I have no idea how much disk space my firm has, but I did hear an apocryphal tale of installing multiple truckfuls of disks every week pretty much indefinitely (now, of course, older smaller disks are also being removed, but even if it's one for one the service life of enterprise disks means the total is continuously growing). But the firm and the total space can't be disclosed. I'm not trying to make any claims - it could well be smaller than the five mentioned, but the point is nobody knows. I'm sure lots of other firms have very big SANs too.
It did come across a bit boastful, didn't it? Sorry about that. I just thought that if I said "I've got a REALLY fast computer, this still runs slow" it wouldn't get the point across.
I'm sitting at a Sun Ultra 40 running Windows XP 64-bit edition. 4 CPU cores. 8GB RAM. NVIDIA QuadroFX 3450 graphics. I'm on a gigabit connection to a major internet link (major financial news and data company). Even so, ajaxWindows recreates that, sticky, gummy, nasty feeling of running X-Windows when the machine isn't really powerful enough and dragging solid windows around is still a reckless waste of horsepower.
The early 90s called, they want their SPARCstation 5s back.
and don't forget Descent 3 and Rune. I don't know if you can still get them, but I definitely had some fun with those. I loved Rune in particular. I completed it on my Athlon 750MHz/GeForce 256 PC back when that was a pretty fast machine.
Not even close. the JVM does not implement a filesystem, or a network stack, or virtual memory management system, or any device drivers, or threading, or low-level graphics operations, or...
Java is fine, but don't confuse it with an entire modern operating system.
Thank you! I knew I'd seen a beowulf in a lunchbox in Linux Journal, and I haven't even been reading it for a couple of years or so.
Some pure speculation now : I also wonder whether this "Microwulf" would even outperform a dual dual-core PC. One can buy a motherboard supporting 2 dual-core Opterons for $500, dual-core opterons are about $200 each, and 8GB fast RAM should be about $500 or so (8x1GB). If we allow the same ($2500) budget, I bet you can make a 4-core PC that outperforms this cluster on EVERYTHING. Each CPU-CPU communication is either on chip or over hypertransport - each massively faster than gigabit ethernet. If you really want a cluster programming model use a virtualisation solution to get 4 copies of the OS running on the box and lock each one to its own core.
I found that mounting using the noatime option on Solaris 8 and later was good for performance on filesystems which I mirrored via rsync. For example we had a cvs server that was struggling a little to keep up because rsyncing the files out to the mirrors was expensive. Changing its filesystem options to include noatime meant that a lot of the unnecessary I/O disappeared (previously I guess the very act of checking if a file had changed would have incurred a disk write, and we were syncing every all the files many times an hour). Other operations (e.g. cvs check-in, check-out) would normally operate on a smaller range of files, rsync was the worst case.
This is one of the best points I've read in this debate. I wet from a RAZR to an iPhone. The RAZR has a few features that the iphone lacked (e.g. taking video), so one might think it was a downgrade, but even if they were on par feature for feature, the iPhone gets more done for me, because when I use it is fun and productive, and when I used my RAZR it made me want to smash it on my desk.
Tony is a replicant - an organised crime model for optimum viciousness. Everything went black because of the built-in deactivation circuit. He actually is the 7th of the replicants that escaped in the back story of Blade Runner and will appear in the upcoming 27-disc "Blade Runner, Binge and Purge" box set (with apologies to Metallica). He is shown whacking Gaff and remarking "F**king origami, fuggedaboutid!". Skynet sent him back in time to defeat the Architect before he could construct the Matrix, but he took a detour into some rather lucrative crime.
This was not the ending you're looking for. You don't need to see this TiVo. You can go about your business. Move Along.
My feeling is any debate like this is mostly dominated by people who ran into problems. I'd like to provide just one data point in the opposite camp.
I got a Sony Bravia 42" LCD TV late last year which does 1080p. I connect my DVD player to it with a 15ft HDMI cable which cost me about $100 at the Sony Store. Quite an expensive cable, however that setup has worked perfectly for me ever since.
More recently I got an AppleTV and hooked it up to the second HDMI port on the TV using a short $20 XtremeMac HDMI cable. That setup also worked perfectly.
So that's two cable lengths, two cable vendors (I'd say makers but who knows who really made either cable), two signal sources of vastly different types from two manufacturers. Even two resolutions (the DVD player seems to be nice at 1080i upscaled, the AppleTV seems better at 720p). Everything works seamlessly. My only complaint is that switching inputs on the TV is slow.
This doesn't prove anything, I know that. This debate is just a numbers game anyway. Even the best regarded, most reliable, most consumer friendly technologies sometimes have problems, it's just the nature of engineering. What matters is not whether the technology is perfect, but that your odds of success should be high.
HDMI has been a breeze for me, and replaced a thick bundle of analog cables snaking around the fireplace in our living room.
By the way, my actual Apple TV failed within 48 hours. The replacement is fine however. Technology remains a crapshoot at consumer prices.
I can't match 20 years in IT, but I do have 18 years programming for a living (i considered saying "programming professionally" but that would be a very bold claim regarding my career). I think if I had no commitments (family to feed, mortgage, car payment etc) I would still like to program but I'd just like to work on better code. No matter which job I'm in, which company, which continent, I have somehow ended up dealing with a lot of legacy junk, often written by people in an awful hurry many years ago, some of whom i never even met. I understand that that's where the money is for many people, since "legacy junk" can also be the same thing as "established product with healthy userbase, making boatloads of cash", however I would just really like to get away from that at some point. I think the "plan" such as it is, would be to win millions on the lottery and then retire, and entertain myself on some open source project that I could be happy with.
Maybe my entire view of things is wrong, but isn't strict layering responsible for most of the things that set Linux apart from less-efficient and less-secure operating systems?
I think strict layering is more of a design discipline than a direct facilitator of efficiency or security. It is supposed to be overall a better tradeoff, not simultaneously better on all fronts. Layers impose costs and limitations, but done well they're often worth it.
The most "efficient" operating system, in terms of, say, maximum filesystem throughput might well be a monolithic "filesystem O/S" which is all compiled together into one big lump, but that would likely have worse security. The layering that Linux does have seems to have successfully permitted very good performance, very good security, maintainability, extensibility, virtualisation, etc etc all together in one design. This is amazing and I run my server using slackware very gratefully.
However, the Linux kernel itself is not as internally layered as some micro-kernel operating systems, or perhaps the barriers between the layers are fliimsier. When challenged on this Linus was famously quite confident that his "layering violation" was worth it. Go read the discussions between him and Tanenbaum on that topic (monolithic vs micro-kernel design for operating systems).
In the world of car magazines, here in the US, the mainstream ones can be had for under a dollar per issue if you subscribe. Meanwhile Evo magazine is about $10/issue because it's independent. It does have US readers (every now and then they even write in complaining about anti-american prejudice), but I'm sure they could sell more over here if they made a US edition and cosied up to the industry in Michigan. Instead they tell it like it is slating cars that neglect driver involvement, no matter how many bhp or cupholders they may have. Even for Porsche and Audi who advertise heavily in the magazine, their bloated and ugly SUVs are regularly blasted (although any progress is duly noted). I wouldn't say Evo magazine is "fair" (who on earth wants that?) but the biases come from the personalities of the staff and not from a sales goal at GM headquarters.
For example, they were even fairly hard on the Bugatti Veyron, leaving it to the designer of the McLaren F1 to point out that where throttle response is key the Veyron would be woefully deficient (consider a marginal overtaking manoeuver, the F1 would be your best bet, the Veyron would still be spooling up its turbos when you wanted to get back in your lane).
I don't know how we can get out of the vicious circle of declining expectations.
I know nobody believes it, but there was a time when beta versions were called betas, and Version 1.0 meant a product that was finally finished, SQA-ed, and working.
Users have a right to a version 1.0 that works. Shrugging your shoulders and saying "hey, what do you expect, it's version 1.0" wouldn't be tolerable in any other product.
Maybe it's true in some other industry, but shoddiness is business as usual in the PC market. QDOS was a quick 6 week hack job knocking off CP/M before Bill Gates bought it and turned it into MS-DOS.
If you don't prioritise time to market (first mover advantage) over nearly everything else you are dead in the mass market software industry.
To put a more favorable spin on this, the computer industry spent most of its first few decades in a race to the bottom (cost-wise). Each "revolutionary" new paradigm can be viewed as an 20%/80% tradeoff approximation of the previous generation. 80% of the functionality, 20% of the cost. Mainframes were smacked by minicomputers, workstations smacked minicomputers, PCs gutted workstations. PCs ruled for a long time, but now they're under threat from a new type of computer which is "Anything-with-a-browser-and-net-access" (AWABANA) and it's much cheaper even than a pc - it comes built into your phone, your games console, perhaps your TV, car etc. Sure there are still some apps that don't work in an AWABANA, but if you ignore the rest and stick to those you are back in expensive niche computing. So the bits of Office, Outlook etc that free browser based apps DON'T do will shortly seem about as relevant as mainframes.
This is true in a sense but the most important thing to understand is that trains are in a sorry state in this country (the US) because nobody wants to pay for them. It takes political will to get a modern train system built. It takes heavy investment and a lot of legislaton to get the land etc. It might be possible to run a train system commercially in this country, but building one is, in my opinion, impossible.
In my view it's not Amtrak's fault, it's just a cultural difference. We (the taxpayers, the landowners, the nature preservationists etc etc) don't support a train system enough to get a good one. That's it really.
I had a game for my Acorn Electron called Blagger. It was a 2d platform game. This was back in the 80s.
You wandered around rooms negotiating obstacles and (I think) stealing things. You had to go as fast as possible to complete the level before the time ran out. Time available was a thin vertical line to the right of the display which started off the full height of the graphical area and shrunk vertically downwards as you used time. When you completed the level the remaining time bar would shrink much faster giving you bonus points proportional to the remaining length, so the quicker you did it the more points you got.
Just one time, however, I did a level where I think I completed it simultaneously with running out of time, as the time bar decremented... from zero downwards. I scored a large number of points.
Slashdot Mirror
Is your firm Goldman Sachs as per your resume on your website?
Not since they let me go in 1998
I have no idea how much disk space my firm has, but I did hear an apocryphal tale of installing multiple truckfuls of disks every week pretty much indefinitely (now, of course, older smaller disks are also being removed, but even if it's one for one the service life of enterprise disks means the total is continuously growing). But the firm and the total space can't be disclosed. I'm not trying to make any claims - it could well be smaller than the five mentioned, but the point is nobody knows. I'm sure lots of other firms have very big SANs too.
If you hate getting ripped off in Britain and think things are much better in the US, well then move to the US.
You can earn a lot more and buy a lot more electronic gadgets.
How do I know all this? Well, because that's what I did, ten years ago.
Is it all quite that simple? Well, no, but this IS slashdot.
Paul Baran is one of the fathers of the Internet and says something along the lines of "each man added a brick and says 'I built a cathedral'"
If you search Google books for "Where Wizards stay up late" you can find the exact quote (but I can't cut and paste it).
You're so badass!
It did come across a bit boastful, didn't it? Sorry about that. I just thought that if I said "I've got a REALLY fast computer, this still runs slow" it wouldn't get the point across.
I'm sitting at a Sun Ultra 40 running Windows XP 64-bit edition. 4 CPU cores. 8GB RAM. NVIDIA QuadroFX 3450 graphics. I'm on a gigabit connection to a major internet link (major financial news and data company). Even so, ajaxWindows recreates that, sticky, gummy, nasty feeling of running X-Windows when the machine isn't really powerful enough and dragging solid windows around is still a reckless waste of horsepower.
The early 90s called, they want their SPARCstation 5s back.
and don't forget Descent 3 and Rune. I don't know if you can still get them, but I definitely had some fun with those. I loved Rune in particular. I completed it on my Athlon 750MHz/GeForce 256 PC back when that was a pretty fast machine.
The OS is just a bootloader for the Java VM.
Not even close. the JVM does not implement a filesystem, or a network stack, or virtual memory management system, or any device drivers, or threading, or low-level graphics operations, or ...
Java is fine, but don't confuse it with an entire modern operating system.
Thank you! I knew I'd seen a beowulf in a lunchbox in Linux Journal, and I haven't even been reading it for a couple of years or so.
Some pure speculation now : I also wonder whether this "Microwulf" would even outperform a dual dual-core PC. One can buy a motherboard supporting 2 dual-core Opterons for $500, dual-core opterons are about $200 each, and 8GB fast RAM should be about $500 or so (8x1GB). If we allow the same ($2500) budget, I bet you can make a 4-core PC that outperforms this cluster on EVERYTHING. Each CPU-CPU communication is either on chip or over hypertransport - each massively faster than gigabit ethernet. If you really want a cluster programming model use a virtualisation solution to get 4 copies of the OS running on the box and lock each one to its own core.
Next thinhg you'll tell me The X Journal is under threat...
I found that mounting using the noatime option on Solaris 8 and later was good for performance on filesystems which I mirrored via rsync. For example we had a cvs server that was struggling a little to keep up because rsyncing the files out to the mirrors was expensive. Changing its filesystem options to include noatime meant that a lot of the unnecessary I/O disappeared (previously I guess the very act of checking if a file had changed would have incurred a disk write, and we were syncing every all the files many times an hour). Other operations (e.g. cvs check-in, check-out) would normally operate on a smaller range of files, rsync was the worst case.
Oh come on! Even if you don't like Neuromancer, finding worse pseudo-intellectual garbage is not at all hard, so how can it be the "worst sort"?
In future in any slashdot discussion where you intend to trash the iPhone, please include the following information :
Have you actually used an iPhone : yes [] no []
thank you
This is one of the best points I've read in this debate. I wet from a RAZR to an iPhone. The RAZR has a few features that the iphone lacked (e.g. taking video), so one might think it was a downgrade, but even if they were on par feature for feature, the iPhone gets more done for me, because when I use it is fun and productive, and when I used my RAZR it made me want to smash it on my desk.
Brian Lam discussed this in a pretty level-headed way:
Brian's great, I like his writing style and he even replies to emails from complete strangers...
Tony is a replicant - an organised crime model for optimum viciousness. Everything went black because of the built-in deactivation circuit. He actually is the 7th of the replicants that escaped in the back story of Blade Runner and will appear in the upcoming 27-disc "Blade Runner, Binge and Purge" box set (with apologies to Metallica). He is shown whacking Gaff and remarking "F**king origami, fuggedaboutid!". Skynet sent him back in time to defeat the Architect before he could construct the Matrix, but he took a detour into some rather lucrative crime.
This was not the ending you're looking for. You don't need to see this TiVo. You can go about your business. Move Along.
My feeling is any debate like this is mostly dominated by people who ran into problems. I'd like to provide just one data point in the opposite camp.
I got a Sony Bravia 42" LCD TV late last year which does 1080p. I connect my DVD player to it with a 15ft HDMI cable which cost me about $100 at the Sony Store. Quite an expensive cable, however that setup has worked perfectly for me ever since.
More recently I got an AppleTV and hooked it up to the second HDMI port on the TV using a short $20 XtremeMac HDMI cable. That setup also worked perfectly.
So that's two cable lengths, two cable vendors (I'd say makers but who knows who really made either cable), two signal sources of vastly different types from two manufacturers. Even two resolutions (the DVD player seems to be nice at 1080i upscaled, the AppleTV seems better at 720p). Everything works seamlessly. My only complaint is that switching inputs on the TV is slow.
This doesn't prove anything, I know that. This debate is just a numbers game anyway. Even the best regarded, most reliable, most consumer friendly technologies sometimes have problems, it's just the nature of engineering. What matters is not whether the technology is perfect, but that your odds of success should be high.
HDMI has been a breeze for me, and replaced a thick bundle of analog cables snaking around the fireplace in our living room.
By the way, my actual Apple TV failed within 48 hours. The replacement is fine however. Technology remains a crapshoot at consumer prices.
I can't match 20 years in IT, but I do have 18 years programming for a living (i considered saying "programming professionally" but that would be a very bold claim regarding my career). I think if I had no commitments (family to feed, mortgage, car payment etc) I would still like to program but I'd just like to work on better code. No matter which job I'm in, which company, which continent, I have somehow ended up dealing with a lot of legacy junk, often written by people in an awful hurry many years ago, some of whom i never even met. I understand that that's where the money is for many people, since "legacy junk" can also be the same thing as "established product with healthy userbase, making boatloads of cash", however I would just really like to get away from that at some point. I think the "plan" such as it is, would be to win millions on the lottery and then retire, and entertain myself on some open source project that I could be happy with.
Chris
-helpful and constructive as always
Maybe my entire view of things is wrong, but isn't strict layering responsible for most of the things that set Linux apart from less-efficient and less-secure operating systems?
I think strict layering is more of a design discipline than a direct facilitator of efficiency or security. It is supposed to be overall a better tradeoff, not simultaneously better on all fronts. Layers impose costs and limitations, but done well they're often worth it.
The most "efficient" operating system, in terms of, say, maximum filesystem throughput might well be a monolithic "filesystem O/S" which is all compiled together into one big lump, but that would likely have worse security. The layering that Linux does have seems to have successfully permitted very good performance, very good security, maintainability, extensibility, virtualisation, etc etc all together in one design. This is amazing and I run my server using slackware very gratefully.
However, the Linux kernel itself is not as internally layered as some micro-kernel operating systems, or perhaps the barriers between the layers are fliimsier. When challenged on this Linus was famously quite confident that his "layering violation" was worth it. Go read the discussions between him and Tanenbaum on that topic (monolithic vs micro-kernel design for operating systems).
Chris -fan of Linux AND ZFS both at the same time
Does integrity exist anywhere?
In the world of car magazines, here in the US, the mainstream ones can be had for under a dollar per issue if you subscribe. Meanwhile Evo magazine is about $10/issue because it's independent. It does have US readers (every now and then they even write in complaining about anti-american prejudice), but I'm sure they could sell more over here if they made a US edition and cosied up to the industry in Michigan. Instead they tell it like it is slating cars that neglect driver involvement, no matter how many bhp or cupholders they may have. Even for Porsche and Audi who advertise heavily in the magazine, their bloated and ugly SUVs are regularly blasted (although any progress is duly noted). I wouldn't say Evo magazine is "fair" (who on earth wants that?) but the biases come from the personalities of the staff and not from a sales goal at GM headquarters.
For example, they were even fairly hard on the Bugatti Veyron, leaving it to the designer of the McLaren F1 to point out that where throttle response is key the Veyron would be woefully deficient (consider a marginal overtaking manoeuver, the F1 would be your best bet, the Veyron would still be spooling up its turbos when you wanted to get back in your lane).
Does anyone have pictures of this "horrible" video playing on a TV so people can actually make a judgment
Why yes, I have a 160x160 pixel animated GIF taken on my cellphone and then downscaled using 'xv' and it clearly shows some artefacts.
Chris
-always helpful
I don't know how we can get out of the vicious circle of declining expectations.
I know nobody believes it, but there was a time when beta versions were called betas, and Version 1.0 meant a product that was finally finished, SQA-ed, and working.
Users have a right to a version 1.0 that works. Shrugging your shoulders and saying "hey, what do you expect, it's version 1.0" wouldn't be tolerable in any other product.
Maybe it's true in some other industry, but shoddiness is business as usual in the PC market. QDOS was a quick 6 week hack job knocking off CP/M before Bill Gates bought it and turned it into MS-DOS.
If you don't prioritise time to market (first mover advantage) over nearly everything else you are dead in the mass market software industry.
To put a more favorable spin on this, the computer industry spent most of its first few decades in a race to the bottom (cost-wise). Each "revolutionary" new paradigm can be viewed as an 20%/80% tradeoff approximation of the previous generation. 80% of the functionality, 20% of the cost. Mainframes were smacked by minicomputers, workstations smacked minicomputers, PCs gutted workstations. PCs ruled for a long time, but now they're under threat from a new type of computer which is "Anything-with-a-browser-and-net-access" (AWABANA) and it's much cheaper even than a pc - it comes built into your phone, your games console, perhaps your TV, car etc. Sure there are still some apps that don't work in an AWABANA, but if you ignore the rest and stick to those you are back in expensive niche computing. So the bits of Office, Outlook etc that free browser based apps DON'T do will shortly seem about as relevant as mainframes.
Chris (some straying off-topic acknowledged)
This is true in a sense but the most important thing to understand is that trains are in a sorry state in this country (the US) because nobody wants to pay for them. It takes political will to get a modern train system built. It takes heavy investment and a lot of legislaton to get the land etc. It might be possible to run a train system commercially in this country, but building one is, in my opinion, impossible.
In my view it's not Amtrak's fault, it's just a cultural difference. We (the taxpayers, the landowners, the nature preservationists etc etc) don't support a train system enough to get a good one. That's it really.
This train cruised at significantly higher than the max terminal velocity of an NHRA top fuel dragster.
I had a game for my Acorn Electron called Blagger. It was a 2d platform game. This was back in the 80s.
You wandered around rooms negotiating obstacles and (I think) stealing things. You had to go as fast as possible to complete the level before the time ran out. Time available was a thin vertical line to the right of the display which started off the full height of the graphical area and shrunk vertically downwards as you used time. When you completed the level the remaining time bar would shrink much faster giving you bonus points proportional to the remaining length, so the quicker you did it the more points you got.
Just one time, however, I did a level where I think I completed it simultaneously with running out of time, as the time bar decremented ... from zero downwards. I scored a large number of points.
Never could do it again.