The goal is to breed animals with digestive systems that can create less methane
Is this really necessary? My local natural gas company buys cow produced methane and then sells it as natural gas. It doesn't really matter if the methane I burn to heat my house comes from an oil well or from cow dung, and apparently according to the gas companies brochure, the cow dung is cheap enough to be price competitive with mined methane.
Why don't we just use the cows we already have as both a source of food as well as a source of carbon neutral natural gas? Unlike carbon buried deep underground millions of years ago, cow dung is made of carbon on the surface of the planet that is already active in the carbon cycle, and hence won't result in long term increase of the ppm concentration of CO2.
I wonder whether it's really a lack of ideas, and or worse engineering staff.
Sure there are a bunch of crappy engineers out there. Most of us (myself included) have to deal with them every day. However, I don't think that's the reason why it takes more staff to get the same results these days. I think the real reason is because all of the easier ways of making advancements that use less staff have already been done.
This is especially true in semiconductor process development. Remember when high-k gate insulators came out and gave us a huge performance boost? We didn't get a boost like that again until tri-gate transistors came out. Developing tri-gate tech is way more difficult, but we got the same bump out of it. Now compare that to the difficulty and cost of developing extreme ultraviolet lithography, where you literally need to figure out how to ionize tungsten plasma to a high enough eV to emit light at the correct wavelength with enough power to efficiently process a high volume of wafers. Your probably putting in 1000000W of electricity to get 1000W of light.
We have hit the point of diminishing returns on CMOS semiconductor R&D. This has nothing to do with laziness.
Considering that Chrome uses more or less the same API, and Chrome has an equivalent called Tampermonkey, I'm pretty sure Geasemonkey will be fine. The more important thing is that Mozilla has been very open to adding stuff to the WebExtension API if there is something an existing addon needs that isn't there yet.
WebExtensions is a good long term direction, the problem is Mozilla is forcing it WAY too fast. It's pretty clear the community is not ready for the switch yet, but as usual Mozilla doesn't give a crap about thier user base.
I gave it a try with Ubuntu 14.04 a couple years ago (before 16.04 came out,) basic 2D graphics do work fine, but even with XFCE I get a lot of screen tearing, much more than I got with Ubuntu 6.06 back in the day. The display drivers appear to be much slower than they used to be. It seems like 3D acceleration support no longer works either.
That said I didn't put a ton of effort in to editing xorg.conf files... I just don't have the patience for it like I used to... esp. since It Just Works(tm) on most laptops built in the last 5 years given 90% of them all have the same standard issue mediocre Intel HD Graphics.
To this day my ancient T-40 has my favorite laptop keyboard. I wrote soooo much code on that thing. Alas it is nowhere near fast enough for today's software, and the drivers haven't been updated since the WinXP days. Even modern Linux distros don't really work with its antique ATI gfx anymore. A new laptop with that keyboard would be awesome. I'd buy one.
I saw some ads for this actually. How the hell is this one hour of TV worth 10 months of Netflix? The fight is fake anyway, its just a soap opera for men... I just don't get it.
I agree with you on the fundamental premise of "put Linux on it" to get better privacy and usually longer support lifetimes... but this is actually bad advice in this particular narrow scenario. These Clover Trail SOCs don't have Linux drivers! Moreover, all Clover Trail systems shipped with 32-bit UEFI with no legacy boot support (aka no CSM). None of the major distros have put any effort in to supporting this platform. These computers are pretty much Windows only, the only sane option seems to be to run Windows 8.1 on them.
The thing that is special about Apple is that it has become a luxury brand. Being able to afford Apple products is a sign of wealth and social status, just the same as buying a Ford vs. buying a BMW. Sure, there are technical things about Mac that are nice, the UNIX based OS for example, but that only matters to geeks on/. Build quality was a major reason to buy Apple 5 years ago when PCs were a sea of cheap plastic, but with machines like the Surface, Spectre, XPS 13, and Yoga the other OEMs have more or less cleaned up their act.
For the average consumer, they start their computer, they click the Chrome icon and then they type in facebook.com and look at the daily gossip. For that type of usage, the only difference between Windows and macOS is that the location of the Chrome icon is slightly different. The reason they buy the Apple system is because the machine looks fancy and expensive and it has that Apple logo on the lid so they fit in with everyone else when they bring it to Starbucks.
Now I agree with him that a lot of people in OSS do not act like professionals and they make petty arguments that a most always boil down to an emotional attachment to the code that one has written. That really annoys me. However, the lack of support for Mir to me comes down to the fact that it is redundant with Wayland and isn't as big of a technical step forward as Wayland. There is a real cost to needing to implement 3 different display drivers for each GPU instead of 1 or 2 (X11, Wayland, Mir.) So it is entirely understandable to me that there would be some pushback. Once Intel decided to only provide drivers for X11 and Wayland that really should have been the wakeup call to just switch to Wayland and be done with it instead of trying to reinvent absolutely everything.
I suspect that is going on here is that with the advent of Netflix and tablets we now have most people using tablets as nothing more than a small portable TV that will also show them their Facebook feed. For the average suburbanite slob it has it all, your trashy, fake TV and your trashy, fake friends!
When writing C code where exceptions are not available, I'll often use goto statements to perform error handling. Its useful because you only have to write any necessary memory de-allocation code once. Its pretty easy to simulate a try/finally block using them and its a way better and more readable way to write that sort of code then the other alternatives you have in C. All that said, I prefer exceptions, they more or less get the same job done, its easier to nest them, and IMHO are a bit more readable than the goto method. Only downside of exceptions is the execution overhead is often very bad compared to a goto.
Sounds like MSFT has realized that only putting the DRM/lockdown features on the ARM Windows RT tablets only had the effect of making those tablets not sell. By making it universal, now it doesn't matter if you have X86 or ARM, everyone gets locked down the same way. My guess is the primary reason for this move is so Microsoft can charge manufactures $10 per Windows license instead of $60, and then if you want to unlock your device and actually make it useful you have to go login to the Windows store and shell out the $50 to make all the software you actually care about work. Now that they have the X86->ARM emulator, they can treat all devices the same way from a licensing standpoint and turn this is to a pure profit generator.
I also suspect that we will now see the subscription charges for Windows 10 materialize. Everyone who upgraded from Windows 7 has either the Home or Pro edition right now. But once Jan. 14, 2020 comes (end of Win7 support) anyone who upgraded to Win10 gets downgraded to Cloud edition. On Jan. 10, 2023 everyone who upgraded from Windows 8 (end of Win8 support) will have the same thing happen. At that point, you can expect to pay $50 to the Windows store get another 10 years on the Home edition or $100 for the Pro edition. Or you can just go buy a new computer which comes with 10 years of either Home or Pro from the time of initial activation.
Also, you can expect on Jan. 10, 2023 (end of Win8 support) that MSFT will completely drop support for booting on non-UEFI computers.
GDP isn't the right metric. If we used GDP the price would still be too high in developing countries. The correct metric is GDP per capita, or purchasing power parity.
Problem is you would need to region lock the software then through yet more DRM.
The thing is, the Qualcomm product is demonstrably better than the competing options...We can see that right now with the intel chipset iPhones falling flat on their faces compared to the superior Qualcomm iPhones.
The Intel modems are inferior, but saying they fall flat on their face is a vast over exaggeration. I have a Zenfone 2 and my wife has a HTC 10. Her Snapdragon 820 is 2 years newer than the XMM 7262 in my phone. The 820 is LTE Cat 12, the 7262 Cat 6, we are both on T-Mobile LTE. I just did a speed test on each phone same place and time, mine got 52 Mb/s down, 27 Mb/s up. Hers gets 89 Mb/s down, 15 Mb/s up. Yeah the 2 years newer Qualcomm chip is better, but if you compare it with the XMM 7480 (same year as the 820) its Cat 9 downlink, Cat 13 uplink. I don't have a XMM 7480, but chances are its pretty close to the 820 and might even be better upload speed. Regardless, a 52 Mb/s connection to my cell phone is pretty damn good, I'm not going to use that bandwidth reading email on my phone.
More importantly, it is imperative that Intel, Mediatek, etc. continue to invest in modems and challenge Qualcomm. Most of us remember the consequences of a tech monopoly and don't want it to happen again. I know its almost a sin on/. but I'm rooting for Intel to keep getting big design wins like the iPhone. There used to be a lot more players in the modem market... most of them have dropped out now. We need the retain the few that are left.
I think it goes further in that Microsoft is most likely intentionally sabotaging Windows 7. It seems that almost every Windows 7 computer I encounter has svchost.exe fully consuming a CPU core and consuming massive amounts of memory for no reason other than a failed update.
This issue is because the dependency resolution algorithm in Windows Update is NP-hard. Its not a big deal until the number of updates gets large, and the dependency graph gets reset every time MSFT releases a service pack. Recently those resets have been done by the Win8-->Win8.1 upgrade and on Win10 every ~6 months they release a new OS image (at time of writing, Win10 TH1-->Win10 TH2-->Win10 RS1, pretty soon we will have RS2, and so on.) So the new Win10 model effectively masks the problem since they will have very frequent resets of the dependency graph now. Also, the cumulative updates further help reduce the growth of that graph. That doesn't help Win7 of course. What they really should do is release a Win7 SP2, reset the dependency graph, and make everyone's life easier, but with how aggressively they are pushing the Win10 upgrade you can bet they won't do anything to make life on Win7 easier.
The reason why is because The Walt Disney Company owns both ABC and ESPN. So they force you to buy all the channels they own, or none. Nothing in between. Disney isn't the only one that does that of course, every company that owns multiple networks does. The only way that will change is if the US Government forces them to offer a la carte. The government would also have to force them to not set pricing such that the a la carte cost for 1 channel is the same as the cost of the bundle (maybe a legally mandated 10% max bundle discount or something.)
As the CEO of T-Mobile Germany stated, the "uncarrier" model that T-Mobile USA has been selling for the last couple years is unsustainable. To be fair, they were selling service so low it really was cheaper than it should have been. It was only a matter of time before a rate hike of some sort would happen. They are trying really really hard to spin the rate hike with marketing to make it sounds like a bigger deal than before with the whole "unlimited" wording, but truth is its no different than their previous plans, except for the extra cost. Don't let them convince you that the rate hike would not have happened had they been allowed to merge with Sprint by the way. I'm just glad I got in while they were still selling the ridiculously low price plans.
The article you link is misleading... The Intel chips being compared with the A9X in that article are Broadwell... 2 generations behind compared to the A9X which is the current chip. The A10X will have to be compete with Cannonlake, not the Skylake chips you see in the current Mac lineup. Assuming the A10X scales from the A10 about the same as the A9X scales from the A9, and Intel IPC generational improvements are about the same as past ones... single core ~6W TDP 2.2 GHz A10X is going to be close to raw single core performance of the ~4.5W TDP ~1.2 GHz Cannonlake, but the even the low power Cannonlake CPU will retain a commanding lead on multi core performance.
The one place where Apple ARM does actually dominate Intel is integrated graphics performance. It's incredible to me how Intel graphics are still so far behind everyone else. At least the recent Intel GPUs render mostly correct now and the graphics drivers are pretty stable at this point. The one advantage that Apple (and other ARM vendors) have on the GPU front that they don't need to implement DirectX + OpenGL in their silicon, a luxury that nVidia/AMD/Intel don't get.
Intel tried to do that a few years ago through the OEM channel when they were pushing aggressively in to the tablet market. Being able to run Windows and Android is a natural advantage for x86. Both Google AND Microsoft killed it by refusal to certify the resulting device. Even if Microsoft is cool with it now it is highly unlikely Google's stance has changed.
but once it works on full steam it munches close to ~90% of its TDP.
Are you seriously saying that if a workload start using 100% CPU would would rather Intel make its CPUs throttle instead of actually running at the frequency that you paid for?
I'm sorry but Intel is the LAST thing you can blame for the MBP battery life sucking. If it was Intel's fault then we would see it in all the Skylake Windows laptops, which we definitely do not (most Skylake laptops have 10+ hour battery life.)
Blame for this issue falls squarely on Apple. My best guess is its a combination of small battery capacity along with issues with their power management software. One area that would be particularly suspect is some problem with the switchable graphics control software being too aggressive with powering up the dGPU when its not actually needed. Or maybe there is a bug where they power it up but then in some cases forget to shut it down after it is no longer being used.
Taking the Lumia off the market is a very intentional maneuver on Microsoft's part. They are hoping that by the time they launch the Surface Phone in Q4 next year with full support for X86 Win32 apps on ARM that we will have largely forgetten about how much the previous Windows Phone underwhelmed us. Also making a short but clean break gives them an excuse not to upgrade all the phones on the market right now to the new OS that supports X86 Win32 apps.
Slashdot Mirror
Is this really necessary? My local natural gas company buys cow produced methane and then sells it as natural gas. It doesn't really matter if the methane I burn to heat my house comes from an oil well or from cow dung, and apparently according to the gas companies brochure, the cow dung is cheap enough to be price competitive with mined methane.
Why don't we just use the cows we already have as both a source of food as well as a source of carbon neutral natural gas? Unlike carbon buried deep underground millions of years ago, cow dung is made of carbon on the surface of the planet that is already active in the carbon cycle, and hence won't result in long term increase of the ppm concentration of CO2.
I wonder whether it's really a lack of ideas, and or worse engineering staff.
Sure there are a bunch of crappy engineers out there. Most of us (myself included) have to deal with them every day. However, I don't think that's the reason why it takes more staff to get the same results these days. I think the real reason is because all of the easier ways of making advancements that use less staff have already been done.
This is especially true in semiconductor process development. Remember when high-k gate insulators came out and gave us a huge performance boost? We didn't get a boost like that again until tri-gate transistors came out. Developing tri-gate tech is way more difficult, but we got the same bump out of it. Now compare that to the difficulty and cost of developing extreme ultraviolet lithography, where you literally need to figure out how to ionize tungsten plasma to a high enough eV to emit light at the correct wavelength with enough power to efficiently process a high volume of wafers. Your probably putting in 1000000W of electricity to get 1000W of light.
We have hit the point of diminishing returns on CMOS semiconductor R&D. This has nothing to do with laziness.
Considering that Chrome uses more or less the same API, and Chrome has an equivalent called Tampermonkey, I'm pretty sure Geasemonkey will be fine. The more important thing is that Mozilla has been very open to adding stuff to the WebExtension API if there is something an existing addon needs that isn't there yet.
WebExtensions is a good long term direction, the problem is Mozilla is forcing it WAY too fast. It's pretty clear the community is not ready for the switch yet, but as usual Mozilla doesn't give a crap about thier user base.
I gave it a try with Ubuntu 14.04 a couple years ago (before 16.04 came out,) basic 2D graphics do work fine, but even with XFCE I get a lot of screen tearing, much more than I got with Ubuntu 6.06 back in the day. The display drivers appear to be much slower than they used to be. It seems like 3D acceleration support no longer works either.
That said I didn't put a ton of effort in to editing xorg.conf files... I just don't have the patience for it like I used to... esp. since It Just Works(tm) on most laptops built in the last 5 years given 90% of them all have the same standard issue mediocre Intel HD Graphics.
To this day my ancient T-40 has my favorite laptop keyboard. I wrote soooo much code on that thing. Alas it is nowhere near fast enough for today's software, and the drivers haven't been updated since the WinXP days. Even modern Linux distros don't really work with its antique ATI gfx anymore. A new laptop with that keyboard would be awesome. I'd buy one.
I saw some ads for this actually. How the hell is this one hour of TV worth 10 months of Netflix? The fight is fake anyway, its just a soap opera for men... I just don't get it.
I saw a couple ads for this actually. How the hell is this one hour of TV worth 10 months of Netflix? The fight is fake anyway... I just don't get it.
I agree with you on the fundamental premise of "put Linux on it" to get better privacy and usually longer support lifetimes... but this is actually bad advice in this particular narrow scenario. These Clover Trail SOCs don't have Linux drivers! Moreover, all Clover Trail systems shipped with 32-bit UEFI with no legacy boot support (aka no CSM). None of the major distros have put any effort in to supporting this platform. These computers are pretty much Windows only, the only sane option seems to be to run Windows 8.1 on them.
So, would YOU have thought it reasonable to use an unsigned 32-bit integer for the number of chess games? ... Two BILLION of them?
Umm... an unsigned 32 bit integer would allow 4 BILLION (2^32 to be precise) games, not 2 BILLION. They are using a signed 32 bit integer (2^31 - 1).
The thing that is special about Apple is that it has become a luxury brand. Being able to afford Apple products is a sign of wealth and social status, just the same as buying a Ford vs. buying a BMW. Sure, there are technical things about Mac that are nice, the UNIX based OS for example, but that only matters to geeks on /. Build quality was a major reason to buy Apple 5 years ago when PCs were a sea of cheap plastic, but with machines like the Surface, Spectre, XPS 13, and Yoga the other OEMs have more or less cleaned up their act.
For the average consumer, they start their computer, they click the Chrome icon and then they type in facebook.com and look at the daily gossip. For that type of usage, the only difference between Windows and macOS is that the location of the Chrome icon is slightly different. The reason they buy the Apple system is because the machine looks fancy and expensive and it has that Apple logo on the lid so they fit in with everyone else when they bring it to Starbucks.
Want to keep unapproved changes out and have a cross geo org? Easy, just make Gerrit code reviews mandatory.
Now I agree with him that a lot of people in OSS do not act like professionals and they make petty arguments that a most always boil down to an emotional attachment to the code that one has written. That really annoys me. However, the lack of support for Mir to me comes down to the fact that it is redundant with Wayland and isn't as big of a technical step forward as Wayland. There is a real cost to needing to implement 3 different display drivers for each GPU instead of 1 or 2 (X11, Wayland, Mir.) So it is entirely understandable to me that there would be some pushback. Once Intel decided to only provide drivers for X11 and Wayland that really should have been the wakeup call to just switch to Wayland and be done with it instead of trying to reinvent absolutely everything.
I suspect that is going on here is that with the advent of Netflix and tablets we now have most people using tablets as nothing more than a small portable TV that will also show them their Facebook feed. For the average suburbanite slob it has it all, your trashy, fake TV and your trashy, fake friends!
I don't.
When writing C code where exceptions are not available, I'll often use goto statements to perform error handling. Its useful because you only have to write any necessary memory de-allocation code once. Its pretty easy to simulate a try/finally block using them and its a way better and more readable way to write that sort of code then the other alternatives you have in C. All that said, I prefer exceptions, they more or less get the same job done, its easier to nest them, and IMHO are a bit more readable than the goto method. Only downside of exceptions is the execution overhead is often very bad compared to a goto.
Sounds like MSFT has realized that only putting the DRM/lockdown features on the ARM Windows RT tablets only had the effect of making those tablets not sell. By making it universal, now it doesn't matter if you have X86 or ARM, everyone gets locked down the same way. My guess is the primary reason for this move is so Microsoft can charge manufactures $10 per Windows license instead of $60, and then if you want to unlock your device and actually make it useful you have to go login to the Windows store and shell out the $50 to make all the software you actually care about work. Now that they have the X86->ARM emulator, they can treat all devices the same way from a licensing standpoint and turn this is to a pure profit generator.
I also suspect that we will now see the subscription charges for Windows 10 materialize. Everyone who upgraded from Windows 7 has either the Home or Pro edition right now. But once Jan. 14, 2020 comes (end of Win7 support) anyone who upgraded to Win10 gets downgraded to Cloud edition. On Jan. 10, 2023 everyone who upgraded from Windows 8 (end of Win8 support) will have the same thing happen. At that point, you can expect to pay $50 to the Windows store get another 10 years on the Home edition or $100 for the Pro edition. Or you can just go buy a new computer which comes with 10 years of either Home or Pro from the time of initial activation.
Also, you can expect on Jan. 10, 2023 (end of Win8 support) that MSFT will completely drop support for booting on non-UEFI computers.
GDP isn't the right metric. If we used GDP the price would still be too high in developing countries. The correct metric is GDP per capita, or purchasing power parity.
Problem is you would need to region lock the software then through yet more DRM.
The thing is, the Qualcomm product is demonstrably better than the competing options...We can see that right now with the intel chipset iPhones falling flat on their faces compared to the superior Qualcomm iPhones.
The Intel modems are inferior, but saying they fall flat on their face is a vast over exaggeration. I have a Zenfone 2 and my wife has a HTC 10. Her Snapdragon 820 is 2 years newer than the XMM 7262 in my phone. The 820 is LTE Cat 12, the 7262 Cat 6, we are both on T-Mobile LTE. I just did a speed test on each phone same place and time, mine got 52 Mb/s down, 27 Mb/s up. Hers gets 89 Mb/s down, 15 Mb/s up. Yeah the 2 years newer Qualcomm chip is better, but if you compare it with the XMM 7480 (same year as the 820) its Cat 9 downlink, Cat 13 uplink. I don't have a XMM 7480, but chances are its pretty close to the 820 and might even be better upload speed. Regardless, a 52 Mb/s connection to my cell phone is pretty damn good, I'm not going to use that bandwidth reading email on my phone.
More importantly, it is imperative that Intel, Mediatek, etc. continue to invest in modems and challenge Qualcomm. Most of us remember the consequences of a tech monopoly and don't want it to happen again. I know its almost a sin on /. but I'm rooting for Intel to keep getting big design wins like the iPhone. There used to be a lot more players in the modem market... most of them have dropped out now. We need the retain the few that are left.
I think it goes further in that Microsoft is most likely intentionally sabotaging Windows 7. It seems that almost every Windows 7 computer I encounter has svchost.exe fully consuming a CPU core and consuming massive amounts of memory for no reason other than a failed update.
This issue is because the dependency resolution algorithm in Windows Update is NP-hard. Its not a big deal until the number of updates gets large, and the dependency graph gets reset every time MSFT releases a service pack. Recently those resets have been done by the Win8-->Win8.1 upgrade and on Win10 every ~6 months they release a new OS image (at time of writing, Win10 TH1-->Win10 TH2-->Win10 RS1, pretty soon we will have RS2, and so on.) So the new Win10 model effectively masks the problem since they will have very frequent resets of the dependency graph now. Also, the cumulative updates further help reduce the growth of that graph. That doesn't help Win7 of course. What they really should do is release a Win7 SP2, reset the dependency graph, and make everyone's life easier, but with how aggressively they are pushing the Win10 upgrade you can bet they won't do anything to make life on Win7 easier.
The reason why is because The Walt Disney Company owns both ABC and ESPN. So they force you to buy all the channels they own, or none. Nothing in between. Disney isn't the only one that does that of course, every company that owns multiple networks does. The only way that will change is if the US Government forces them to offer a la carte. The government would also have to force them to not set pricing such that the a la carte cost for 1 channel is the same as the cost of the bundle (maybe a legally mandated 10% max bundle discount or something.)
As the CEO of T-Mobile Germany stated, the "uncarrier" model that T-Mobile USA has been selling for the last couple years is unsustainable. To be fair, they were selling service so low it really was cheaper than it should have been. It was only a matter of time before a rate hike of some sort would happen. They are trying really really hard to spin the rate hike with marketing to make it sounds like a bigger deal than before with the whole "unlimited" wording, but truth is its no different than their previous plans, except for the extra cost. Don't let them convince you that the rate hike would not have happened had they been allowed to merge with Sprint by the way. I'm just glad I got in while they were still selling the ridiculously low price plans.
The article you link is misleading... The Intel chips being compared with the A9X in that article are Broadwell... 2 generations behind compared to the A9X which is the current chip. The A10X will have to be compete with Cannonlake, not the Skylake chips you see in the current Mac lineup. Assuming the A10X scales from the A10 about the same as the A9X scales from the A9, and Intel IPC generational improvements are about the same as past ones... single core ~6W TDP 2.2 GHz A10X is going to be close to raw single core performance of the ~4.5W TDP ~1.2 GHz Cannonlake, but the even the low power Cannonlake CPU will retain a commanding lead on multi core performance.
The one place where Apple ARM does actually dominate Intel is integrated graphics performance. It's incredible to me how Intel graphics are still so far behind everyone else. At least the recent Intel GPUs render mostly correct now and the graphics drivers are pretty stable at this point. The one advantage that Apple (and other ARM vendors) have on the GPU front that they don't need to implement DirectX + OpenGL in their silicon, a luxury that nVidia/AMD/Intel don't get.
Intel tried to do that a few years ago through the OEM channel when they were pushing aggressively in to the tablet market. Being able to run Windows and Android is a natural advantage for x86. Both Google AND Microsoft killed it by refusal to certify the resulting device. Even if Microsoft is cool with it now it is highly unlikely Google's stance has changed.
but once it works on full steam it munches close to ~90% of its TDP.
Are you seriously saying that if a workload start using 100% CPU would would rather Intel make its CPUs throttle instead of actually running at the frequency that you paid for?
I'm sorry but Intel is the LAST thing you can blame for the MBP battery life sucking. If it was Intel's fault then we would see it in all the Skylake Windows laptops, which we definitely do not (most Skylake laptops have 10+ hour battery life.)
Blame for this issue falls squarely on Apple. My best guess is its a combination of small battery capacity along with issues with their power management software. One area that would be particularly suspect is some problem with the switchable graphics control software being too aggressive with powering up the dGPU when its not actually needed. Or maybe there is a bug where they power it up but then in some cases forget to shut it down after it is no longer being used.
Taking the Lumia off the market is a very intentional maneuver on Microsoft's part. They are hoping that by the time they launch the Surface Phone in Q4 next year with full support for X86 Win32 apps on ARM that we will have largely forgetten about how much the previous Windows Phone underwhelmed us. Also making a short but clean break gives them an excuse not to upgrade all the phones on the market right now to the new OS that supports X86 Win32 apps.