MP3 players are a mature market. There isn't much in the way of new features to add to them so there is little point to developing new models. But it is profitable to continue to produce the old ones. SanDisk isn't spending R&D money on the Sansa, but they will continue making them so long as people keep buying. Apple had a long run with the iPod Classic as a cash cow; they only discontinued it because some key parts (notably the tiny hard disk drive) were no longer available.
The $9 does not include a power adapter or battery. It can be powered from USB so any standard USB charger should work as a power source. They are selling batteries if you want to go that way. It also does not include a case.
WiFi is built into the board. So is Bluetooth. But no wired Ethernet.
Yes, in a way it's a cheaper Raspberry Pi. Overall it's comparable to the original Pi - pluses are a faster CPU and built-in WiFi and Bluetooth, the big minus is no built-in HDMI. It falls short of the power of the Pi 2, which has a quad core CPU and 1GB RAM, but we shouldn't expect a $9 computer to be as powerful as a $35 computer. The $9 price point makes it appealing for embedded applications, and the fact that you're getting networking with that $9 makes it even better.
The shared FPU on the FX series is rarely a problem. Sure, there is only one that is shared between a pair of cores - but it is a 128 bit FPU that can do two 64 bit operations simultaneously. It's not likely to be a bottleneck unless you are working with long double or binary128 data, in which case it can only do one operation at a time.
The most common problem with poor utilization of the core pairs in the FX series is cache contention. That's where the updated process scheduler comes in. To oversimplify a bit, you want to assign multiple threads of the same program to the same core pair, and assign different programs to different pairs whenever possible. If different programs get assigned to the same pair, they should be programs with small working sets. You need a scheduler that understands the unique constraints of the core pairs and does not consider all threads to be equal for scheduling purposes.
Similar constraints apply to scheduling hyperthreaded cores on Intel processors. The important difference is that Windows 7 knows how to use those well, but does not know how to use the FX to best advantage. Microsoft made improvements in Windows 8 but did not backport them to older versions of Windows, whereas Linux systems get the improvements with their kernel updates.
Not so clear cut. The Phenom has a slight edge in single-core performance at the same speed, but the FX can give you more cores (8 are readily available but the Phenom topped out at 6) and is now available with higher clock speeds.
One catch with the FX is that you need updates to the scheduling algorithm in the OS. In the case of Windows that means you have to run Windows 8 (or presumably the Windows 10 preview) because earlier Windows versions never got updated to take full advantage of the FX.
The other catch with any AMD processor is the prevalence of Intel's C compiler. As one might expect, it is designed to produce code that runs really well on Intel's own CPUs; performance on AMD is at best an afterthought, and some have accused it of being deliberately pessimized.
Glad that upgrade worked out for you. Doing a similar upgrade to an older laptop with a PATA drive is more challenging because PATA SSDs have vanished from the market. But you can get PATA to mSATA bridge cards for under $10 so it should be feasible if you find a good deal on an mSATA SSD, or have one left over from an upgrade of a Macbook Air.
I'm guessing that there will eventually be a special upgrade path for small tablets; it will be even more necessary for the models with only 16GB flash. Instead of doing the usual install to the primary file system, they will replace the recovery/boot partition. On small tablets the recovery partition serves both purposes; they boot directly from the compressed system image in the recovery partition rather than installing Windows in the usual way.
The tricky thing will be to implement it so that user data and installed applications can be preserved across the upgrade. On the other hand, small tablets usually have sufficiently small amounts of those things that backing up the data and reinstalling apps after the upgrade would not be horrible, and they could automate the reinstall for any apps that come from Microsoft's store.
Another thing that would improve the small tablet experience would be frequent patch rollup updates for them. A problem with the scheme of booting from the recovery partition is that patches from Windows Update quickly chew up space. When you run Windows Update on a normal Windows install, the patched file replaces the old one, and the net loss of space is just the difference in size. When you run Windows Update on a small tablet, the new file is installed in the primary file system but the old one is still part of the compressed system image, so the net loss of space is the full size of the updated file. A patch rollup every three months or so would alleviate the problem; that is, release a new system image that includes all the patches to date. It should be possible to do these patch rollups in a way that would preserve user data and installed apps, even if a full version upgrade can't be done that way.
The study only looked at the fuel consumption of the two modes of travel. Unsurprisingly, a packed plane did well compared to a car with one person in it. The ground travel can do better in a number of ways: a more fuel-efficient car, putting more people in the car, instead studying multi-passenger vehicles like buses and trains.
The current study also fails to look at another important issue: the high impact of air travel on the ozone layer. Because it takes place at high altitudes (and thus releases pollutants into the upper atmosphere), studies have shown that flying causes a lot of damage to the ozone - much more than the same level of polluting activity at ground level.
The LF spectrum in question (135-137 and 472-479 KHz) is not used for BPL (broadband over power lines). The power line communication (PLC) on those frequencies is lower bandwidth data that is used by the utility itself, to control the grid and for the remotely readable power meters that many utilities have installed. Citation for the proposed band sharing: http://www.arrl.org/news/fcc-p...
Power line networking is the least of our problems on HF. A more serious problem is the huge number of broadband noise generators in every modern home. Computing devices like computers, phones, and tablets are the starting point but far from the entire story. Microcontrollers are being put into EVERYTHING these days - if it has a digital display, control buttons, and/or a remote control, it probably has one or more microcontrollers in it. Most modern electronic gadgets have switching power supplies. We are installing CFL and LED light bulbs, so even our LIGHTS have switching power supplies.
The amount of noise that any one device can create is regulated but the cumulative effect of all of them in a dense urban neighborhood is devastating; the noise floor in my house has been raised at least 20dB by all of it. In other words, you have to send 100 times as loud a signal for me to hear you, and that puts a serious crimp in ham radio HF reception. If there were ever a major emergency where I needed to help out on HF, I would probably trip the main circuit breakers in my house and run the radio on batteries. (And I'd still have to deal with the noise from the neighbors.)
The non technically minded hams can also be valuable in emergency communications. Sure, you need somebody to keep the ham repeater, internet link, or whatever going. But you also need people out in the field with radios, talking to people, taking their stories, and passing information along to friends, loved ones, and authorities. Those people don't need to know how their radio works to do their service. They need to know how to USE their radio, even in times of stress and heavy demands. They have an important skill and mission: taking messages and passing them along accurately under challenging conditions.
I was a ham radio operator at the 2013 Boston Marathon. I was well away from ground zero (my station was in Newton, about 10 miles from the finish line) but we also had hams who were in the thick of the disaster. I had a long day because we held runners at our stations and later moved them to temporary shelters; we had to coordinate the plans for getting people on the buses and moving them to the shelters. I think my single most important job that day was to help keep the runners calm, assure them that things were being handled, and help them contact their friends and family so they could get home. The last was done with my cell phone because that was the appropriate tool for the job, and fortunately I had a Sprint phone which was working at my location - Verizon and AT&T towers were mostly overloaded and useless. But even when I wasn't using my ham equipment, the interpersonal and communication skills that I developed in ham radio were key.
Oh... why do hams do communications at public service events like the Marathon? In part because they are excellent training exercises for what to do when disaster strikes. In a normal year we coordinate medical supplies, ambulance pickups, and transport buses for runners who can't complete the course; we could do all of those things with cell phones. (Trained COMMUNICATORS would still be valuable; the medical and logistics people are busy doing their jobs and don't have time to spend on communications.) In 2013 the training suddenly turned into the real thing and we were ready - we wouldn't have been able to use cell phones, because the infrastructure was hopelessly overloaded and it was difficult or impossible to make a call. (Mine worked 10 miles from the finish, but things were far worse near ground zero.)
9/11 was especially bad because most of the telecommunications infrastructure for Lower Manhattan - landline, cellular, and internet - was in the World Trade Center. So were the primary broadcast sites of most of the TV and FM radio stations in New York City. (Some but not all of the broadcasters had backup sites in New Jersey.)
Putting so many of our communication eggs in one basket may be dangerous, but it is also appealing for economic reasons; building a separate tower for each broadcaster would be a lot more expensive, and data interchange between networks is easier if they all have communication nodes in one place. I know the broadcast situation here in Boston isn't much better; toppling a couple of towers in Needham would take out most of our stations.
The one on the belt of the police officer also has a few other things that make it much more expensive. First, it's more rugged than the $60 ham version; it has to be, because it is carried many hours a day and because people don't handle their working tools as carefully as they handle their personally owned equipment. Second, it has features for secure communication that aren't in the ham radio - they CAN'T be, ham radio has to be open to all listeners by law. Third, many commercial radio systems have a feature called trunking that allows many users to share a small number of radio channels while keeping the transmissions of each secure from the others; hams don't do that. (See point #2 for the reason.) Finally, it has to receive a higher level of certification by the FCC, because it is being used for critical communications and because it is being operated by untrained people who cannot be held responsible for technical failures.
Yes, some hams are using digital voice modes now, which will foil a garden variety scanner. But ham transmissions must use open and published standards and be unencrypted. (And the digital ham radios don't cost $60. Yet.)
The OpenOffice / LibreOffice split can be mostly blamed on Oracle. When Oracle bought Sun, OpenOffice became one of their properties, but they really didn't have a clue how to handle it properly. The development process became largely closed off to outside open source developers - sure they could download the code and change it, but Oracle wasn't doing anything useful with their code submissions. The result was that a group of people forked the code and created LibreOffice. Many users and most Linux distributions quickly switched.
Eventually Oracle concluded that OpenOffice wasn't a strategic fit for them. But they weren't willing to admit they had made a mistake and transfer the code ownership and trademarks to The Document Foundation (the LibreOffice organization), so instead they worked out a deal with the Apache Foundation to take over OpenOffice. I allocate a side order of blame to Apache; they should have refused the deal and insisted that Oracle turn everything over to The Document Foundation.
Word 6.0 for the Mac was a badly done port of Word for Windows 6.0, which wasn't all that well written to begin with. It was a dog on the Mac, plus it lacked features that were in the previous Word for Macintosh, 5.1, which was a native Macintosh application. Microsoft's desire to move to a common code base for both GUI platforms was laudable, but they probably would have done better to port the Mac program to Windows like they did with Excel.
Since then Microsoft has fixed the worst of the coding problems, and increased computing power has taken care of the rest. But they still haven't fixed the fact that Word has user interface problems; the shift to the Ribbon in place of the traditional menus wasn't an improvement.
The original Nexus 7 had a hardware Achilles heel: the charging port was very weak. To make matters worse, the replacement part to fix it cost about $50 until Chinese sources got around to cloning them. Aside from that it was a decent piece of hardware, though like all Android devices with only 1GB RAM it started to get bogged down as both the system and the apps got larger over time. The Facebook app is a prime offender; it runs all the time and uses around 100MB all by itself. Lollipop 5.0 was especially bad; 5.1 is much better and seems to be a bit of an improvement over KitKat.
One good solution was to buy the dock. I got one after they were discontinued, on clearance for about $20.
I don't have much interest in doing business with a computer company that is incapable of keeping its web site up to date. I should have no need to talk to their sales people, ever, for any reason.
Dell consumer systems come with the usual bloatware, though Dell is far from the worst offender. If you order a customizable system from their web site rather than one of the stock configurations you can opt out of a lot of it. Dell business systems can be configured with absolutely nothing but Windows itself and the appropriate drivers for your system if that's what you want, and some models can be ordered with Ubuntu Linux instead of Windows.
Chrome is DESIGNED to use lots of storage if you have it. That makes it faster, especially if you open multiple tabs or windows. But it will scale back its memory usage if you don't have it, by doing things like discarding the memory for tabs that you are not currently looking at. It is then slower to revisit those tabs because they have to be re-rendered, and in extreme cases reloaded from the server.
The message found its way to Slashdot over the internet. That almost certainly involved multiple relays; for example, when I did a traceroute from my laptop to hardware.slashdot.org, it showed that my packets went through 15 nodes, divided almost evenly between comcast.net and savvis.net. Everything other than the end points is probably either running Linux or Cisco IOS.
Eventually the packets arrive at slashdot. Slashdot is a database-driven system, so everything you do here is probably being touched by more than one computer - typically the database doesn't run on the same system as the web server.
So... 10-20 systems is NORMAL for the internet. It's how the thing works.
Some cars are already there. If you drive a hybrid or an electric car, the only thing you touch that has a mechanical linkage to anything is the steering wheel; everything else is just a control input for the computer, including the accelerator and the brake pedal. Cars that also do steering by wire have been proposed, so even that last vestige of direct control may disappear within a few years.
In any modern car, the operation of the engine is heavily moderated by computer control. That's the only way you can make a car that meets modern pollution standards while also having acceptable performance and fuel efficiency. The days of direct unmoderated links between the gas pedal and the amount of fuel that flows into the engine are long gone, as are the days of engine timing that is controlled mechanically.
In many ways, the modern era is far superior. Aside from the efficiency issues, they are more reliable. I remember when it was a question whether your car would start on a cold or wet morning, and you often had to crank the engine repeatedly to get it going. Modern cars pretty much just work, except when they break. But the downside is that the technology is more complex and difficult to repair.
A beige toaster (original Macintosh) was also easier to repair than a Macbook Air. But do you really want to go back?
The average rainfall for the entire state of California is about 10 inches per year, but that includes drier areas like the Central Valley and some actual desert areas like Death Valley. The annual average for San Francisco is 23.64 inches, so rainwater collection there will be more effective than your estimate.
Reusing water that lands on roadways is difficult. It tends to be polluted with various byproducts of cars, like petroleum products and worn-off synthetic rubber from the tires. In places that get snow you also have to deal with road salt. Road runoff is not suitable for direct agricultural use; it would have to be cleaned up first.
Rainwater cisterns are only effective if you have rain. That will work in some places like the Bay Area, but it would do little good in places like the Central Valley that are near-deserts in terms of rainfall.
Slashdot Mirror
MP3 players are a mature market. There isn't much in the way of new features to add to them so there is little point to developing new models. But it is profitable to continue to produce the old ones. SanDisk isn't spending R&D money on the Sansa, but they will continue making them so long as people keep buying. Apple had a long run with the iPod Classic as a cash cow; they only discontinued it because some key parts (notably the tiny hard disk drive) were no longer available.
The $9 does not include a power adapter or battery. It can be powered from USB so any standard USB charger should work as a power source. They are selling batteries if you want to go that way. It also does not include a case.
WiFi is built into the board. So is Bluetooth. But no wired Ethernet.
Yes, in a way it's a cheaper Raspberry Pi. Overall it's comparable to the original Pi - pluses are a faster CPU and built-in WiFi and Bluetooth, the big minus is no built-in HDMI. It falls short of the power of the Pi 2, which has a quad core CPU and 1GB RAM, but we shouldn't expect a $9 computer to be as powerful as a $35 computer. The $9 price point makes it appealing for embedded applications, and the fact that you're getting networking with that $9 makes it even better.
Does Coursera have to charge G.S.T. for Canadian customers? If they do that could account for a lot of that spread.
The shared FPU on the FX series is rarely a problem. Sure, there is only one that is shared between a pair of cores - but it is a 128 bit FPU that can do two 64 bit operations simultaneously. It's not likely to be a bottleneck unless you are working with long double or binary128 data, in which case it can only do one operation at a time.
The most common problem with poor utilization of the core pairs in the FX series is cache contention. That's where the updated process scheduler comes in. To oversimplify a bit, you want to assign multiple threads of the same program to the same core pair, and assign different programs to different pairs whenever possible. If different programs get assigned to the same pair, they should be programs with small working sets. You need a scheduler that understands the unique constraints of the core pairs and does not consider all threads to be equal for scheduling purposes.
Similar constraints apply to scheduling hyperthreaded cores on Intel processors. The important difference is that Windows 7 knows how to use those well, but does not know how to use the FX to best advantage. Microsoft made improvements in Windows 8 but did not backport them to older versions of Windows, whereas Linux systems get the improvements with their kernel updates.
Not so clear cut. The Phenom has a slight edge in single-core performance at the same speed, but the FX can give you more cores (8 are readily available but the Phenom topped out at 6) and is now available with higher clock speeds.
One catch with the FX is that you need updates to the scheduling algorithm in the OS. In the case of Windows that means you have to run Windows 8 (or presumably the Windows 10 preview) because earlier Windows versions never got updated to take full advantage of the FX.
The other catch with any AMD processor is the prevalence of Intel's C compiler. As one might expect, it is designed to produce code that runs really well on Intel's own CPUs; performance on AMD is at best an afterthought, and some have accused it of being deliberately pessimized.
Glad that upgrade worked out for you. Doing a similar upgrade to an older laptop with a PATA drive is more challenging because PATA SSDs have vanished from the market. But you can get PATA to mSATA bridge cards for under $10 so it should be feasible if you find a good deal on an mSATA SSD, or have one left over from an upgrade of a Macbook Air.
I'm guessing that there will eventually be a special upgrade path for small tablets; it will be even more necessary for the models with only 16GB flash. Instead of doing the usual install to the primary file system, they will replace the recovery/boot partition. On small tablets the recovery partition serves both purposes; they boot directly from the compressed system image in the recovery partition rather than installing Windows in the usual way.
The tricky thing will be to implement it so that user data and installed applications can be preserved across the upgrade. On the other hand, small tablets usually have sufficiently small amounts of those things that backing up the data and reinstalling apps after the upgrade would not be horrible, and they could automate the reinstall for any apps that come from Microsoft's store.
Another thing that would improve the small tablet experience would be frequent patch rollup updates for them. A problem with the scheme of booting from the recovery partition is that patches from Windows Update quickly chew up space. When you run Windows Update on a normal Windows install, the patched file replaces the old one, and the net loss of space is just the difference in size. When you run Windows Update on a small tablet, the new file is installed in the primary file system but the old one is still part of the compressed system image, so the net loss of space is the full size of the updated file. A patch rollup every three months or so would alleviate the problem; that is, release a new system image that includes all the patches to date. It should be possible to do these patch rollups in a way that would preserve user data and installed apps, even if a full version upgrade can't be done that way.
The study only looked at the fuel consumption of the two modes of travel. Unsurprisingly, a packed plane did well compared to a car with one person in it. The ground travel can do better in a number of ways: a more fuel-efficient car, putting more people in the car, instead studying multi-passenger vehicles like buses and trains.
The current study also fails to look at another important issue: the high impact of air travel on the ozone layer. Because it takes place at high altitudes (and thus releases pollutants into the upper atmosphere), studies have shown that flying causes a lot of damage to the ozone - much more than the same level of polluting activity at ground level.
The LF spectrum in question (135-137 and 472-479 KHz) is not used for BPL (broadband over power lines). The power line communication (PLC) on those frequencies is lower bandwidth data that is used by the utility itself, to control the grid and for the remotely readable power meters that many utilities have installed. Citation for the proposed band sharing: http://www.arrl.org/news/fcc-p...
Power line networking is the least of our problems on HF. A more serious problem is the huge number of broadband noise generators in every modern home. Computing devices like computers, phones, and tablets are the starting point but far from the entire story. Microcontrollers are being put into EVERYTHING these days - if it has a digital display, control buttons, and/or a remote control, it probably has one or more microcontrollers in it. Most modern electronic gadgets have switching power supplies. We are installing CFL and LED light bulbs, so even our LIGHTS have switching power supplies.
The amount of noise that any one device can create is regulated but the cumulative effect of all of them in a dense urban neighborhood is devastating; the noise floor in my house has been raised at least 20dB by all of it. In other words, you have to send 100 times as loud a signal for me to hear you, and that puts a serious crimp in ham radio HF reception. If there were ever a major emergency where I needed to help out on HF, I would probably trip the main circuit breakers in my house and run the radio on batteries. (And I'd still have to deal with the noise from the neighbors.)
The non technically minded hams can also be valuable in emergency communications. Sure, you need somebody to keep the ham repeater, internet link, or whatever going. But you also need people out in the field with radios, talking to people, taking their stories, and passing information along to friends, loved ones, and authorities. Those people don't need to know how their radio works to do their service. They need to know how to USE their radio, even in times of stress and heavy demands. They have an important skill and mission: taking messages and passing them along accurately under challenging conditions.
I was a ham radio operator at the 2013 Boston Marathon. I was well away from ground zero (my station was in Newton, about 10 miles from the finish line) but we also had hams who were in the thick of the disaster. I had a long day because we held runners at our stations and later moved them to temporary shelters; we had to coordinate the plans for getting people on the buses and moving them to the shelters. I think my single most important job that day was to help keep the runners calm, assure them that things were being handled, and help them contact their friends and family so they could get home. The last was done with my cell phone because that was the appropriate tool for the job, and fortunately I had a Sprint phone which was working at my location - Verizon and AT&T towers were mostly overloaded and useless. But even when I wasn't using my ham equipment, the interpersonal and communication skills that I developed in ham radio were key.
Oh... why do hams do communications at public service events like the Marathon? In part because they are excellent training exercises for what to do when disaster strikes. In a normal year we coordinate medical supplies, ambulance pickups, and transport buses for runners who can't complete the course; we could do all of those things with cell phones. (Trained COMMUNICATORS would still be valuable; the medical and logistics people are busy doing their jobs and don't have time to spend on communications.) In 2013 the training suddenly turned into the real thing and we were ready - we wouldn't have been able to use cell phones, because the infrastructure was hopelessly overloaded and it was difficult or impossible to make a call. (Mine worked 10 miles from the finish, but things were far worse near ground zero.)
9/11 was especially bad because most of the telecommunications infrastructure for Lower Manhattan - landline, cellular, and internet - was in the World Trade Center. So were the primary broadcast sites of most of the TV and FM radio stations in New York City. (Some but not all of the broadcasters had backup sites in New Jersey.)
Putting so many of our communication eggs in one basket may be dangerous, but it is also appealing for economic reasons; building a separate tower for each broadcaster would be a lot more expensive, and data interchange between networks is easier if they all have communication nodes in one place. I know the broadcast situation here in Boston isn't much better; toppling a couple of towers in Needham would take out most of our stations.
The one on the belt of the police officer also has a few other things that make it much more expensive. First, it's more rugged than the $60 ham version; it has to be, because it is carried many hours a day and because people don't handle their working tools as carefully as they handle their personally owned equipment. Second, it has features for secure communication that aren't in the ham radio - they CAN'T be, ham radio has to be open to all listeners by law. Third, many commercial radio systems have a feature called trunking that allows many users to share a small number of radio channels while keeping the transmissions of each secure from the others; hams don't do that. (See point #2 for the reason.) Finally, it has to receive a higher level of certification by the FCC, because it is being used for critical communications and because it is being operated by untrained people who cannot be held responsible for technical failures.
Yes, some hams are using digital voice modes now, which will foil a garden variety scanner. But ham transmissions must use open and published standards and be unencrypted. (And the digital ham radios don't cost $60. Yet.)
The OpenOffice / LibreOffice split can be mostly blamed on Oracle. When Oracle bought Sun, OpenOffice became one of their properties, but they really didn't have a clue how to handle it properly. The development process became largely closed off to outside open source developers - sure they could download the code and change it, but Oracle wasn't doing anything useful with their code submissions. The result was that a group of people forked the code and created LibreOffice. Many users and most Linux distributions quickly switched.
Eventually Oracle concluded that OpenOffice wasn't a strategic fit for them. But they weren't willing to admit they had made a mistake and transfer the code ownership and trademarks to The Document Foundation (the LibreOffice organization), so instead they worked out a deal with the Apache Foundation to take over OpenOffice. I allocate a side order of blame to Apache; they should have refused the deal and insisted that Oracle turn everything over to The Document Foundation.
Word 6.0 for the Mac was a badly done port of Word for Windows 6.0, which wasn't all that well written to begin with. It was a dog on the Mac, plus it lacked features that were in the previous Word for Macintosh, 5.1, which was a native Macintosh application. Microsoft's desire to move to a common code base for both GUI platforms was laudable, but they probably would have done better to port the Mac program to Windows like they did with Excel.
Since then Microsoft has fixed the worst of the coding problems, and increased computing power has taken care of the rest. But they still haven't fixed the fact that Word has user interface problems; the shift to the Ribbon in place of the traditional menus wasn't an improvement.
And let's not forget Windows tablets. The mobile apps might also make sense there, especially for tablets with cellular data capability.
The original Nexus 7 had a hardware Achilles heel: the charging port was very weak. To make matters worse, the replacement part to fix it cost about $50 until Chinese sources got around to cloning them. Aside from that it was a decent piece of hardware, though like all Android devices with only 1GB RAM it started to get bogged down as both the system and the apps got larger over time. The Facebook app is a prime offender; it runs all the time and uses around 100MB all by itself. Lollipop 5.0 was especially bad; 5.1 is much better and seems to be a bit of an improvement over KitKat.
One good solution was to buy the dock. I got one after they were discontinued, on clearance for about $20.
I don't have much interest in doing business with a computer company that is incapable of keeping its web site up to date. I should have no need to talk to their sales people, ever, for any reason.
Dell consumer systems come with the usual bloatware, though Dell is far from the worst offender. If you order a customizable system from their web site rather than one of the stock configurations you can opt out of a lot of it. Dell business systems can be configured with absolutely nothing but Windows itself and the appropriate drivers for your system if that's what you want, and some models can be ordered with Ubuntu Linux instead of Windows.
Chrome is DESIGNED to use lots of storage if you have it. That makes it faster, especially if you open multiple tabs or windows. But it will scale back its memory usage if you don't have it, by doing things like discarding the memory for tabs that you are not currently looking at. It is then slower to revisit those tabs because they have to be re-rendered, and in extreme cases reloaded from the server.
The message found its way to Slashdot over the internet. That almost certainly involved multiple relays; for example, when I did a traceroute from my laptop to hardware.slashdot.org, it showed that my packets went through 15 nodes, divided almost evenly between comcast.net and savvis.net. Everything other than the end points is probably either running Linux or Cisco IOS.
Eventually the packets arrive at slashdot. Slashdot is a database-driven system, so everything you do here is probably being touched by more than one computer - typically the database doesn't run on the same system as the web server.
So... 10-20 systems is NORMAL for the internet. It's how the thing works.
Some cars are already there. If you drive a hybrid or an electric car, the only thing you touch that has a mechanical linkage to anything is the steering wheel; everything else is just a control input for the computer, including the accelerator and the brake pedal. Cars that also do steering by wire have been proposed, so even that last vestige of direct control may disappear within a few years.
In any modern car, the operation of the engine is heavily moderated by computer control. That's the only way you can make a car that meets modern pollution standards while also having acceptable performance and fuel efficiency. The days of direct unmoderated links between the gas pedal and the amount of fuel that flows into the engine are long gone, as are the days of engine timing that is controlled mechanically.
In many ways, the modern era is far superior. Aside from the efficiency issues, they are more reliable. I remember when it was a question whether your car would start on a cold or wet morning, and you often had to crank the engine repeatedly to get it going. Modern cars pretty much just work, except when they break. But the downside is that the technology is more complex and difficult to repair.
A beige toaster (original Macintosh) was also easier to repair than a Macbook Air. But do you really want to go back?
The average rainfall for the entire state of California is about 10 inches per year, but that includes drier areas like the Central Valley and some actual desert areas like Death Valley. The annual average for San Francisco is 23.64 inches, so rainwater collection there will be more effective than your estimate.
Reusing water that lands on roadways is difficult. It tends to be polluted with various byproducts of cars, like petroleum products and worn-off synthetic rubber from the tires. In places that get snow you also have to deal with road salt. Road runoff is not suitable for direct agricultural use; it would have to be cleaned up first.
Rainwater cisterns are only effective if you have rain. That will work in some places like the Bay Area, but it would do little good in places like the Central Valley that are near-deserts in terms of rainfall.