse?
I've been vaguely looking around for more information on that. From what I can gather, it's a reproducible observation, but other effects haven't been ruled out, so no one's sure what to make of it.
Yes, I believe they ship OS X Server with the Xserves. Also, if you're doing double precision floating point stuff the PPC970 is the best chip you can buy this side of Itanium.
However, those things have nothing to do with the mainframe type stuff. The whole point of clusters is to have relatively inexpensive interchangable machines. The whole point of a mainframe is to have one bulletproof system. The CPUs are hotpluggable, the filesystems can move from disk to disk without being taken offline, etc. MacOS X cannot do those things. Similarly, the mainframe OSes don't bother trying to have advanced graphical capabilities, or any of a hundred things MacOS does better than anyone.
I'm not saying Solaris or AIX is "better", only that they're better at what they're designed for.
oh, jeeze... Please, for all of our sakes, get your head out of the clouds.
MacOS tries to be a desktop OS, and it succeeds brilliantly. However, MacOS does not try to be a mainframe OS. It doesn't run on mainframes, and would't do a very good job if it could. This is what Solaris and AIX do, and they do do a good job.
The two things aren't even remotely comparable.
Re:Old school hackers vs. new school hackers.
on
Good Bad Attitude
·
· Score: 4, Insightful
There's a distinction here no one else seems to see.
I think the actual use of P2P software to upload or download files might be unethical, but the phenomenon of file sharing in general is nothing more and nothing less than the invisible hand attempting to correct the price of the media being distributed. It's no more unethical than the phenomenon of male Lions killing cubs when they take over a pride. It might be nasty when each individual does it, but the practice has helped Lions survive as a species.
The simple fact is that the market is hardwired into humans. Trade is in our nature. We may argue with actions people take, but there's very little that can be done to stop anything completely. We can no more stop file sharing than we can the drug trade or prostitution. There's demand, there's supply, the market takes care of the rest whether we like it or not.
The problem with those trying to stop it is that they're fighting human nature. Human nature won't change. It's not that the **AA is wrong (or at least, exclusively wrong), it's that their goal is not achievable. They will either continue to fight, capitulate, or compromise. My money's on compromise, but no bets on when.
The price of the media in question is higher than the market is willing to bear. I'm not saying it is or is not a fair price, only that it's higher than people are willing to pay. As a result, there were a lot of people that wanted the media but didn't want to pay for it. To the tune of several times the total volume of legitimate sales.
Napster was growing 15% a day for several months. And I've got news for you, Napster sucked. That just doesn't happen unless there's a huge unmet demand.
It's the **AA's fault for ignoring the market. While I might consider the actions of the file traders themselves to be unethical, I recognize that they're following human nature.
Safari is not a great browser. It has significant compatability problems, at least with the pages I used it with.
Camino is the way to go. Once loaded, it's as fast as Safari, and it's MUCH better at rendering correctness. It's designed for OS X, and looks like a native app. It has had a minimalist interface similar to Safari's since before Safari existed. It's easily the best browser for Mac.
"They now offer a single processor low-end G5. This lowers the price of entry into their pro-range. It helps keep up with demand, given that supply of G5 processors is an issue."
The dual-only PowerMacs were the thing keeping Apple from being a consideration for me. I'm glad I waited. I can only assume G5 availability from IBM was not good enough to start selling a lower margin machine.
I'll tell you what though, a thousand frenzied Apple fans chanting "Sure it costs more, but it's still worth it!" is about the most annoying thing I've ever seen on Slashdot. That argument is so intellectually bankrupt that I didn't even get modded down yelling at them. It was a lot more money without greater utility, because the second CPU does nothing unless you're CPU-bound. Some of us just want a computer that we can upgrade in a few years.
Now that there's a machine that does what people like me need, it can be truthfully said they have a well rounded lineup. It will be a little while for me, but I think I will recommend a single-1.8 for my parents. I have no desire to get dragged back there for tech support, no matter how good my mothers cooking is.
"I don't care nearly as much as I used to about the what the kernel does; it's so easy to emulate your way back to a familiar state."
DragonFlyBSD has a system call layer that would allow potentially very different interfaces to be presented to userspace stuff with essentially no penalty. This may allow newer ideas to be explored in a familiar environment.
It prevents spammers from sending spam with a faked "From:" header that claims it originates from Gmail. This is a very obvious step to maintain the value of an @gmail.com address. Other big webmail sites get filtered aggressively because they get a bad reputation.
If Google makes it easy to identify legitimate e-mails from them, they will not suffer the same fate.
"I certainly can understand how it would bother someone else, especially if the people that ought to be helping are trying to convince you there isn't even a problem."
Yeah. It's not that I'm incapable of getting something working, it's just that I don't see why I should spend my sunday afternoon figuring why something suddenly doesn't work. Overtime pay when fixing my own computer is not great.
I prefer (not as a plug, just statement of fact) OpenBSD for server stuff and Suse Linux for desktop stuff. Both have significant weaknesses, but they DO WHAT THEY'RE TOLD.
I had major problems with Gentoo solely as a result of insufficient testing (eg, stuff that would have been caught if anyone had tried it before releasing it). My sig before this one was actually "I know I sound like I might like Gentoo. I didn't. Please don't suggest it.".
Unfortunately, that was more of a lightning rod for Gentoo zealots who took it upon themselves to convince me that my experience wasn't bad. That doesn't make much sense to me, but they did it consistently. As if there was one little thing I missed that would have made my experience perfect.
It turns out the Gentoo zealots love the speed and adaptability so much that they don't even notice the stuff that bugs the rest of us, so they think we're wrong when we have criticisms.
But don't many toasters only have 1 bit of memory? And that bit tends to zero itself after a minute or two. Can NetBSD be ported to a machine with only one bit?
I've found that the Windows scheduler is remarkably stupid at responsiveness. CPU bound threads are not kept under control, they can make the whole system unresponsive.
Another flow of control lets other stuff do SOMETHING when the CPU bound thread is still busy, so the system "feels" faster. It's really just a reduction in latency. Windows will benefit more than anyone else because they're the worst at it right now.
Most of the *nixen are fairly good at responsiveness because they've always given tasks that sleep more higher priority. They stand to benefit less because they're not that bad at it right now.
Slashdot Mirror
Generally, we don't hate companies for being too big, we hate companies when they start acting like asshats. For example, IBM has been well behaved recently, so most people like them.
Besides, Google has been big for a long time, big enough that they've been able to work on some projects that didn't generate any revenue for quite a while. Some of these things are starting to come together, so people are starting to notice, but their size is not a new thing.
Remember, they've been the default search engine for most of the web for years. It's impossible to do that without being big.
"Unlike some of the other posts though, I'm not worried about RAM or hard drive space, I'm worried about L1 and L2 cache space. If a sizable amount of the data cached goes from 32 to 64 bits, the processors could take a big performance hit unless the caches are bigger too."
This is a valid concern, but the hit isn't that big. IIRC, 64-bit code and data tends to be about 10-15% bigger, doubling the cache is good enough to overcome this. Looking at the cache size of major x86 processors, the Prescott has gone from 8k to 16k L1 and 512k to 1024k L2. AMD64 processors are constant (starting to move to 1m L2 on the newest ones), but they had a lot of L1 to begin with.
Have you been watching cache sizes lately? They were constant at around 256k-512 k for close to 10 years (my P2 from 1996 has the same cache as my P4 from 2004). They're just beginning to increase now with 90 nm because cache increase is the most beneficial thing to do with the transistors becoming available.
"Here's hoping programmers keep using 32 bit data structtures when they don't need 64 bits!"
This is not a new thing. How many people now use a 32-bit int when an 8-bit char would do? Pretty much everyone, in my experience. They won't bother optimizing. When they do, it's because it needs a resource that's very scarce (network traffic, or huge numbers of the data structure being used). Also consider is that there's usually a penalty to loading values from a non aligned address. I'm not saying this dominates, but using the smallest type you can is almost never free.
In the case of PowerPC (and SPARC, MIPS, anything else I can think of), the opcodes is still 32 bits. The number of registers stays the same, the number of instructions stays (more or less) the same. There is no reason to extend the size of opcodes.
"You didn't click on the link did you? Sure a little scripting of GMail can make it fast, but as far as rich interfaces go, I don't think it would look so good in comparison to a XUL interface. Go look at the demo, and spend some time clicking around and using all the features of it."
I did click in fact, and I don't really see much difference in user experience between gmail and the link you posted. The one you posted looks more like a native app, but to me they seem about as fast as each other,
"Now, imagine if GMail started doing that... IE users of GMail get the standard webmail interface, but Firefox users get a full fast XUL interface. Have a look at that demo site again, and do some clicking around ... then tell me that that wouldn't be an absolutely killer app for Firefox."
Gmail does do some kind of local scripting. I don't know the specifics, but it's very fast, much faster than it would be if it had to access anything online for every click. Also, there's autocomplete boxes that pop up when you start typing in a From: box that narrow down the list with each letter you type. That and other stuff makes me sure that a lot of what it does happens locally.
For that reason, I kinda doubt XUL would be much faster.
It's not a bit more heat. It's a lot more heat.
The only time they're worth it is when you're trying to achieve a temperature below the ambient temperature. Otherwise, it's easier to put a heat spreader on the chip (as AMD and Intel already do) so the contact area with the heat sink is bigger.
These peltiers aren't going to go in any general purpose CPUs as we know them.
Kerry might not be much better than Bush, but he is better than Bush. One thing at a time.
They're faster. I didn't say they were worth it. :)
If the Diebold machines were perfectly accurate, they'd still only be recording the votes of idiots.
While I think the grandparent with his head in the clouds deserved to get yelled at for being wrong (as OS X does not rival the high end *nixen), smart aquisitions are part of business and buying up NeXT was a very good idea. Just because it was around before doesn't mean Apple didn't make very good progress in a short period of time.
"Good luck with that. Especially since it's illegal under the DMCA."
That's funny, I thought it was explicitly allowed. Granted that's just Canada... but the US is not the only country out there.
"And since future versions of Intel hardware and Microsoft software will put a hard block on your ability to do this."
pffft
That assumes that every stage of the setup secure, and that details are never leaked. Given that the whole thing is designed by committee on a deadline, and that they're going to be dealing with people that can sniff the bus, I find that unlikely. Indeed, Microsoft claiming they will be able to provide unbreakable DRM is equivilant to Microsoft claiming they can provide perfect security. For example, one of the vulnerabilities that allowed an XBox to play pirated games was in the firmware itself.
And then, at the end of the day, the best case they can hope for is forcing everyone to use the analog hole.
Compared to MacOS X, I would say that Solaris/X11/Gnome does not have advanced graphical capabilites. While Solaris/X11/Gnome does have OpenGL support, it does not (AFAIK) have anything similar to Quartz Extreme.
Is this at all related to that effect where the gravitational influence of the sun is slightly reduced at the point of totality in a solar eclipse?
I've been vaguely looking around for more information on that. From what I can gather, it's a reproducible observation, but other effects haven't been ruled out, so no one's sure what to make of it.
Yes, I believe they ship OS X Server with the Xserves. Also, if you're doing double precision floating point stuff the PPC970 is the best chip you can buy this side of Itanium.
However, those things have nothing to do with the mainframe type stuff. The whole point of clusters is to have relatively inexpensive interchangable machines. The whole point of a mainframe is to have one bulletproof system. The CPUs are hotpluggable, the filesystems can move from disk to disk without being taken offline, etc. MacOS X cannot do those things. Similarly, the mainframe OSes don't bother trying to have advanced graphical capabilities, or any of a hundred things MacOS does better than anyone.
I'm not saying Solaris or AIX is "better", only that they're better at what they're designed for.
oh, jeeze... Please, for all of our sakes, get your head out of the clouds.
MacOS tries to be a desktop OS, and it succeeds brilliantly. However, MacOS does not try to be a mainframe OS. It doesn't run on mainframes, and would't do a very good job if it could. This is what Solaris and AIX do, and they do do a good job.
The two things aren't even remotely comparable.
There's a distinction here no one else seems to see.
I think the actual use of P2P software to upload or download files might be unethical, but the phenomenon of file sharing in general is nothing more and nothing less than the invisible hand attempting to correct the price of the media being distributed. It's no more unethical than the phenomenon of male Lions killing cubs when they take over a pride. It might be nasty when each individual does it, but the practice has helped Lions survive as a species.
The simple fact is that the market is hardwired into humans. Trade is in our nature. We may argue with actions people take, but there's very little that can be done to stop anything completely. We can no more stop file sharing than we can the drug trade or prostitution. There's demand, there's supply, the market takes care of the rest whether we like it or not.
The problem with those trying to stop it is that they're fighting human nature. Human nature won't change. It's not that the **AA is wrong (or at least, exclusively wrong), it's that their goal is not achievable. They will either continue to fight, capitulate, or compromise. My money's on compromise, but no bets on when.
The price of the media in question is higher than the market is willing to bear. I'm not saying it is or is not a fair price, only that it's higher than people are willing to pay. As a result, there were a lot of people that wanted the media but didn't want to pay for it. To the tune of several times the total volume of legitimate sales.
Napster was growing 15% a day for several months. And I've got news for you, Napster sucked. That just doesn't happen unless there's a huge unmet demand.
It's the **AA's fault for ignoring the market. While I might consider the actions of the file traders themselves to be unethical, I recognize that they're following human nature.
Safari is not a great browser. It has significant compatability problems, at least with the pages I used it with.
Camino is the way to go. Once loaded, it's as fast as Safari, and it's MUCH better at rendering correctness. It's designed for OS X, and looks like a native app. It has had a minimalist interface similar to Safari's since before Safari existed. It's easily the best browser for Mac.
"They now offer a single processor low-end G5. This lowers the price of entry into their pro-range. It helps keep up with demand, given that supply of G5 processors is an issue."
The dual-only PowerMacs were the thing keeping Apple from being a consideration for me. I'm glad I waited. I can only assume G5 availability from IBM was not good enough to start selling a lower margin machine.
I'll tell you what though, a thousand frenzied Apple fans chanting "Sure it costs more, but it's still worth it!" is about the most annoying thing I've ever seen on Slashdot. That argument is so intellectually bankrupt that I didn't even get modded down yelling at them. It was a lot more money without greater utility, because the second CPU does nothing unless you're CPU-bound. Some of us just want a computer that we can upgrade in a few years.
Now that there's a machine that does what people like me need, it can be truthfully said they have a well rounded lineup. It will be a little while for me, but I think I will recommend a single-1.8 for my parents. I have no desire to get dragged back there for tech support, no matter how good my mothers cooking is.
"I don't care nearly as much as I used to about the what the kernel does; it's so easy to emulate your way back to a familiar state."
DragonFlyBSD has a system call layer that would allow potentially very different interfaces to be presented to userspace stuff with essentially no penalty. This may allow newer ideas to be explored in a familiar environment.
It prevents spammers from sending spam with a faked "From:" header that claims it originates from Gmail. This is a very obvious step to maintain the value of an @gmail.com address. Other big webmail sites get filtered aggressively because they get a bad reputation.
If Google makes it easy to identify legitimate e-mails from them, they will not suffer the same fate.
"I certainly can understand how it would bother someone else, especially if the people that ought to be helping are trying to convince you there isn't even a problem."
Yeah. It's not that I'm incapable of getting something working, it's just that I don't see why I should spend my sunday afternoon figuring why something suddenly doesn't work. Overtime pay when fixing my own computer is not great.
I prefer (not as a plug, just statement of fact) OpenBSD for server stuff and Suse Linux for desktop stuff. Both have significant weaknesses, but they DO WHAT THEY'RE TOLD.
I had major problems with Gentoo solely as a result of insufficient testing (eg, stuff that would have been caught if anyone had tried it before releasing it). My sig before this one was actually "I know I sound like I might like Gentoo. I didn't. Please don't suggest it.".
Unfortunately, that was more of a lightning rod for Gentoo zealots who took it upon themselves to convince me that my experience wasn't bad. That doesn't make much sense to me, but they did it consistently. As if there was one little thing I missed that would have made my experience perfect.
It turns out the Gentoo zealots love the speed and adaptability so much that they don't even notice the stuff that bugs the rest of us, so they think we're wrong when we have criticisms.
Not only are you an (apparently) GPL supporter that correctly characterizes the BSD crowd, but I like your signature.
But don't many toasters only have 1 bit of memory? And that bit tends to zero itself after a minute or two. Can NetBSD be ported to a machine with only one bit?
I'd like to see that.
I've found that the Windows scheduler is remarkably stupid at responsiveness. CPU bound threads are not kept under control, they can make the whole system unresponsive.
Another flow of control lets other stuff do SOMETHING when the CPU bound thread is still busy, so the system "feels" faster. It's really just a reduction in latency. Windows will benefit more than anyone else because they're the worst at it right now.
Most of the *nixen are fairly good at responsiveness because they've always given tasks that sleep more higher priority. They stand to benefit less because they're not that bad at it right now.