I don't really care whether Linux is good enough for the masses these days, other than to satisfy intellectual curiosity. It's good enough for me, my family and friends, and I appreciate that. I remember the days when uber hackers (i.e. hacking addicts) were the only people I knew running Linux.
It wasn't a matter of pure IQ per se. I knew several extraordinary Linux hackers who failed calculus multiple times (eventually giving up) but ran Linux as a desktop OS way back in 1997 and earlier. It was IQ and perhaps an early introduction by a parent or peer to understanding computing minutiae, along with the ability to just grok Linux. I was always in awe of them.
The bar has been lowered repeatedly since then. Only 9 years later or so, a version of Linux was dumbed down enough to suit my need for a tool that didn't require too much hacking to get to work. It was not only the distribution, it was also the moderation policy and community of ubuntuforums that enabled me to do any necessary hacking with a minimum of pain - a newbie had asked the question once and searching revealed the answer.
I fully expect that Microsoft will bundle their operating for next to nothing on commodity hardware in order to maintain mindshare. I will be surprised if MS doesn't manage to muscle their way to the point where their monopoly is maintained. There comes a point on the bell curve where no one does their own install, no matter how easy it is. This is what protects MS, because the fraction under that bell curve incapapable/unwilling to install is a clear majority.
Meanwhile, Linux distributions (probably Ubuntu based) will keep getting better to the point that they do everything a user could want. Commoditization of PCs and componentry means that supporting the bleeding edge with drivers will become less important as cost, suitability and other factors become more important than pure speed. If people want to switch, it will get easier. Those who see the advantage and have the skills to install an OS will do so if they perceive it to be in their interest. Once the growth in mindshare from those who can install and will benefit from a Linux install is maxed out (probably subject to network effect), growth will come from one of at least three sources.
1) Reduction in Profit Margin * Volume for Windows OEM sales means that MS is less able to dictate terms to hardware manufacturers, and Linux factory installs start to proliferate. 2) Linux becomes cool, and power users start installing it on the computers belonging to the rest of the bell curve. 3) A third party decides it is in their financial interest to back hardware manufacturers into pre-installing Linux, stymying the best efforts of MS.
Anyway, I look forward to the release of Hardy Heron. I expect it to be better than the previous releases and to "just work" for a long time. If you are going to install Linux on a friend's computer, I expect that this will be very stable and so reduce need for your support being a Long Term Support (LTS) release.
"Oh, and I may be a power user, but I'm also a gamer, and I want games that run natively on Linux. Besides a tiny subset of games, that's not happening until Linux is the average desktop."
Oh great, when that happens I'll truly never get any work done. Hopefully that day is soon followed by an Ubuntu variant that disables any access to games in the repos.
"People have said as a joke that OpenOffice.org or similar programs will take over once they have their own clippy, but may a true word is said in jest."
Thanks for the link, that was interesting to read.
Another thing to realize is that security companies and consultants have an inherent desire to inflate the size and probability of any security risk. The more fear they can build in the client, the bigger the more expensive the service they can sell.
"If MS really was serious about making life easy for the customer they would've bundled MS Office and a decent mailing client to name a few."
If they did want to bundle Office, would that be against the law? Just wondering. I can see how it could be a viable strategy for them if they decided to embrace the low margin for the sake of higher volume. If the choice is between death and bundling a major cash cow, I think they'd sacrifice the cow.
"Neither can Dell, HP or any other hardware manufacturer. This trend impacts them every bit as much as Microsoft, although on the whole I think hardware manufacturers should be able to adapt easier than Microsoft. "
I think both MS and the hardware manufacturers have known about this for a long time. The eternal upgrade cycle was driven by obsolescence. The moment that faster CPUs would not obsolete older, slower CPUs because of the lack of killer apps requiring faster CPU speed, the profits would drop off. Changes would have to be made to the business model. It could go several ways, and no doubt there have been contingency plans drawn up on this very subject.
If you've ever seen "Letters from Iwo Jima", you will see that MS, Intel and the like are in much the same position as General Kuribayashi, facing the inevitable defeat at the hands of commoditization. If they are smart, their goal will be to bleed the consumer for as much as they can on the way down.
Intel will be trying to beat Via and AMD with just enough performance increase with a reduced power requirement, but not to produce something so good that they can't make something a bit better in another year or two. Eventually I would expect them to look for a way to start increasing obsolescence in other ways. e.g. CPUs will be designed to fail after a minimum number of years. It's tough though, as the technology plateaus they won't be the only one manufacturing CPUs and by doing this they will get a reputation for reduced reliability which will feed sales of the competition.
Another way is of course to include a Microsoft operating system that will be overcome with malware given enough time. This is probably more likely.
What has Microsoft done? They have waited until someone forced their hand. Microsoft has retooled XP and is ready to sacrifice their margins for increased volume. Once the market really takes off, I would expect them to drop the margins of Office as well in order to properly compete with something like Ubuntu that includes OpenOffice. If they don't do this, Ubuntu will be perceived as more useful and gain adherents. The last thing MS wants is a large consumer base happy to buy from a hardware manufacturer who is unwilling to sell MS and Linux systems for the same price (likely by paying MS a drawback for Linux installs).
If I were a greasy monopolist in the shoes of MS, I would see Asus and cut a deal with them. You either install XP on some of your systems and pay us a small fee for ANY sale of a computer (including that of Linux, perhaps even more in the case of Linux), or we will partner with your nearest competitor and subsidize them until you are making no money on your Eee PC. Do we have a deal?
This bluff might be called. I can see that it would be in the long term interest of a country like China to subsidize Linux in a price war until they gained enough mindshare, and then their hardware manufacturers would be free of the Microsoft Tax.
Unfortunately, the biggest problem Ubuntu faces is being in opposition to the long term interest of manufacturers, software vendors, and MS. It might be possible that given enough marketshare, spyware on Ubuntu will become rampant (more money in exploiting vulnerabilities than finding fixes) and people will buy a new computer rather than reinstall. This would be good for the hardware manufacturers and ironically increase support of Ubuntu.
Traditional software vendors (those that sell the install rather than the support) must see the repository as a threat, since the repository is just so much easier and also safer. Most of the large vendors devoid of FOSS religion (e.g. Adobe) will see the threat of their mindshare being eroded more easily via the repo and instinctively avoid encouraging Linux by making their software available. (The exception is the gaming company that sells content that is much more expensive to create than the Open Source community can compete with. e.g. WoW.)
"Had he not laid all his cards out on the table, he might have got the fight he was looking for, and showed Monster they can't bully everyone around. Part of me is afraid that won't happen now."
Just because he wrote that he was itching for a fight doesn't mean that he was actually itching for one. It's a good bluff though. If he enjoyed litigation so much, chances are he'd still be doing it. Still, a cool letter.
The reason this hasn't been tried commercially is that there aren't commercial quantities of energy available. A bit of googling yields a figure of 250kWh per strike, or less than my personal energy bill for a quarter. It's a bit like shuffling your shoes on the floor to generate static electricity - most of the energy goes towards heating up your shoes, not transfering electrons.
I never had the patience for running a linux desktop despite trying multiple times since 1997 or so. You must have been one of those rare people with enough patience. My opinion changed about a year ago, when I installed Ubuntu for the second time. If I had left Linux in 2003, I would have had the same opinion as you.
Any maintenance issues I've had have lasted less than half an hour a month. It is certainly less of a pain to use than Windows was, a bit of work up front for (so far) a year of working extremely smoothly. The only issues have been reinstalling video drivers each time the kernel upgrades. The fix takes less than 10 minutes. That is the ONLY routine maintenance I have to do (other than click and install updates when they come, which would be the same on a Mac).
Since I don't edit video, I can't comment on that aspect of Ubuntu.
"I don't understand why modern cars get such lousy mileage."
Looking at first order causes... cars get poor mileage because energy is lost in: 1) Having to brake. 2) Having to move air around, which eventually ends up as heat. 3) Deforming tyres.
1) is as much a problem of infrastructure design (frequent stops that can't be anticipated) as it is a problem of lack of driver training (for stops that can be anticipated) and obese vehicles.
2) is a problem of convenience and marketability. (Try implementing stylistic obsolescence when every year the shape must stay the same - a modified teardrop.) It is also a problem of overly high speed limits (legislation).
3) is not really much of a problem in comparison with 1 & 2. Can be solved by pumping up tyres and dealing with the bumpier ride.
Of course, all 3 come at the problem from an engineering perspective. If you look at the problem from an oil company perspective and ask the question "How can I ensure that my mode of transport is more convenient (by far) than all alternate forms of transport, that I get another entity (taxation) to fund the infrastructure, and that I partner/own the automobile companies such that they also extract the maximum dollars/year from consumers", you are asking the right question.
It's then a matter of engineering the situation such that no other forms of transport can compete with gasoline/diesel's high energy density and also relatively low cost/volume. And sitting back to enjoy the profits while they last.
I can't help but think that this is a stop-gap measure. I used to read up about all the various methods of silencing a computer (with the intention to implement myself) but for consumer-grade applications I'd prefer to wait for a variant of Moore's Law to do its work - the propensity for performance per watt to keep increasing until it nears whatever limits are predicted by information theory.
At that stage there will be an option to cool with no moving parts for typical desktop/laptop applications, and it will be a superior solution in all aspects compared with any combination of cooling and sound minimization.
Another welcome change will be for the idle power consumption to drop, which it certainly can.
Efficiency is inherently hard - it's all about approaching zero loss, and losses seem to almost never present a target suitable for a single silver bullet approach. That means work - killing all the losses on your loss budget one by one, and figuring how to integrate all the changes so it still works. The good thing is that once you come up with a design that lops off as many heads of the efficiency hydra as possible, it's just a matter of mass production.
"Windows is so big, with such a huge established base, that its decline will resemble that of the old IBM mainframe environment -- which is still doing fine, decades after the death of the mainframe was predicted. This ain't going to happen overnight."
It certainly won't happen overnight. There will be a period when MS is a "dead man walking". But it is all about the revenue stream. The revenue stream comes from Office and Windows. If no one (relatively speaking) is upgrading or buying new computers, Microsoft's army of programmers, marketers, lobbyists and salespeople start looking for work elsewhere.
Network effect is a double-edged sword. Great on the way up, great at consolidating a monopoly. When the rats start leaving the sinking ship, they leave in droves. Consider Lotus 1-2-3, or Amiga.
Mainframes still had a very valid use in defense and other big government type problems. Until they get a working strong AI, there will probably still be use for more high-end grunt by governments all over the world. The average consumer is a different market. That market depends on disposable income, as you made excellent reference to, and also need. Until now, CPU power has been "good, but never enough". That is changing.
We are very near the point where the choice will be between something more expensive, large, louder, requiring maintenance and with performance only a minority will actually use, or something cheap (both initial purchase price and ongoing), small, silent, no maintenance that does the job well. The latter ticks the relevant boxes. And there is not much of a profit margin in that world for MS.
It is an open question as to how many people would buy something like that at a particular price, because the upgrade will probably come as a result of a malware infestation for someone computer illiterate, or a hardware failure for someone who knows what they are doing, or the price is so low that the payback period for the electricity cost difference is good. Maybe it will be possible for MS to slash prices, push Intel into the high volume/low margin game as well, and extract the last cash from their cow. It seems like that is what they are reluctantly doing with XP on the Eee PC. It is as if GM broke down and started manufacturing 150cc motorbikes and bicycles.
I suppose it's fully possible that this strategy well executed will result in there never being a year of the linux desktop. I don't really care so long as Linux and especially Ubuntu maintains or improves over time.
I'm not sure I like the bike analogy. The ultimate security in a bike rack is still not very secure compared to a home computer installation. A professional thief can still wait until no one is there to come along and nab your bike using inexpensive tools. If he does that, it's relatively low risk.
It takes a bit more effort, marginally more cost, and you can end up with a home network that is much more secure (at the router/modem interface) than leaving your bike at any bike rack with any sort of security. There are numerous FOSS distros specifically for this application, IPCop, Monowall, PfSense, OpenBSD, Zeroshell to name a few. Of course, if you are going to go to that effort you may as well run linux/bsd and firefox with noscript, no sense turning your front door into a safe and leaving the window open.
That is an investment in time and effort that can be done once, and will likely fend off even a concerted attack. The next step up will require either government or large corporation resources. This would be comparable to leaving your bike either inside your house or inside your place of work, and never at a bike rack... but with a little larger investment up front but more convenience after you have done the work.
Also, until software routinely kills people, it will not be subject to bridge-building type regulations, and if it does, it will only be software written for those things. In fact, I believe that defense applications, airplane software, things where lives or important things are at stake do submit software to the same degree of rigorous testing by qualified people that the design and construction of a bridge would have to withstand.
Games also have the downside of being highly addictive. Some things in life just aren't easy, and require significant investment of frustration in the beginning but pay off over time. In order to be highly addictive (and hence popular), games are structured to eliminate a lot of the grind that daily life involves. The end result is that even though valuable learning may have taken place, the brain gets rewired to run back to the refuge of the addiction rather than deal with the grind and accomplish something.
Getting most of your knowledge through games (or movies, or secondary source books) also puts you at risk of learning the media creator's version of the world rather than the world as it is, or the world as primary sources had it. For example, you play Civilization and learn that Neville Chamberlain was one of the worst world leaders of all time (because that equates to a low civilization score). And being a kid, it's easier to just add that to your list of facts about the world along with most of the other stuff that seems mostly right. Another thing you "learn" is that the US is a peaceful isolationist civilization.
Very often some things are "good enough" for a long period of time. Some examples: -AK-47, built in 1947 -Subsonic passenger jets -The horse, fastest way to get around for thousands of years. -C, SQL -The car, versus the "flying car".
Why development of something plateaus has everything to do with limits to optimization, efficiency, network effect, cost benefit analysis, diminishing marginal returns, return on investment, political and legislative situations. Complacency and familiarity are important, but there are certainly many, many more factors involved.
Sure I'd like an infinitely fast CPU, a commercially viable fusion reactor and a flying car while I'm at it. Some things are hard, and breakthroughs are difficult to schedule.
"As gigahertz race is over and mobility takes over, size, power and price are becoming more important than performance"
Don't forget the silence and no-maintenance aspect that going completely passively cooled and solid-state affords you. And even in a desktop system all the other issues apart from computing speed become important once you experience the difference.
Such other concerns are the whole raison d'etre of silentpcreview.com. There have been some clever cases designed for silence, but they lack the elegence of a small enclosed box that never needs to have filters cleaned or the worry that a fan will seize at an inopportune time.
With the release of the Intel Atom and the Via Isaiah I suspect that it will be only a matter of time before we get the desktop system with essentially no downside. Which is why I'm waiting for it, because at that point the upgrade cycle will likely be over for me. Maybe there will be a killer app coming along, but we are 4 cores into the parallelization path of more CPU horsepower and I haven't seen it yet.
"Linux is just not a good platform for something like this as it currently stands. I for one never want to worry about whether or not my glibc is the right fucking version before I install software"
It's a lot easier to just click on synaptic, search for what you want, click a checkbox and install. I've installed loads of software on Ubuntu without once scrap of dependency hell. The model works, is easier and is superior to the buy/warez method of installing software.
(My comment got mangled before, this is a shorter paraphrasing.)
640k should be enough... meh. At some point, businesses approach diminishing returns with technology for the mass market. e.g. typical passenger jet versus Concorde. Another example, the AK-47 should be good enough for everyone, and was built in, you guessed it, 1947.
My daily needs have been well and truly provided for since the P4. In fact, I don't own a faster computer than a P4. This has been the case for something like 4 years now. I'm at the point where a typical consumer in the classic Moore's Law years would have upgraded. I don't see any need to.
Office suite, music, video, internet, games. That's a superset of the average requirements for a computer. For a large number of cores, what possible benefit would you gain? I'm a long-term computer addict, so I'm an early adopter for most uses, and anticipate others. More cores could be useful for a limited number of applications suited to the average computer user.
Any replacement is about reducing the quibbles, not gaining functionality. I want something with the same capabilities as has been on the market for the last 4 years or so. I want it cheap to purchase, small, passively cooled, very low power (i.e. 20W or less), solid state. Hence it will be silent, no maintenance, cheap to run, unintrusive.
The last two computers I bought were under 300 Mhz and under $400. Neither has a video card, both completely solid state, completely silent, and fit in a coat pocket (not that they are portable of course). Both consume less than 10W. I don't anticipate buying another personal computer without similar characteristics.
Slashdot Mirror
Now we have tyranny by anyone who can afford to buy lobbyists, politicians and advertising campaigns.
Wow, 3 brain operations for epilepsy? I'm surprised that you are writing this and not merely someone's research paper.
"What were they expecting? Cavemen who recited poetry?"
I suppose "Developers! Developers! Developers!" could be considered a primitive form of poetry.
That's why the next revision will come with a warning label:
"For front pocket ONLY"
I don't really care whether Linux is good enough for the masses these days, other than to satisfy intellectual curiosity. It's good enough for me, my family and friends, and I appreciate that. I remember the days when uber hackers (i.e. hacking addicts) were the only people I knew running Linux.
It wasn't a matter of pure IQ per se. I knew several extraordinary Linux hackers who failed calculus multiple times (eventually giving up) but ran Linux as a desktop OS way back in 1997 and earlier. It was IQ and perhaps an early introduction by a parent or peer to understanding computing minutiae, along with the ability to just grok Linux. I was always in awe of them.
The bar has been lowered repeatedly since then. Only 9 years later or so, a version of Linux was dumbed down enough to suit my need for a tool that didn't require too much hacking to get to work. It was not only the distribution, it was also the moderation policy and community of ubuntuforums that enabled me to do any necessary hacking with a minimum of pain - a newbie had asked the question once and searching revealed the answer.
I fully expect that Microsoft will bundle their operating for next to nothing on commodity hardware in order to maintain mindshare. I will be surprised if MS doesn't manage to muscle their way to the point where their monopoly is maintained. There comes a point on the bell curve where no one does their own install, no matter how easy it is. This is what protects MS, because the fraction under that bell curve incapapable/unwilling to install is a clear majority.
Meanwhile, Linux distributions (probably Ubuntu based) will keep getting better to the point that they do everything a user could want. Commoditization of PCs and componentry means that supporting the bleeding edge with drivers will become less important as cost, suitability and other factors become more important than pure speed. If people want to switch, it will get easier. Those who see the advantage and have the skills to install an OS will do so if they perceive it to be in their interest. Once the growth in mindshare from those who can install and will benefit from a Linux install is maxed out (probably subject to network effect), growth will come from one of at least three sources.
1) Reduction in Profit Margin * Volume for Windows OEM sales means that MS is less able to dictate terms to hardware manufacturers, and Linux factory installs start to proliferate.
2) Linux becomes cool, and power users start installing it on the computers belonging to the rest of the bell curve.
3) A third party decides it is in their financial interest to back hardware manufacturers into pre-installing Linux, stymying the best efforts of MS.
Anyway, I look forward to the release of Hardy Heron. I expect it to be better than the previous releases and to "just work" for a long time. If you are going to install Linux on a friend's computer, I expect that this will be very stable and so reduce need for your support being a Long Term Support (LTS) release.
"Oh, and I may be a power user, but I'm also a gamer, and I want games that run natively on Linux. Besides a tiny subset of games, that's not happening until Linux is the average desktop."
Oh great, when that happens I'll truly never get any work done. Hopefully that day is soon followed by an Ubuntu variant that disables any access to games in the repos.
"People have said as a joke that OpenOffice.org or similar programs will take over once they have their own clippy, but may a true word is said in jest."
They do have one, it's called the "Help Agent".
"I'm sure some Linux experts will say I'm just stupid, but I tried to install Hardy Heron on a fairly stock, high-end Dell desktop earlier today."
Does the word 'beta' mean anything to you? Release date for Hardy Heron is April 24th.
If you go on the ubuntu site to download it, it tells you:
"Note: This is still a beta release. Do not install it on production machines. The final stable version will be released in April 2008."
Thanks for the link, that was interesting to read.
Another thing to realize is that security companies and consultants have an inherent desire to inflate the size and probability of any security risk. The more fear they can build in the client, the bigger the more expensive the service they can sell.
"If MS really was serious about making life easy for the customer they would've bundled MS Office and a decent mailing client to name a few."
If they did want to bundle Office, would that be against the law? Just wondering. I can see how it could be a viable strategy for them if they decided to embrace the low margin for the sake of higher volume. If the choice is between death and bundling a major cash cow, I think they'd sacrifice the cow.
"Neither can Dell, HP or any other hardware manufacturer. This trend impacts them every bit as much as Microsoft, although on the whole I think hardware manufacturers should be able to adapt easier than Microsoft. "
I think both MS and the hardware manufacturers have known about this for a long time. The eternal upgrade cycle was driven by obsolescence. The moment that faster CPUs would not obsolete older, slower CPUs because of the lack of killer apps requiring faster CPU speed, the profits would drop off. Changes would have to be made to the business model. It could go several ways, and no doubt there have been contingency plans drawn up on this very subject.
If you've ever seen "Letters from Iwo Jima", you will see that MS, Intel and the like are in much the same position as General Kuribayashi, facing the inevitable defeat at the hands of commoditization. If they are smart, their goal will be to bleed the consumer for as much as they can on the way down.
Intel will be trying to beat Via and AMD with just enough performance increase with a reduced power requirement, but not to produce something so good that they can't make something a bit better in another year or two. Eventually I would expect them to look for a way to start increasing obsolescence in other ways. e.g. CPUs will be designed to fail after a minimum number of years. It's tough though, as the technology plateaus they won't be the only one manufacturing CPUs and by doing this they will get a reputation for reduced reliability which will feed sales of the competition.
Another way is of course to include a Microsoft operating system that will be overcome with malware given enough time. This is probably more likely.
What has Microsoft done? They have waited until someone forced their hand. Microsoft has retooled XP and is ready to sacrifice their margins for increased volume. Once the market really takes off, I would expect them to drop the margins of Office as well in order to properly compete with something like Ubuntu that includes OpenOffice. If they don't do this, Ubuntu will be perceived as more useful and gain adherents. The last thing MS wants is a large consumer base happy to buy from a hardware manufacturer who is unwilling to sell MS and Linux systems for the same price (likely by paying MS a drawback for Linux installs).
If I were a greasy monopolist in the shoes of MS, I would see Asus and cut a deal with them. You either install XP on some of your systems and pay us a small fee for ANY sale of a computer (including that of Linux, perhaps even more in the case of Linux), or we will partner with your nearest competitor and subsidize them until you are making no money on your Eee PC. Do we have a deal?
This bluff might be called. I can see that it would be in the long term interest of a country like China to subsidize Linux in a price war until they gained enough mindshare, and then their hardware manufacturers would be free of the Microsoft Tax.
Unfortunately, the biggest problem Ubuntu faces is being in opposition to the long term interest of manufacturers, software vendors, and MS. It might be possible that given enough marketshare, spyware on Ubuntu will become rampant (more money in exploiting vulnerabilities than finding fixes) and people will buy a new computer rather than reinstall. This would be good for the hardware manufacturers and ironically increase support of Ubuntu.
Traditional software vendors (those that sell the install rather than the support) must see the repository as a threat, since the repository is just so much easier and also safer. Most of the large vendors devoid of FOSS religion (e.g. Adobe) will see the threat of their mindshare being eroded more easily via the repo and instinctively avoid encouraging Linux by making their software available. (The exception is the gaming company that sells content that is much more expensive to create than the Open Source community can compete with. e.g. WoW.)
Certainly interesting times.
"Had he not laid all his cards out on the table, he might have got the fight he was looking for, and showed Monster they can't bully everyone around. Part of me is afraid that won't happen now."
Just because he wrote that he was itching for a fight doesn't mean that he was actually itching for one. It's a good bluff though. If he enjoyed litigation so much, chances are he'd still be doing it. Still, a cool letter.
The reason this hasn't been tried commercially is that there aren't commercial quantities of energy available. A bit of googling yields a figure of 250kWh per strike, or less than my personal energy bill for a quarter. It's a bit like shuffling your shoes on the floor to generate static electricity - most of the energy goes towards heating up your shoes, not transfering electrons.
I never had the patience for running a linux desktop despite trying multiple times since 1997 or so. You must have been one of those rare people with enough patience. My opinion changed about a year ago, when I installed Ubuntu for the second time. If I had left Linux in 2003, I would have had the same opinion as you.
Any maintenance issues I've had have lasted less than half an hour a month. It is certainly less of a pain to use than Windows was, a bit of work up front for (so far) a year of working extremely smoothly. The only issues have been reinstalling video drivers each time the kernel upgrades. The fix takes less than 10 minutes. That is the ONLY routine maintenance I have to do (other than click and install updates when they come, which would be the same on a Mac).
Since I don't edit video, I can't comment on that aspect of Ubuntu.
"I don't understand why modern cars get such lousy mileage."
Looking at first order causes... cars get poor mileage because energy is lost in:
1) Having to brake.
2) Having to move air around, which eventually ends up as heat.
3) Deforming tyres.
1) is as much a problem of infrastructure design (frequent stops that can't be anticipated) as it is a problem of lack of driver training (for stops that can be anticipated) and obese vehicles.
2) is a problem of convenience and marketability. (Try implementing stylistic obsolescence when every year the shape must stay the same - a modified teardrop.) It is also a problem of overly high speed limits (legislation).
3) is not really much of a problem in comparison with 1 & 2. Can be solved by pumping up tyres and dealing with the bumpier ride.
Of course, all 3 come at the problem from an engineering perspective. If you look at the problem from an oil company perspective and ask the question "How can I ensure that my mode of transport is more convenient (by far) than all alternate forms of transport, that I get another entity (taxation) to fund the infrastructure, and that I partner/own the automobile companies such that they also extract the maximum dollars/year from consumers", you are asking the right question.
It's then a matter of engineering the situation such that no other forms of transport can compete with gasoline/diesel's high energy density and also relatively low cost/volume. And sitting back to enjoy the profits while they last.
I can't help but think that this is a stop-gap measure. I used to read up about all the various methods of silencing a computer (with the intention to implement myself) but for consumer-grade applications I'd prefer to wait for a variant of Moore's Law to do its work - the propensity for performance per watt to keep increasing until it nears whatever limits are predicted by information theory.
At that stage there will be an option to cool with no moving parts for typical desktop/laptop applications, and it will be a superior solution in all aspects compared with any combination of cooling and sound minimization.
Another welcome change will be for the idle power consumption to drop, which it certainly can.
Efficiency is inherently hard - it's all about approaching zero loss, and losses seem to almost never present a target suitable for a single silver bullet approach. That means work - killing all the losses on your loss budget one by one, and figuring how to integrate all the changes so it still works. The good thing is that once you come up with a design that lops off as many heads of the efficiency hydra as possible, it's just a matter of mass production.
So by your own definition you are a fake nerd?
"Windows is so big, with such a huge established base, that its decline will resemble that of the old IBM mainframe environment -- which is still doing fine, decades after the death of the mainframe was predicted. This ain't going to happen overnight."
It certainly won't happen overnight. There will be a period when MS is a "dead man walking". But it is all about the revenue stream. The revenue stream comes from Office and Windows. If no one (relatively speaking) is upgrading or buying new computers, Microsoft's army of programmers, marketers, lobbyists and salespeople start looking for work elsewhere.
Network effect is a double-edged sword. Great on the way up, great at consolidating a monopoly. When the rats start leaving the sinking ship, they leave in droves. Consider Lotus 1-2-3, or Amiga.
Mainframes still had a very valid use in defense and other big government type problems. Until they get a working strong AI, there will probably still be use for more high-end grunt by governments all over the world. The average consumer is a different market. That market depends on disposable income, as you made excellent reference to, and also need. Until now, CPU power has been "good, but never enough". That is changing.
We are very near the point where the choice will be between something more expensive, large, louder, requiring maintenance and with performance only a minority will actually use, or something cheap (both initial purchase price and ongoing), small, silent, no maintenance that does the job well. The latter ticks the relevant boxes. And there is not much of a profit margin in that world for MS.
It is an open question as to how many people would buy something like that at a particular price, because the upgrade will probably come as a result of a malware infestation for someone computer illiterate, or a hardware failure for someone who knows what they are doing, or the price is so low that the payback period for the electricity cost difference is good. Maybe it will be possible for MS to slash prices, push Intel into the high volume/low margin game as well, and extract the last cash from their cow. It seems like that is what they are reluctantly doing with XP on the Eee PC. It is as if GM broke down and started manufacturing 150cc motorbikes and bicycles.
I suppose it's fully possible that this strategy well executed will result in there never being a year of the linux desktop. I don't really care so long as Linux and especially Ubuntu maintains or improves over time.
I'm not sure I like the bike analogy. The ultimate security in a bike rack is still not very secure compared to a home computer installation. A professional thief can still wait until no one is there to come along and nab your bike using inexpensive tools. If he does that, it's relatively low risk.
It takes a bit more effort, marginally more cost, and you can end up with a home network that is much more secure (at the router/modem interface) than leaving your bike at any bike rack with any sort of security. There are numerous FOSS distros specifically for this application, IPCop, Monowall, PfSense, OpenBSD, Zeroshell to name a few. Of course, if you are going to go to that effort you may as well run linux/bsd and firefox with noscript, no sense turning your front door into a safe and leaving the window open.
That is an investment in time and effort that can be done once, and will likely fend off even a concerted attack. The next step up will require either government or large corporation resources. This would be comparable to leaving your bike either inside your house or inside your place of work, and never at a bike rack... but with a little larger investment up front but more convenience after you have done the work.
Also, until software routinely kills people, it will not be subject to bridge-building type regulations, and if it does, it will only be software written for those things. In fact, I believe that defense applications, airplane software, things where lives or important things are at stake do submit software to the same degree of rigorous testing by qualified people that the design and construction of a bridge would have to withstand.
That's a very good point.
Games also have the downside of being highly addictive. Some things in life just aren't easy, and require significant investment of frustration in the beginning but pay off over time. In order to be highly addictive (and hence popular), games are structured to eliminate a lot of the grind that daily life involves. The end result is that even though valuable learning may have taken place, the brain gets rewired to run back to the refuge of the addiction rather than deal with the grind and accomplish something.
Getting most of your knowledge through games (or movies, or secondary source books) also puts you at risk of learning the media creator's version of the world rather than the world as it is, or the world as primary sources had it. For example, you play Civilization and learn that Neville Chamberlain was one of the worst world leaders of all time (because that equates to a low civilization score). And being a kid, it's easier to just add that to your list of facts about the world along with most of the other stuff that seems mostly right. Another thing you "learn" is that the US is a peaceful isolationist civilization.
Very often some things are "good enough" for a long period of time. Some examples:
-AK-47, built in 1947
-Subsonic passenger jets
-The horse, fastest way to get around for thousands of years.
-C, SQL
-The car, versus the "flying car".
Why development of something plateaus has everything to do with limits to optimization, efficiency, network effect, cost benefit analysis, diminishing marginal returns, return on investment, political and legislative situations. Complacency and familiarity are important, but there are certainly many, many more factors involved.
Sure I'd like an infinitely fast CPU, a commercially viable fusion reactor and a flying car while I'm at it. Some things are hard, and breakthroughs are difficult to schedule.
"As gigahertz race is over and mobility takes over, size, power and price are becoming more important than performance"
Don't forget the silence and no-maintenance aspect that going completely passively cooled and solid-state affords you. And even in a desktop system all the other issues apart from computing speed become important once you experience the difference.
Such other concerns are the whole raison d'etre of silentpcreview.com. There have been some clever cases designed for silence, but they lack the elegence of a small enclosed box that never needs to have filters cleaned or the worry that a fan will seize at an inopportune time.
With the release of the Intel Atom and the Via Isaiah I suspect that it will be only a matter of time before we get the desktop system with essentially no downside. Which is why I'm waiting for it, because at that point the upgrade cycle will likely be over for me. Maybe there will be a killer app coming along, but we are 4 cores into the parallelization path of more CPU horsepower and I haven't seen it yet.
http://en.wikipedia.org/wiki/VIA_Isaiah
"Linux is just not a good platform for something like this as it currently stands. I for one never want to worry about whether or not my glibc is the right fucking version before I install software"
It's a lot easier to just click on synaptic, search for what you want, click a checkbox and install. I've installed loads of software on Ubuntu without once scrap of dependency hell. The model works, is easier and is superior to the buy/warez method of installing software.
(My comment got mangled before, this is a shorter paraphrasing.)
640k should be enough... meh. At some point, businesses approach diminishing returns with technology for the mass market. e.g. typical passenger jet versus Concorde. Another example, the AK-47 should be good enough for everyone, and was built in, you guessed it, 1947.
My daily needs have been well and truly provided for since the P4. In fact, I don't own a faster computer than a P4. This has been the case for something like 4 years now. I'm at the point where a typical consumer in the classic Moore's Law years would have upgraded. I don't see any need to.
Office suite, music, video, internet, games. That's a superset of the average requirements for a computer. For a large number of cores, what possible benefit would you gain? I'm a long-term computer addict, so I'm an early adopter for most uses, and anticipate others. More cores could be useful for a limited number of applications suited to the average computer user.
Any replacement is about reducing the quibbles, not gaining functionality. I want something with the same capabilities as has been on the market for the last 4 years or so. I want it cheap to purchase, small, passively cooled, very low power (i.e. 20W or less), solid state. Hence it will be silent, no maintenance, cheap to run, unintrusive.
The last two computers I bought were under 300 Mhz and under $400. Neither has a video card, both completely solid state, completely silent, and fit in a coat pocket (not that they are portable of course). Both consume less than 10W. I don't anticipate buying another personal computer without similar characteristics.
This new petroleum deposit is actually formed from the remains of ancient long, thin reptiles known to slither on the ground.