First off, he needs to shoot more selectively and learn how to cull. I've worked with professional photographers (and have been one) and they know not to keep everything they shoot. There are tools specifically made to help the issue. There is no need to keep images you know you'll never use (and it's easy to know) especially in a capacity-intensive application.
Second, write performance of RAID 5 is certainly adequate for the application and larger spindles would have been a far better choice. He could have achieved the same capacity with 3 or 4 drives and still had totally adequate performance.
That is incorrect. A 5+5 mirrored array can tolerate, at most, 1 arbitrary drive failure. A correct alternative would be RAID 5 in a 9+1 configuration. Capacity would be much higher with RAID 5 but write performance would suffer. Write performance probably doesn't matter in a professional photographer's application however.
Hot spares only make sense in a system that can tolerate little or no downtime. Such would not be the case in a workstation so drives dedicated to that would be wasted.
Assuming 1500GB was the desired capacity, the right choice currently would be to use 500GB spindles in a 3+1 RAID 5 configuration. Much smaller, cheaper, quieter, cooler, and more reliable than what was done. The RAID 6 config with hot spares is a total waste of money. If performance is a concern, a 3+3 mirrored array would be an alternative.
Some capacity on a drive is already used for redundancy and the techniques used are far more powerful than simple mirroring or XOR that RAID uses. In fact, RAID 1, 4, and 5 take advantage of the fact that such redundant coding exists, otherwise there would be no method of detecting certain kinds of failures. RAID 2 and 3 are unpopular but they do include such error detection capabilities.
It would be INCREDIBLY stupid to implement RAID parity across platters of a single drive. The result would be entirely useless, wasteful, and would slow the drive to a crawl. Leave the design of these things to people who know what they are doing. What you desire is already done inside all drives and in a manner far superior than you suggest.
Touchscreen keyboards don't take the place of thumb keyboards unless you rarely need a keyboard at all (then T9 works). Furthermore, they take up precious screen real estate. Besides, why doesn't this device come with one already?
Apple doesn't face the massive piracy issues that MS does and has no need to take such measures. If Apple decided to sell OS X for generic PC's you can bet they'd end up with registration.
Apple is the only company to use DRM to a lock its OS to its hardware. Where's the trust?
Both of your quoted sources use "RISC" rather loosely. Internal ops of modern processors have nothing to do with RISC instruction sets and it's absurd to suggest that they do. Modern RISC processors, the G5 for example, have instruction set decoders that feed "RISC engines", too. If they were RISC in the traditional sense, why would the G5 bother?
It is a fact that you can't execute instructions in microcode if you want them to go fast. RISC concepts, when they mattered, were all about making that job easier to do. They accomplished this by reducing or eliminating instructions that required microcode, by making instructions easy to decode, and by pushing off work onto the compilers. RISC processors, when the term was all the rage, were very primitive in their capabilities. These days RISC tends to be defined differently but that's revisionist history. RISC stripped out everything hard and added it back as their capabilities grew. It wasn't about load-store, it was about eliminating addressing modes to make decoding easier. It was the 80/20 rule and RISC optimized the big piece and discarded the rest.
These days, there's not a single threadbare RISC processor left in the desktop space. The last of the great RISC champions, the PowerPC, has an instruction set so powerful it's hard to argue there's anything simplified or reduced to it. The RISC vs CISC argument raged and the fact was that it ultimately didn't matter. If a winner must be declared then it has to be the only one left. All modern processors are designed the same way save Itanium.
Pentium Pro became PII, then PIII. Pentium M bears much resemblance to the PIII and Core/Core 2 are decendents of the Pentium M. The so called "RISC cores" of these processors are not all the same though.
"PS: If you think RISC lost the war, then remember that modern x86 processors consist of a RISC core with a translator stage to handle all those pesky, legacy CISC instructions."
The "core" of modern x86 processor is nothing like RISC processors from back when the arguments raged. All modern processors implement designs similar to x86 rather than execute their instruction sets directly. Claiming that these internal engines are "RISC" is preposterous.
The "IS" in RISC and CISC stands for "Instruction Set" and it's been shown through the current performance of the x86 that the choice of instruction set is not the largest determiner of performance. Obviously, the x86 IS isn't sufficiently bad for it to be unable to compete. Back when transistor budgets were small these things mattered more. Too bad RISC processors didn't run Lotus 123 then.
The RISC vs. CISC war is over and it turned out to be unimportant.
Back in the day, my company hired an IBM fellow to run R&D. He was a brilliant mind and an expert at processor design having been instrumental on early RISC development in his previous job. One of the first things he explained to us was the wall that x86 was about to hit due to its CISC design. It was clear to him that the 386 had one more generation of meaningful performance left and that we would need to prepare for a shift to RISC processors and workstation capabilities. He couldn't have been more wrong.
Arguing RISC vs CISC is like arguing between a horse and a mule. The horse lovers simply didn't understand the possibility of future invention. The x86 today doesn't have a horse under the hood, it has a jet engine.
The "naive conclusion" was unsupported by the facts due to the "correct conclusion". Your additional argument, that Alpha was 3x more "powerful" than a corresponding VAX, is also meaningless. What constituted a "corresponding VAX"? Alphas where to be used in future VAX'es so there were inherent generational differences in play. Finally, there is no reason to believe that VAX'es were representative of the best of breed in CISC design.
All that limiting instructions to gain performance means is that the instructions you avoid are poorly implemented. You may benefit from an alternate design that excludes the avoid instructions or you may benefit more by fixing the slow instructions. You ust also consider why the compiler produced slow code to begin with (as the original poster said).
Alpha's were great performers in their day but they were also incredibly power-hungry and their performance was not sufficiently compelling for them to succeed.
how is this an option? suppliers negotiate their contracts so if they lose money it's their fault.
Dell is better at negotiating with suppliers than the competition. That's good business, not abuse. A vendor can choose to not do business with Dell and some do.
The thread is there for all to see, including you. I acknowledged being corrected on the Sorenson/Linux matter. It remains a fact that the sorenson codecs were supported without the efforts of Apple or Sorenson, both of whom kept those formats undocumented.
The original claim in this thread is that Apple is a champion of open standards. Sorenson is an example of that not being the case regardless of whether a team had reverse-engineered it or not.
MS has proprietary codecs because they develop them. Apple licensed Sorenson then replaced it with open standards. Both are examples of Apple using other people's work. Perhaps that's in the best interests of the user but the fact is that MS is contributing more to codec development even if they want to lock it to their platform.
I won't defend Windows and I didn't offer comments on it, but there are attempts on the internet to benchmark the performance of the kernels of popular OS'es, and those results show OS X to be slower than others including Windows. Your examples include booting, shutdowns, and explorer (finder) performance, none of which are the kernel itself.
I have a MBP and it is quite slow and unresponsive at times. Yes, if you don't have a color wheel then it responds. I see the color wheel quite often and for long periods. I don't make great demands of the machine and have 1G or RAM as well. My MBP is not fast by my measure.
"but the impartials do seem to prefer mac"
according to you they do. Good is subjective as both of us have said. I don't agree that OS X is better nor do I claim it is worse. I feel it is prettier and is frequently more consistent. I don't care for it's clickiness nor its tendency to require lots of mouse movement. I can live with OS X but I don't feel burdened by Windows.
It is true. If you only download formats that your mac can play then it's not true for you. I have video that cannot be played on a mac.
WMV may not be necessary but that's the choice of the content provider. If you have interest in that content then being able to play it IS necessary.
I haven't forgotten that people use iTunes and people have quicktime. It's irrelevant when it's the content providers choosing the format and they are only providing WMV. Flip4mac doesn't play DRM'ed WMV files and DRM is precisely the reason some content chooses to distribute in WMV.
I was not aware of that since I haven't encoded video for years. Here's an excerpt:
"As Apple began to move away from proprietary codecs with its embrace of MPEG-4, Sorenson Media next licensed the newest version of the codec to Macromedia as Sorenson Spark (Sorenson H.263), released with Macromedia Flash 6/MX on March 4, 2002. [1]
The specifications of the codec were not public, and for a long time the only way to play back Sorenson video was to use Apple's QuickTime player, or the MPlayer for Unix/Linux, which in turn piggy-backed Microsoft Windows DLL-files extracted from Apple's player.
According to an anonymous developer1 of FFmpeg, reverse engineering of the SVQ3 codec revealed it as a tweaked version of H.264. The same developer also added support for this codec to FFmpeg, making native playback on all platforms supported by FFmpeg possible."
It's not clear to me that all versions of Sorenson are supported since earlier ones predated H.264. Nevertheless, here's the original statement that I responded to:
"Hurray about those Open Standards huh? Good thing Apple's been a champion of them for years, with Quicktime supporting the book of Open Standards and VLC to support the rest of the nutso formats and encodings."
Sorenson was never an open standard and for many years it was Quicktime's main video codec. The fact that it is playable now is not due to any open efforts by Apple or Sorenson. Apple always does what's in their business interest just like MS does. MS simply does their own codec development while Apple uses industry standards.
Easy. Make it suitable for the living room but with far greater IO capability than a mini. Use a Merom to keep it silent and it won't compete with the Mac Pro. It may destroy the market for the mini but the mini isn't that great anyway.
The problem with Apple's headless solution is that it doesn't have sufficient IO. The size of the mini isn't sufficiently attractive to make up for that. If it were, Apple could afford to offer both machines. As it stands, Apple doesn't have a media PC offering at all.
It make Windows a better system if playing DRM'ed WMV files is important to you. Specifically, it means that MCE is the only game in town unless you are willing to accept the loss of certain content. By certain content I mean downloaded porn;)
Apple is a solutions company, not simply hardware or software. Their business isn't viable just competing with machines or just competing with software. They know that from experience.
"fast, good, cheap, pick two." that assumes Apple offers any of these.
Fast? No, used to be slow and now is the same. OS X is somewhat slower than competing OS'es. Good? Subjective, but made from the same parts bins but with polish. Software choices are good in some ways and bad in others. Cheap? Of course not. Not always expensive though.
Hip, pretty, alternative. Those are what Apple offers. Fast, good, cheap, only in Apple marketing campaigns.
Yes, Fortune 500 businesses are all set up in trailer parks.
"In case you havent noticed the new Intel based macs" are 100% PC with Apple skins. That's opposed to 98% that the previous Mac's were. Makes sense that they're sold right next to the Gucci boutique.
Good to know you're such a discriminating shopper.
Slashdot Mirror
First off, he needs to shoot more selectively and learn how to cull. I've worked with professional photographers (and have been one) and they know not to keep everything they shoot. There are tools specifically made to help the issue. There is no need to keep images you know you'll never use (and it's easy to know) especially in a capacity-intensive application.
Second, write performance of RAID 5 is certainly adequate for the application and larger spindles would have been a far better choice. He could have achieved the same capacity with 3 or 4 drives and still had totally adequate performance.
That is incorrect. A 5+5 mirrored array can tolerate, at most, 1 arbitrary drive failure. A correct alternative would be RAID 5 in a 9+1 configuration. Capacity would be much higher with RAID 5 but write performance would suffer. Write performance probably doesn't matter in a professional photographer's application however.
Hot spares only make sense in a system that can tolerate little or no downtime. Such would not be the case in a workstation so drives dedicated to that would be wasted.
Assuming 1500GB was the desired capacity, the right choice currently would be to use 500GB spindles in a 3+1 RAID 5 configuration. Much smaller, cheaper, quieter, cooler, and more reliable than what was done. The RAID 6 config with hot spares is a total waste of money. If performance is a concern, a 3+3 mirrored array would be an alternative.
Some capacity on a drive is already used for redundancy and the techniques used are far more powerful than simple mirroring or XOR that RAID uses. In fact, RAID 1, 4, and 5 take advantage of the fact that such redundant coding exists, otherwise there would be no method of detecting certain kinds of failures. RAID 2 and 3 are unpopular but they do include such error detection capabilities.
It would be INCREDIBLY stupid to implement RAID parity across platters of a single drive. The result would be entirely useless, wasteful, and would slow the drive to a crawl. Leave the design of these things to people who know what they are doing. What you desire is already done inside all drives and in a manner far superior than you suggest.
Touchscreen keyboards don't take the place of thumb keyboards unless you rarely need a keyboard at all (then T9 works). Furthermore, they take up precious screen real estate. Besides, why doesn't this device come with one already?
Furthermore, would there be any point to an agreement that wasn't "legally binding"?
It's not that surprising that it didn't happen at all. There's no journalistic integrity here.
Apple doesn't face the massive piracy issues that MS does and has no need to take such measures. If Apple decided to sell OS X for generic PC's you can bet they'd end up with registration.
Apple is the only company to use DRM to a lock its OS to its hardware. Where's the trust?
There were LiIon batteries in laptops before 1995. Dell had an exclusive with Sony in 92 or 93.
http://www.answers.com/topic/lithium-ion-battery
That article claims that Toshiba was the first in 1993 but Dell predated Toshiba as I recall. Apple was nowhere near first.
Who can argue with Wikipedia?
Both of your quoted sources use "RISC" rather loosely. Internal ops of modern processors have nothing to do with RISC instruction sets and it's absurd to suggest that they do. Modern RISC processors, the G5 for example, have instruction set decoders that feed "RISC engines", too. If they were RISC in the traditional sense, why would the G5 bother?
It is a fact that you can't execute instructions in microcode if you want them to go fast. RISC concepts, when they mattered, were all about making that job easier to do. They accomplished this by reducing or eliminating instructions that required microcode, by making instructions easy to decode, and by pushing off work onto the compilers. RISC processors, when the term was all the rage, were very primitive in their capabilities. These days RISC tends to be defined differently but that's revisionist history. RISC stripped out everything hard and added it back as their capabilities grew. It wasn't about load-store, it was about eliminating addressing modes to make decoding easier. It was the 80/20 rule and RISC optimized the big piece and discarded the rest.
These days, there's not a single threadbare RISC processor left in the desktop space. The last of the great RISC champions, the PowerPC, has an instruction set so powerful it's hard to argue there's anything simplified or reduced to it. The RISC vs CISC argument raged and the fact was that it ultimately didn't matter. If a winner must be declared then it has to be the only one left. All modern processors are designed the same way save Itanium.
Pentium Pro became PII, then PIII. Pentium M bears much resemblance to the PIII and Core/Core 2 are decendents of the Pentium M. The so called "RISC cores" of these processors are not all the same though.
"PS: If you think RISC lost the war, then remember that modern x86 processors consist of a RISC core with a translator stage to handle all those pesky, legacy CISC instructions."
The "core" of modern x86 processor is nothing like RISC processors from back when the arguments raged. All modern processors implement designs similar to x86 rather than execute their instruction sets directly. Claiming that these internal engines are "RISC" is preposterous.
The "IS" in RISC and CISC stands for "Instruction Set" and it's been shown through the current performance of the x86 that the choice of instruction set is not the largest determiner of performance. Obviously, the x86 IS isn't sufficiently bad for it to be unable to compete. Back when transistor budgets were small these things mattered more. Too bad RISC processors didn't run Lotus 123 then.
The RISC vs. CISC war is over and it turned out to be unimportant.
Back in the day, my company hired an IBM fellow to run R&D. He was a brilliant mind and an expert at processor design having been instrumental on early RISC development in his previous job. One of the first things he explained to us was the wall that x86 was about to hit due to its CISC design. It was clear to him that the 386 had one more generation of meaningful performance left and that we would need to prepare for a shift to RISC processors and workstation capabilities. He couldn't have been more wrong.
Arguing RISC vs CISC is like arguing between a horse and a mule. The horse lovers simply didn't understand the possibility of future invention. The x86 today doesn't have a horse under the hood, it has a jet engine.
The "naive conclusion" was unsupported by the facts due to the "correct conclusion". Your additional argument, that Alpha was 3x more "powerful" than a corresponding VAX, is also meaningless. What constituted a "corresponding VAX"? Alphas where to be used in future VAX'es so there were inherent generational differences in play. Finally, there is no reason to believe that VAX'es were representative of the best of breed in CISC design.
All that limiting instructions to gain performance means is that the instructions you avoid are poorly implemented. You may benefit from an alternate design that excludes the avoid instructions or you may benefit more by fixing the slow instructions. You ust also consider why the compiler produced slow code to begin with (as the original poster said).
Alpha's were great performers in their day but they were also incredibly power-hungry and their performance was not sufficiently compelling for them to succeed.
"taking legal action to recover their money"
how is this an option? suppliers negotiate their contracts so if they lose money it's their fault.
Dell is better at negotiating with suppliers than the competition. That's good business, not abuse. A vendor can choose to not do business with Dell and some do.
The thread is there for all to see, including you. I acknowledged being corrected on the Sorenson/Linux matter. It remains a fact that the sorenson codecs were supported without the efforts of Apple or Sorenson, both of whom kept those formats undocumented.
The original claim in this thread is that Apple is a champion of open standards. Sorenson is an example of that not being the case regardless of whether a team had reverse-engineered it or not.
MS has proprietary codecs because they develop them. Apple licensed Sorenson then replaced it with open standards. Both are examples of Apple using other people's work. Perhaps that's in the best interests of the user but the fact is that MS is contributing more to codec development even if they want to lock it to their platform.
I won't defend Windows and I didn't offer comments on it, but there are attempts on the internet to benchmark the performance of the kernels of popular OS'es, and those results show OS X to be slower than others including Windows. Your examples include booting, shutdowns, and explorer (finder) performance, none of which are the kernel itself.
I have a MBP and it is quite slow and unresponsive at times. Yes, if you don't have a color wheel then it responds. I see the color wheel quite often and for long periods. I don't make great demands of the machine and have 1G or RAM as well. My MBP is not fast by my measure.
"but the impartials do seem to prefer mac"
according to you they do. Good is subjective as both of us have said. I don't agree that OS X is better nor do I claim it is worse. I feel it is prettier and is frequently more consistent. I don't care for it's clickiness nor its tendency to require lots of mouse movement. I can live with OS X but I don't feel burdened by Windows.
I'm not really sure how to respond to your post, but I really enjoyed your sig.
It is true. If you only download formats that your mac can play then it's not true for you. I have video that cannot be played on a mac.
WMV may not be necessary but that's the choice of the content provider. If you have interest in that content then being able to play it IS necessary.
I haven't forgotten that people use iTunes and people have quicktime. It's irrelevant when it's the content providers choosing the format and they are only providing WMV. Flip4mac doesn't play DRM'ed WMV files and DRM is precisely the reason some content chooses to distribute in WMV.
by that I meant content only playable using WMP10 or newer. I believe any DRM'ed content requires that.
I have WMV files downloaded from the internet that have DRM and require WMP10+ yet do not include executables.
I was not aware of that since I haven't encoded video for years. Here's an excerpt:
"As Apple began to move away from proprietary codecs with its embrace of MPEG-4, Sorenson Media next licensed the newest version of the codec to Macromedia as Sorenson Spark (Sorenson H.263), released with Macromedia Flash 6/MX on March 4, 2002. [1]
The specifications of the codec were not public, and for a long time the only way to play back Sorenson video was to use Apple's QuickTime player, or the MPlayer for Unix/Linux, which in turn piggy-backed Microsoft Windows DLL-files extracted from Apple's player.
According to an anonymous developer1 of FFmpeg, reverse engineering of the SVQ3 codec revealed it as a tweaked version of H.264. The same developer also added support for this codec to FFmpeg, making native playback on all platforms supported by FFmpeg possible."
It's not clear to me that all versions of Sorenson are supported since earlier ones predated H.264. Nevertheless, here's the original statement that I responded to:
"Hurray about those Open Standards huh? Good thing Apple's been a champion of them for years, with Quicktime supporting the book of Open Standards and VLC to support the rest of the nutso formats and encodings."
Sorenson was never an open standard and for many years it was Quicktime's main video codec. The fact that it is playable now is not due to any open efforts by Apple or Sorenson. Apple always does what's in their business interest just like MS does. MS simply does their own codec development while Apple uses industry standards.
Easy. Make it suitable for the living room but with far greater IO capability than a mini. Use a Merom to keep it silent and it won't compete with the Mac Pro. It may destroy the market for the mini but the mini isn't that great anyway.
The problem with Apple's headless solution is that it doesn't have sufficient IO. The size of the mini isn't sufficiently attractive to make up for that. If it were, Apple could afford to offer both machines. As it stands, Apple doesn't have a media PC offering at all.
It make Windows a better system if playing DRM'ed WMV files is important to you. Specifically, it means that MCE is the only game in town unless you are willing to accept the loss of certain content. By certain content I mean downloaded porn ;)
"WMV and Real are just as good on the Mac as they are in Windows."
No. Not all WMV is supported on the mac. Flip4mac does not do drm'ed WMV files and MS has purposefully eliminated that support.
"They can be played through QuickTime using Flip4Mac, VLC, or mplayer without problems."
Sorry, not all of them.
For a platform that claims it's the ultimate for multimedia, why is it that all the solutions here are 3rd party? I thought macs "just worked"?
Sorenson for one
Apple is a solutions company, not simply hardware or software. Their business isn't viable just competing with machines or just competing with software. They know that from experience.
"fast, good, cheap, pick two." that assumes Apple offers any of these.
Fast? No, used to be slow and now is the same. OS X is somewhat slower than competing OS'es.
Good? Subjective, but made from the same parts bins but with polish. Software choices are good in some ways and bad in others.
Cheap? Of course not. Not always expensive though.
Hip, pretty, alternative. Those are what Apple offers. Fast, good, cheap, only in Apple marketing campaigns.
How much for the 500GB hard drive upgrade? Oh yeah, you can't get one because the mini can't use real hard drives. Great comparison.
The reason you can't buy a mini from HP or Dell is that they don't have a market for it. HP and Dell know their markets better than you or Apple do.
Yes, Fortune 500 businesses are all set up in trailer parks.
"In case you havent noticed the new Intel based macs" are 100% PC with Apple skins. That's opposed to 98% that the previous Mac's were. Makes sense that they're sold right next to the Gucci boutique.
Good to know you're such a discriminating shopper.