3-4 years is the cycle life of a server, generally. In the old days, disk arrays might outlive their servers, as they were potentially many times more expensive than their servers, and in those cases, You'd probably be more likely to add a live mirror than swap disks.
Today, with SANs, most of this discussion is moot. You'd have hot spare servers ready, if you were seriously interested in uptime, complete with the last full back up (you did test your backup by doing a restore, no?) and would merely need to update the transaction logs from then to present. (Lots of assumptions there, it really depends upon the particulars of your situation.)
As for Google, I'd really love to see the breakdown of drive manufacturers and types. From my own experience with 500 or so Seagate SCSI HDs my failure rate was closer to 0.1% across 4 years (I believe we only lost 3 or 4 drives out of 500). The 200 or so variety of IDE drives in lesser servers failed at a much higher rate, especially considering we had a bad batch of Corsair IDE drives in a set of 30 Dells we bought, with 5 HDs failing during the burn in test (72 hours of continuous full load testing). All of this was during the period of 93 through 99, across several batches of servers. The SGI boxes we bought in 91 didn't suffer a single drive failure in 9 years (total of 8 drives, all SCSI II, think they were Seagate).
Back then, you bought Seagate if you wanted to sleep easy. I personally own several Seagate drives, one even from 90, that I believe still works although it's been retired for the last 4 or so years. (It was serving as a scratch disk). I've had 2 Hitachi (pre Deathstar) 18GB SCSI drives both pretty much fail after 8 years, one catastrophically. IBM drives are another story, I received a large batch of OEM/refurb drives (twenty 9GB, fifteen 18GB) that had several failures within a 3 year period. I still have 5 36 GB drives that are currently idle, having been replaced by 320GB SATA drives. (Much cheaper, quieter, cooler, and a little slower but much much larger;)
In the 90s, drive manufacturer was very important. I do not have enough data to make definitive statements about today's drives other than the less expensive IDE and SATA drives definitely don't seem as reliable or long-lived as last decade's SCSI drives.
Vista is hardly failing. Were Vista any company's product other than MS, it would be an utter and complete failure and be a footnote today. The only reason it's alive today is because of MS's monopoly (but weakening) hold on the distribution chain. There have been questions raised as to how many "real" copies have actually been sold, and how many are marketing arithmetic.
I'm not exactly sure what you mean by that the "main UI issues that made them not want Vista in the first place will not be available in 2-3 years either". Firstly, I highly doubt that the reason why businesses are hesitant to adopt Vista is UI related, and even if it was, one would think that the absence of show-stopping issues would be a positive thing. You under-estimate the power of inertia. Businesses don't want new interfaces. New interfaces mean losses due to training and lowered productivity due to inexperience. From a business perspective, software UI changes are bad. You have to remember that the bulk of users in this category are largely computer illiterate, and might even have problems turning their computer on or using the cup holder.
In 2-3 years, Vista's interface will still be Vista's, and not XPs. Linux, however, will probably provide a desktop almost indistinguishable from today's XP. So, if your choice was to pay MS large sums of cash for Vista and suffer the losses due to training and lowered productivity, or switch vendors and install Linux with lowered costs and no losses in productivity and training, which way would you go?
In either case, "closely mimicing" UIs and applications is not what businesses want. They want the actual software they've gotten comfortable with, and they want an OS compatible with their existing infrastructure. Not some alternative that does less or exactly the same, just in a different way. After all, the training and initial drop in productivity per worker is far more expensive to a Microsoft-only company than a Windows XP license and a copy of Office 2007. I'm not sure why you're arguing, as you've just supported my argument, and possible presented it better than my first post. Read the above with "Vista" in mind as the alternative. BTW, Office 2007 is also not as widely adopted as some would like. I'm using it at work, and I must admit that I dislike it greatly. It's not merely the UI menu changes that suck, but the entire interface just reminds me of Fisher Price.
It's all conditional. Until it fails, do you really know it's going to fail? If I knew which drive was going to fail and when, I wouldn't need RAID in the first place.
Putting in process and regular refreshes of hardware is the surest way to prevent a multi-disk loss. (Yes, you'd want to do this offline)
Your supposition is that MS will continue to sell XP. They've stated that their desire is to EOL it ASAP. The failure of Vista is apparent in the undesired extended sales cycle of XP demanded by customers.
Those customers know that XP will end. They also know that barring an asteroid or hell freezing over, the main UI issues that made them not want Vista in the first place will not be available in 2-3 years either. What will be available are most likely very close clones of the apps they want on OSes that mimmic their known UIs. It won't be a MS OS, but most likely Linux varieties.
Such is how monopolies screw up and lose their market in the digital age.
Right, but when you build production systems, your mirror halves will be from different lots, and you might even preemptively swap half out on a time schedule, depending upon how crucial your data is. So the odds of 2 drives in a mirror set failing simultaneously, or near enough to affect the full data set becomes astronomical.
...The thing is that Microsoft isn't losing money on Vista. On the contrary. They're making good money off of both their new and their old product, and having two products on a market supplementing eachother like XP and Vista do is only a positive thing for Microsoft. It helps retain the userbase through to the next generation of the OS. Actually, they are losing money on Vista. Failure to upgrade to Vista produces a cascading failure to upgrade other products (Office, some of the DRM products) and also loses them mindshare. In case you haven't been following the magazines that cover these types of things, CIO's not only are now aware of alternatives to MS, but are actively considering the potential to rolling them out instead of Windows. This year is the first year that CIOs responded with the potential for rolling out Macs. That's huge news.
The reasons for this? The utter and abject failure of Vista. Vista means retraining. Vista means large hardware upgrade costs. Vista means a whole slew of new IT issues. When taken in that context, alternatives that seemed untenable just a year ago all of a sudden start looking attractive.
Will MS go broke? Naah. But they're certainly not retaining users (their only real goal since they're a monopoly). What we are seeing in retrospect is the peak of MS about 5-7 years ago. First Linux took the datacenter, and now a combination of IE dropping to under 75% marketshare coupled with increased interest in alternative OSes and office suites and the recent shenanigans of DirectX all weaken MS's monopoly hold on the desktop.
I think you've missed the point - RAID 1+0 allows you to fail up to n drives (with 2 n needed to build the array).
Additionally, should a drive fail, rebuilding will only marginally affect your performance, degrading it by a fraction compared to a RAID 5/6 rebuild. (Only 1 drive is affected out of your stripe set, the rest perform at peak operational speed)
I'm curious, what exactly was contained in the last EULA from MS that scared you so? I've pretty much ignored MS EULA's as mere crap since about 2000, and don't bother reading them at all anymore as they are closer to bathroom stall adolescent scribbles than any actual binding legal document. (Hint - prove I read it and more importantly, agreed to it....)
That's not to say that MS EULAs shouldn't turn people to other alternatives. Considering how abusive their EULAs are, what would make one think that they would be treated any better in any other interaction with MS?
Operator overloading is a decision that makes or breaks the java language for those that write highly arithmatic code. Some prefer to see their standard symbols in code, and find the sort of code you presented unacceptable. I'm not in either camp on that one, since there are certainly very good and fast special purpose libraries in C, FORTRAN and others that would probably suit you better. I've not had the need to research Java's math capabilities, other than I know that a lack of understanding of Java's internals can create some incredibly slow code compared to what is achievable. (Mainly in multi-dimensional matrix algebra).
This is a non-argument. Until they're a monopoly, you can't accuse them of being a monopoly. Look at the Hunt brothers, they tried to corner to silver market. They failed, went broke, and then were prosecuted for some of their illegal practices to add insult to injury.
You can't convict someone for something they might become before they become it, especially if they become one legally. (Although I'm not aware of a single monopoly that got there legally, unless you count those that were established as monopolies via government fiat).
C# is the antithesis of a decent language. It's a wrapper for java-esque syntax and rules on top of C.
And let's not forget that if a library is written and a class has a method not declared with the "override" keyword, it's equivalent to a Java library with a method declared "final". I'm sure libraries have never been inadvertently made useless in C# because someone was too lazy to add override to all methods except those that needed to be final.
Thanks for a really informative bit. That should be required reading on a regular basis for those working with lasers, and definitely before they are used the first time.
First, per others, the red cross symbol was created prior to J&J's trademark application by an international treaty organization, one that the US signed even. Second, this symbol goes back at least a thousand years, so I'm not sure that you can usurp a commonly used symbol for your own personal trademark, much like Microsoft is still trying futilely to claim the word "windows" as a trademark.
J&J should never have been granted a Trademark on this in the first place, everything after that is irrelevant.
I'll check it out as soon as I have time. I suspect it doesn't have the battery life, the lightweight weight, or the screen resolution either, just from initial guesses, but I'll follow up and check it out. (I'm assuming those from a friend that recently bought an HP laptop, but I don't know the model)
I'll give you this though, when I bought my MBP, I bought the 15" refurb'd unit that came in just under $2K. It came with the C2D 2.33GHz and 2GB RAM and the X1600. It was a good deal then, as this same configuration is easily fetching $1700 on ebay. I'll probably sell this one and get a new one when Leopard comes out for that very reason. It'll be cheaper to buy a new one with all the software than it will to upgrade all the software, given the increases in specs.
Well, even if they don't know how to properly secure something, I giuess they at least know improper security when they see it. They've certainly seen a lot of different ways to do it badly!
Maybe that will encourage web designers to not use JS where it's not needed. There's usually little to no reason to have a JS button anyways, for submitting information.
Secondly, several sites have a "You need Javascript Enabled to visit this site" display that's overwritten by JS content. That option takes care of alerting the common Joe that they need to enable JS for a site.
Lastly, enabling JS for a site only occurs on the first visit, so, much like AdBlock's filter list, a list of "good" sites could be provided and used to initialize NoScript if a user wished. For instance, I frequently get Flash and doubleclick.net disallowed messages on pages I visit.
BMI is a completely useless "index". It's similar to comparing a car's safety via a mass/gas mileage rating. Such a ratio may have some meaning, but most likely only comparitively between like models. Look at the human race. There's large variations in bone structure and density within the species, and muscle mass also varies significantly. A water displacement/weight ratio is much more accurrate, but still fails to take into account what exactly is being measured.
At my peak in physical fitness, I had a BMI above 30. This was with me biking 20-30 miles a day, having a 32" waist, swimming several times a week, and being over 6' tall. I'd have laughed, except when I applied for life insurance, I was dumped into the highest cost category. BMI should be debunked as the charlatan snake oil salesman it is.
I just got a HP with a 17" screen, not one but two hard drives, 160 gigs each, firewire, card reader, hdmi out, lightscribe dvd burner, nvidia 7600 graphics, core2duo, 2 gigs of ram. It even pulsates in hibernate mode;)
It cost from 600 to 1k less than a macbook pro, even before adding stuff to try and match the HP specs. Desktop machines are even worse, you can easily find stuff that just blow the macs away. I'd like to see the identity of that machine, then again, I suppose you're specing out a desktop against a laptop. Of course, you realize that the MBP now comes with the 8600 and 2GB of RAM?
Slashdot Mirror
3-4 years is the cycle life of a server, generally. In the old days, disk arrays might outlive their servers, as they were potentially many times more expensive than their servers, and in those cases, You'd probably be more likely to add a live mirror than swap disks.
Today, with SANs, most of this discussion is moot. You'd have hot spare servers ready, if you were seriously interested in uptime, complete with the last full back up (you did test your backup by doing a restore, no?) and would merely need to update the transaction logs from then to present. (Lots of assumptions there, it really depends upon the particulars of your situation.)
As for Google, I'd really love to see the breakdown of drive manufacturers and types. From my own experience with 500 or so Seagate SCSI HDs my failure rate was closer to 0.1% across 4 years (I believe we only lost 3 or 4 drives out of 500). The 200 or so variety of IDE drives in lesser servers failed at a much higher rate, especially considering we had a bad batch of Corsair IDE drives in a set of 30 Dells we bought, with 5 HDs failing during the burn in test (72 hours of continuous full load testing). All of this was during the period of 93 through 99, across several batches of servers. The SGI boxes we bought in 91 didn't suffer a single drive failure in 9 years (total of 8 drives, all SCSI II, think they were Seagate).
Back then, you bought Seagate if you wanted to sleep easy. I personally own several Seagate drives, one even from 90, that I believe still works although it's been retired for the last 4 or so years. (It was serving as a scratch disk). I've had 2 Hitachi (pre Deathstar) 18GB SCSI drives both pretty much fail after 8 years, one catastrophically. IBM drives are another story, I received a large batch of OEM/refurb drives (twenty 9GB, fifteen 18GB) that had several failures within a 3 year period. I still have 5 36 GB drives that are currently idle, having been replaced by 320GB SATA drives. (Much cheaper, quieter, cooler, and a little slower but much much larger;)
In the 90s, drive manufacturer was very important. I do not have enough data to make definitive statements about today's drives other than the less expensive IDE and SATA drives definitely don't seem as reliable or long-lived as last decade's SCSI drives.
In 2-3 years, Vista's interface will still be Vista's, and not XPs. Linux, however, will probably provide a desktop almost indistinguishable from today's XP. So, if your choice was to pay MS large sums of cash for Vista and suffer the losses due to training and lowered productivity, or switch vendors and install Linux with lowered costs and no losses in productivity and training, which way would you go? In either case, "closely mimicing" UIs and applications is not what businesses want. They want the actual software they've gotten comfortable with, and they want an OS compatible with their existing infrastructure. Not some alternative that does less or exactly the same, just in a different way. After all, the training and initial drop in productivity per worker is far more expensive to a Microsoft-only company than a Windows XP license and a copy of Office 2007. I'm not sure why you're arguing, as you've just supported my argument, and possible presented it better than my first post. Read the above with "Vista" in mind as the alternative. BTW, Office 2007 is also not as widely adopted as some would like. I'm using it at work, and I must admit that I dislike it greatly. It's not merely the UI menu changes that suck, but the entire interface just reminds me of Fisher Price.
It's all conditional. Until it fails, do you really know it's going to fail? If I knew which drive was going to fail and when, I wouldn't need RAID in the first place.
Putting in process and regular refreshes of hardware is the surest way to prevent a multi-disk loss. (Yes, you'd want to do this offline)
Your supposition is that MS will continue to sell XP. They've stated that their desire is to EOL it ASAP. The failure of Vista is apparent in the undesired extended sales cycle of XP demanded by customers.
Those customers know that XP will end. They also know that barring an asteroid or hell freezing over, the main UI issues that made them not want Vista in the first place will not be available in 2-3 years either. What will be available are most likely very close clones of the apps they want on OSes that mimmic their known UIs. It won't be a MS OS, but most likely Linux varieties.
Such is how monopolies screw up and lose their market in the digital age.
Right, but when you build production systems, your mirror halves will be from different lots, and you might even preemptively swap half out on a time schedule, depending upon how crucial your data is. So the odds of 2 drives in a mirror set failing simultaneously, or near enough to affect the full data set becomes astronomical.
...The thing is that Microsoft isn't losing money on Vista. On the contrary. They're making good money off of both their new and their old product, and having two products on a market supplementing eachother like XP and Vista do is only a positive thing for Microsoft. It helps retain the userbase through to the next generation of the OS. Actually, they are losing money on Vista. Failure to upgrade to Vista produces a cascading failure to upgrade other products (Office, some of the DRM products) and also loses them mindshare. In case you haven't been following the magazines that cover these types of things, CIO's not only are now aware of alternatives to MS, but are actively considering the potential to rolling them out instead of Windows. This year is the first year that CIOs responded with the potential for rolling out Macs. That's huge news.The reasons for this? The utter and abject failure of Vista. Vista means retraining. Vista means large hardware upgrade costs. Vista means a whole slew of new IT issues. When taken in that context, alternatives that seemed untenable just a year ago all of a sudden start looking attractive.
Will MS go broke? Naah. But they're certainly not retaining users (their only real goal since they're a monopoly). What we are seeing in retrospect is the peak of MS about 5-7 years ago. First Linux took the datacenter, and now a combination of IE dropping to under 75% marketshare coupled with increased interest in alternative OSes and office suites and the recent shenanigans of DirectX all weaken MS's monopoly hold on the desktop.
I think you've missed the point - RAID 1+0 allows you to fail up to n drives (with 2 n needed to build the array).
Additionally, should a drive fail, rebuilding will only marginally affect your performance, degrading it by a fraction compared to a RAID 5/6 rebuild. (Only 1 drive is affected out of your stripe set, the rest perform at peak operational speed)
I'm curious, what exactly was contained in the last EULA from MS that scared you so? I've pretty much ignored MS EULA's as mere crap since about 2000, and don't bother reading them at all anymore as they are closer to bathroom stall adolescent scribbles than any actual binding legal document. (Hint - prove I read it and more importantly, agreed to it....)
That's not to say that MS EULAs shouldn't turn people to other alternatives. Considering how abusive their EULAs are, what would make one think that they would be treated any better in any other interaction with MS?
Operator overloading is a decision that makes or breaks the java language for those that write highly arithmatic code. Some prefer to see their standard symbols in code, and find the sort of code you presented unacceptable. I'm not in either camp on that one, since there are certainly very good and fast special purpose libraries in C, FORTRAN and others that would probably suit you better. I've not had the need to research Java's math capabilities, other than I know that a lack of understanding of Java's internals can create some incredibly slow code compared to what is achievable. (Mainly in multi-dimensional matrix algebra).
This is a non-argument. Until they're a monopoly, you can't accuse them of being a monopoly. Look at the Hunt brothers, they tried to corner to silver market. They failed, went broke, and then were prosecuted for some of their illegal practices to add insult to injury.
You can't convict someone for something they might become before they become it, especially if they become one legally. (Although I'm not aware of a single monopoly that got there legally, unless you count those that were established as monopolies via government fiat).
C# is the antithesis of a decent language. It's a wrapper for java-esque syntax and rules on top of C.
And let's not forget that if a library is written and a class has a method not declared with the "override" keyword, it's equivalent to a Java library with a method declared "final". I'm sure libraries have never been inadvertently made useless in C# because someone was too lazy to add override to all methods except those that needed to be final.
That's just one major flaw with C#.
perhaps we could load more blocks outside of the immediate view? You know, like, perhaps, just maybe, caching them with an intelligent loader?
But I guess that might have been too forward thinking.
And this differs from concepts like Google Maps how? (which also is not an original idea, btw)
but generally too lazy... thus the ready supply of mods for the obvious link.
But how can you license a commonly used symbol? Can you license a crucifix? An ankh? A star of David? A pentagram?
Thanks for a really informative bit. That should be required reading on a regular basis for those working with lasers, and definitely before they are used the first time.
First, per others, the red cross symbol was created prior to J&J's trademark application by an international treaty organization, one that the US signed even. Second, this symbol goes back at least a thousand years, so I'm not sure that you can usurp a commonly used symbol for your own personal trademark, much like Microsoft is still trying futilely to claim the word "windows" as a trademark.
J&J should never have been granted a Trademark on this in the first place, everything after that is irrelevant.
I'll check it out as soon as I have time. I suspect it doesn't have the battery life, the lightweight weight, or the screen resolution either, just from initial guesses, but I'll follow up and check it out. (I'm assuming those from a friend that recently bought an HP laptop, but I don't know the model)
I'll give you this though, when I bought my MBP, I bought the 15" refurb'd unit that came in just under $2K. It came with the C2D 2.33GHz and 2GB RAM and the X1600. It was a good deal then, as this same configuration is easily fetching $1700 on ebay. I'll probably sell this one and get a new one when Leopard comes out for that very reason. It'll be cheaper to buy a new one with all the software than it will to upgrade all the software, given the increases in specs.
Maybe that will encourage web designers to not use JS where it's not needed. There's usually little to no reason to have a JS button anyways, for submitting information.
Secondly, several sites have a "You need Javascript Enabled to visit this site" display that's overwritten by JS content. That option takes care of alerting the common Joe that they need to enable JS for a site.
Lastly, enabling JS for a site only occurs on the first visit, so, much like AdBlock's filter list, a list of "good" sites could be provided and used to initialize NoScript if a user wished. For instance, I frequently get Flash and doubleclick.net disallowed messages on pages I visit.
Maybe NoScript will become a standard FF component. It should be anyways.
The only people that do not screw-up are those that do nothing.
OK, time for some coffee, that was far too Confuscious-like...
BMI is a completely useless "index". It's similar to comparing a car's safety via a mass/gas mileage rating. Such a ratio may have some meaning, but most likely only comparitively between like models. Look at the human race. There's large variations in bone structure and density within the species, and muscle mass also varies significantly. A water displacement/weight ratio is much more accurrate, but still fails to take into account what exactly is being measured.
At my peak in physical fitness, I had a BMI above 30. This was with me biking 20-30 miles a day, having a 32" waist, swimming several times a week, and being over 6' tall. I'd have laughed, except when I applied for life insurance, I was dumped into the highest cost category. BMI should be debunked as the charlatan snake oil salesman it is.
BMI is a completely useless "index". It's similar to comparing a car's safety via a mass/gas mileage rating.
It cost from 600 to 1k less than a macbook pro, even before adding stuff to try and match the HP specs. Desktop machines are even worse, you can easily find stuff that just blow the macs away. I'd like to see the identity of that machine, then again, I suppose you're specing out a desktop against a laptop. Of course, you realize that the MBP now comes with the 8600 and 2GB of RAM?