I should have been a tad clearer in my post. The machine would physically sit at a location I (hopefully) control, so it would be in my physical possession. The reason for a hypervisor is so that the VM used for stashing archived mail would be able to be passed from bare metal to bare metal install as time goes on, without need to rebuild the system. It makes backups easy as well, where I just power the VM off, plug a USB drive into the host, mount a VeraCrypt volume, export the VM as a.OVA file, dismount the.hc file and drive, call it done. This isn't fancy, but snapshots taken often combined with monthly/quarterly exports to offsite media should cover things fairly reliably. If the data is vital, I toss it into one IMAP folder, encrypt that folder via PGP, GPG, VeraCrypt or some other brute-force resistant method, then toss it onto Amazon Glacier to rot as the backup of last resort.
TLAs are really not on my threat model, so I treat hypervisors the same as operating systems. However, I do like keeping communications with clients around for a period of time before dumping it, as a best practice, so I'm mainly concerned with an E-mail provider getting breached and wide swaths of users having their stuff made into torrents.
With OpenStack gaining momentum, RHEV is nice, but with libvirt, a CM, PXE booting and provisioning, it isn't really needed to get compute nodes up.
The one thing about Ansible which is nice is the fact that it doesn't require a client installed. Just throw a SSH key onto the client, and go from there. However, Puppet has more flexibility in the fact that you can run with "puppet master" servers and push out manifests, or masterless and have clients pull manifests from something like a local github install.
As of now, there are a lot of choices in the CM market. The next year or so will do a lot in weeding the competition down. I'd guess Puppet and SCOM/SCVMM/SCCM will survive, but after that, it will be a toss-up beteween Chef and Ansible in what is a mainstream CM.
The key will likely will be Windows support, especially when Windows Server 2016 hits RTM and brings with it a new generation of hypervisor features and distributed storage (think VMWare vSAN.)
I'm still using my Kindle 3/Kindle Keyboard because it has a matte screen, readable in almost any light, and one doesn't need to poke at the screen to turn pages. With cellular and Wi-Fi turned off, the device has a long battery life.
The funny thing is that we see basic, effective devices hit the market, like early Palm PDAs, e-readers, and other items which may not have a ton of bells and whistles... but do a single task very effectively. Then, they start getting stuff added, and wind up just being like everyone else, jammed full of junk, glossy screen, and their main function isn't that great.
When my Kindle Keyboard bites the dust, (likely when the battery gives up the ghost), choices are lackluster. Either buy a full-fledged tablet like a Nexus 9, or a touch screen e-Ink reader, and neither is really the best solution compared to the e-Ink readers offered just a few years back by Kobo, Sony, and Amazon.
The first thing I thought of when I saw these small pigs was to release them into the Texas wilds. If they wind up breeding with the feral hogs, they might just reduce the size (and thus hazard) of those things.
I take an easier approach. If I'm selling something I'll replace the drive.
However, for a machine I'm giving to a friend or family member, what I wind up doing is just a format command, then a pass with cipher/w (assuming Windows.) Since all my volumes are BitLocker protected, a format command overwrites the areas on the hard drive with the volume master key multiple times. Even with the right BitLocker password or recovery key protector, the data is gone, since the master key cannot be retrieved. The cipher/w just does a simple three pass (zeroes, ones, random numbers), which is good enough for almost anything.
SSDs are even easier. A format command zaps the keys, then I boot a Linux live CD, run hdparam to do a secure erase, or at the minimum, a blkdiscard on the entire drive, and call it done. The secure erase or the TRIM command ensures that all data on the drive is zeroed (or at least reported to the reading OS as zeroed), so there is almost no chance of recovery whatsoever. If by chance some data is recovered, it will just be encrypted stuff. If I wanted to, I could run an erase pass on the entire drive, but why shorten the drive's life when the secure erase or TRIM has pretty much ensured the drive will be clean.
Solar panels are going quite well. What would be nice is to see is battery capacity drop in price. Having charge controllers and inverters get cheaper, but still maintain the same level of quality and safety wouldn't be bad either.
Batteries are the weakest link in the solar equation. We get banks that are reasonably priced for individuals, have a long life, can handle charge/discharge cycles, and can store a decent amount of ampere-hours, and that will go a long way in helping with energy issues.
Of course, the ability to pull CO2 from the air and synthesize a fuel using solar wouldn't be bad either, especially if it were ethanol or a synthetic diesel. This would provide for long term storage in an energy-dense manner.
What I have wondered about is something plain and simple:
Canned water. Yes, cans are not perfect, but they are completely recyclable, and can store water indefinitely. Pretty much what Anheuser Busch did for a day for Texas flood victims, but a constant product [1].
Costs per item would be cheaper than soda water, as all that is needed is filtered tap water.
For an added bonus, add a 5-10 cent deposit onto each can. That will pretty much ensure they come back.
[1]: One could always make a joke about it not being a change from their beers, but what they did was a good thing, regardless.
Before updating/upgrading, make an image of the system with wbadmin (wbadmin.exe start backup with the usual options.) This way, you can recover not just the C: drive, but the recovery partition and others with ease.
Of course there is other Windows backup software, but virtually all of it is junk, except the enterprise stuff like NetBackup. Main reason is that most programs are unable to back up open files or make a usable snapshot image without booting from offline media.
AIX also has a "preservation install" which saves/home and non-OS logical volumes, but blows away everything else.
The ironic thing is that I have yet to actually need to have to use that feature with Linux. Usually it is some subsystem that gets trashed, so that is rebuilt. The exception is a security breach, and from there, I copy the data files off to removable/remote media, erase the machine completely, and install from scratch, so the chance malware remaining is extremely low.
You pretty much described it. My photos are never stored in a world accessible place, and if they are stored on the cloud, it is behind an encryption layer like BoxCryptor. Even though it doesn't mean much if the provider itself is compromised, 2FA goes without saying.
One can't control "leakage" like people popping pictures of you and tagging, but what doesn't go to a public forum doesn't get indexed, so just keeping vacation photos private is the best thing. Want to share them with friends? There are means to do it with others privately (well away from mass indexers), as opposed to tossing them onto a social networking site.
AFAIK, in California, they do that for older cars, newer ones get a device that plugs into the OBD II port and they log from that.
Texas is similar. If the vehicle has an OBD II port, they plug in their reader, pull the values from that and call it done. The only time exhaust checking might be actually used is if it is obvious someone did a custom tune (the OBD II port showing clean on a coal roller truck, for example.)
I remember one company that did a dog/pony show for an all-SSD SAN product that did this, although I forgot the name of the company.
Their SAN had fast Intel SSD for the landing zone where the data had one pass at being deduplicated. Then, there was a background task that deduplicated the data a lot better (but took more CPU power) and moved the data to slower, but higher capacity solid-state drives. Both the faster Intel SSD and the slower (but bigger in capacity) Samsungs would definitely have a place in something like this.
Only thing that made me scratch my head was the fact that (IIRC) they didn't factor in for failure on the landing zone SSD. If that failed, you would have to completely down that respective SAN controller to replaced the dead modules. At least they used some redundancy on the main storage array.
I can see another use for these fast SSD units. Windows Server 2016 offers Storage Spaces Direct, which is similar in functionality to VMWare's Virtual SAN, where it presents all the compute nodes' hard disks as one backing storage (in MS's case, a CSV replacement.) Having a SSD like this on all the nodes in a cluster will wind up being useful, just to handle the ton of random I/O that virtualization requires.
3D printers remind me of the beige box PC industry in the 1990s, bicycle parts makers in the 1990s (with everything CNC machined and anodized), and the inexpensive MP3 player market.
What I see is that a bunch of little guys are going to fight amongst themselves, and as soon as there are a few big players, some big company will swoop in, buy them out, and own the playing field, either a single company, or 2-3 firms (just like how paper printing is now, with just a relatively few companies offering models.)
One can be creative with 3D parts, but there is a limit that the plastic from the current generation can handle. At best, it is something to make to hone an injection mold from so "real" parts can be manufactured. Plus, the parts are rough, so they need sanded and coated with something like Smooth-On's epoxy if using them directly for a task.
Done right, it can be useful. Things like manufacturer MAC blocking and having one MAC per port is a way to ensure someone doesn't attach a switch or Wi-Fi AP to an internal network.
On some networks like POS networks, it is one extra security measure, just because someone can't unplug a cash register, plug in their laptop and go at that segment. Not foolproof (as one can figure out the POS's MAC and spoof it), but it does stop the guy who wants to plug into a network jack because the public store wireless is too slow for his video streaming. For networks that have more machines, MAC locking isn't worth the time, but for static networks, it can be a help, similar to a lock on the bank doors leading to the vault.
My sentiments exactly. There is a lot more someone can do who has physical access, is willing to face felony level malicious mischief charges and is willing to end any chance of a meaningful career in IT (heck, a meaningful career anywhere, for that matter.) Just walking up to a rack and yanking all drives out will bring a data center to its knees. Yes, some data centers actually take the time to use the locks on the equipment, but most don't bother since the locks tend to be engineered to hold plastic bezels in place and provide nuisance protection as opposed to actual physical attack resistance.
This is absolutely nothing new. Back in the early 1990s, I worked with a guy who had "adapters" which were 120VAC to coax Ethernet, 120VAC to serial, 120VAC to thicknet, and 120VAC to SCSI.
One place I worked at had someone use customized surge suppressors on Ethernet drops that went from a public area to a private area, because they were afraid of this.
This is nothing new... This is in the same category of stuff like sticking blobs of Superglue into the locks on a building as part of a "denial of service" attack.
These days, the fix is easy... if really worried and wireless isn't an option, go with single mode fiber if concerned that someone is going to use a network drop for an attack. If someone blows out the NIC on the other end with a 100+ laser, it will only blow out the SFP.
What Tesla brought to the table was making electric cars good looking, cool, fast, and fun to drive. Before that, we did have electric cars. However, they looked like the ZAP Sparkee, cute, little, underpowered bubble things with a range of footsteps. Tesla brought interstate travel to the table for electric cars. They also got places to install electric charging stations, and legitimized people plugging into the wall at stores and such [1], which was considered theft previously.
Tesla definitely doesn't sell cars like Toyota... but for what they offer sets a standard for other automakers to follow. Things like vehicle updates that add features, even for vehicles several years old, decent service (even in areas where they are forbidden to sell vehicles), a very good safety record, and excellent customer service. Plus, when you pop the hood of a Tesla Model S, it is awesome -- another place to toss suitcases and other items.
[1]: Well, except for Alaska where stores and other places have outlets to plug into to keep vehicle heaters going.
The trick is to tag an expiration date on all info. John Doe tells the robot about his drink preferences, and the robot will retain those preferences either until the drinks are served and the tab closed, or until there is a certain point in time, where the drink preference info is flagged to expire. Every so often, a garbage collector task runs, purges all robot preferences that are expired and not flagged for retention [1].
In general, expiration timestamps might be something to have in a database row, because when combined with a garbage collection task, it ensures that data will be tossed without having to actively go and delete it. Backup systems do this already, where if I don't flag a backup snapshot, after a certain time, snapshots expire, and eventually get overwritten.
[1]: A transaction could be flagged for retention if the dining parties decide to checkwalk, for example. Of course, this can be abused by setting the threshold for retaining transactions extremely low, but it should be in place if need be.
The problem is that on Marshmallow, the dev has to have a manifest allowing for the switched permissions to happen included with the APK file, otherwise, it will default to its traditional all or nothing.
As mentioned elsewhere, I will plug xPrivacy. If an app wants everything under the sun, it can have it. The camera will happily give black frames, the app can slurp all the video from/dev/null it can handle, the user is always located in the same spot, and the IMEI/UDID is always different each time it is asked.
On iOS, I recommend PMP, but that requires a jailbreak.
We also have long since passed peak coal. Most coal plants are burning lignite coal, which is a small step up from peat moss, which is the dirtiest, most impure coal one can get.
A battery makes sense. It may take some cost to build, but plates and contents are recyclable, and all the amount of pollutants and CO2 are costs which can't be measured.
Here in Austin, I see more Teslas than I do Leafs, however I am starting to see those BMW i3s start popping up.
The one thing I wish Texas has, and Toyota doesn't sell them here are plug-in Priuses.
I wouldn't be surprised to see things go forward and vehicles like pickup trucks start having hybrid drivetrains. For a one ton truck, electric motors will be very useful, especially because their torque is at peak at 0 RPM, and that is where it is needed for towing, and higher RPMs, the gasser or diesel engine can take it from there. Of course, there are practical uses too. Stick a heavy duty inverter that feeds from a truck's batteries, and now you don't need a generator or a PTO driven generator head for construction work in a rural area, or use the batteries for power (even A/C) as opposed to using a a generator.
Or take a Volt and extend that concept, where the drivetrain and engine are electric, but there is a generator that can be of whatever fuel type the customer specifies which will fire on when the batteries get low. This way, only a relatively small subsystem needs changed when fuel needs change.
There are still a ton of Teslas on the roads here (well, namely Austin.) People just buy out of state and bring them here, and Austin has a repair depot for them, so one isn't SOL when they need maintenance.
I do agree, Texas needs to get with the times... if someone is going to buy a Tesla, they will, and the TT&L money is going to go to Texas or another state.
Windows has had the Desktops application and the PowerToys suite before that, which allowed for virtual desktops for decades now (since the NT 4.0 days.)
It is front and center in Windows 10, but it isn't really anything that wasn't able to be fetched before.
As for what I do, I use virtualization a lot, so instead of virtual desktops, I use the column selector to pick the VM I want to use, and go with that. Yes, there is definitely the performance hit (mainly I/O, which can be mitigated by a good amount of RAM and a SSD because multiple operating systems do lots of random reads/writes), but the advantage is separation and security. The VM holding Facebook, if that gets compromised, isn't going to affect the VM with Quicken/Quickbooks/TurboTax.
The ideal fuel wouldn't be hydrogen, but something like propane that is relatively easy to store, is not a greenhouse gas if it leaks, and takes a proper oxygen/fuel ratio to ignite as opposed to being set off by virtually anything.
The ideal would be ethanol. It isn't toxic like methanol, has a decent energy per unit volume (not as good as gasoline or diesel, but not horri-bad.) Alcohol is somewhat corrosive, but nothing that can't be engineered around, and in Brazil, this is quite a solved problem.
Part of the definition should be software that sends or alters data and the machine configuration from a user's machine without explicit authorization and without a direct, primary purpose beneficial to the user.
Something like VirusTotal where a user scans a file against a good amount of AV programs passes these two tests. It has a primary reason to grab and upload a file, and the user explicitly uploaded it.
Browser fingerprinting software, update "services", loopback tunneling services to MITM SSL, and many other items do not benefit the user, nor are they often even authorized, so they are thus malware.
The gray area are processes like Blizzard's Warden and Valve's VAC which scan and upload data to see if a user is cheating.
Slashdot Mirror
I should have been a tad clearer in my post. The machine would physically sit at a location I (hopefully) control, so it would be in my physical possession. The reason for a hypervisor is so that the VM used for stashing archived mail would be able to be passed from bare metal to bare metal install as time goes on, without need to rebuild the system. It makes backups easy as well, where I just power the VM off, plug a USB drive into the host, mount a VeraCrypt volume, export the VM as a .OVA file, dismount the .hc file and drive, call it done. This isn't fancy, but snapshots taken often combined with monthly/quarterly exports to offsite media should cover things fairly reliably. If the data is vital, I toss it into one IMAP folder, encrypt that folder via PGP, GPG, VeraCrypt or some other brute-force resistant method, then toss it onto Amazon Glacier to rot as the backup of last resort.
TLAs are really not on my threat model, so I treat hypervisors the same as operating systems. However, I do like keeping communications with clients around for a period of time before dumping it, as a best practice, so I'm mainly concerned with an E-mail provider getting breached and wide swaths of users having their stuff made into torrents.
With OpenStack gaining momentum, RHEV is nice, but with libvirt, a CM, PXE booting and provisioning, it isn't really needed to get compute nodes up.
The one thing about Ansible which is nice is the fact that it doesn't require a client installed. Just throw a SSH key onto the client, and go from there. However, Puppet has more flexibility in the fact that you can run with "puppet master" servers and push out manifests, or masterless and have clients pull manifests from something like a local github install.
As of now, there are a lot of choices in the CM market. The next year or so will do a lot in weeding the competition down. I'd guess Puppet and SCOM/SCVMM/SCCM will survive, but after that, it will be a toss-up beteween Chef and Ansible in what is a mainstream CM.
The key will likely will be Windows support, especially when Windows Server 2016 hits RTM and brings with it a new generation of hypervisor features and distributed storage (think VMWare vSAN.)
I'm still using my Kindle 3/Kindle Keyboard because it has a matte screen, readable in almost any light, and one doesn't need to poke at the screen to turn pages. With cellular and Wi-Fi turned off, the device has a long battery life.
The funny thing is that we see basic, effective devices hit the market, like early Palm PDAs, e-readers, and other items which may not have a ton of bells and whistles... but do a single task very effectively. Then, they start getting stuff added, and wind up just being like everyone else, jammed full of junk, glossy screen, and their main function isn't that great.
When my Kindle Keyboard bites the dust, (likely when the battery gives up the ghost), choices are lackluster. Either buy a full-fledged tablet like a Nexus 9, or a touch screen e-Ink reader, and neither is really the best solution compared to the e-Ink readers offered just a few years back by Kobo, Sony, and Amazon.
The first thing I thought of when I saw these small pigs was to release them into the Texas wilds. If they wind up breeding with the feral hogs, they might just reduce the size (and thus hazard) of those things.
I take an easier approach. If I'm selling something I'll replace the drive.
However, for a machine I'm giving to a friend or family member, what I wind up doing is just a format command, then a pass with cipher /w (assuming Windows.) Since all my volumes are BitLocker protected, a format command overwrites the areas on the hard drive with the volume master key multiple times. Even with the right BitLocker password or recovery key protector, the data is gone, since the master key cannot be retrieved. The cipher /w just does a simple three pass (zeroes, ones, random numbers), which is good enough for almost anything.
SSDs are even easier. A format command zaps the keys, then I boot a Linux live CD, run hdparam to do a secure erase, or at the minimum, a blkdiscard on the entire drive, and call it done. The secure erase or the TRIM command ensures that all data on the drive is zeroed (or at least reported to the reading OS as zeroed), so there is almost no chance of recovery whatsoever. If by chance some data is recovered, it will just be encrypted stuff. If I wanted to, I could run an erase pass on the entire drive, but why shorten the drive's life when the secure erase or TRIM has pretty much ensured the drive will be clean.
Solar panels are going quite well. What would be nice is to see is battery capacity drop in price. Having charge controllers and inverters get cheaper, but still maintain the same level of quality and safety wouldn't be bad either.
Batteries are the weakest link in the solar equation. We get banks that are reasonably priced for individuals, have a long life, can handle charge/discharge cycles, and can store a decent amount of ampere-hours, and that will go a long way in helping with energy issues.
Of course, the ability to pull CO2 from the air and synthesize a fuel using solar wouldn't be bad either, especially if it were ethanol or a synthetic diesel. This would provide for long term storage in an energy-dense manner.
What I have wondered about is something plain and simple:
Canned water. Yes, cans are not perfect, but they are completely recyclable, and can store water indefinitely. Pretty much what Anheuser Busch did for a day for Texas flood victims, but a constant product [1].
Costs per item would be cheaper than soda water, as all that is needed is filtered tap water.
For an added bonus, add a 5-10 cent deposit onto each can. That will pretty much ensure they come back.
[1]: One could always make a joke about it not being a change from their beers, but what they did was a good thing, regardless.
Before updating/upgrading, make an image of the system with wbadmin (wbadmin.exe start backup with the usual options.) This way, you can recover not just the C: drive, but the recovery partition and others with ease.
Of course there is other Windows backup software, but virtually all of it is junk, except the enterprise stuff like NetBackup. Main reason is that most programs are unable to back up open files or make a usable snapshot image without booting from offline media.
AIX also has a "preservation install" which saves /home and non-OS logical volumes, but blows away everything else.
The ironic thing is that I have yet to actually need to have to use that feature with Linux. Usually it is some subsystem that gets trashed, so that is rebuilt. The exception is a security breach, and from there, I copy the data files off to removable/remote media, erase the machine completely, and install from scratch, so the chance malware remaining is extremely low.
You pretty much described it. My photos are never stored in a world accessible place, and if they are stored on the cloud, it is behind an encryption layer like BoxCryptor. Even though it doesn't mean much if the provider itself is compromised, 2FA goes without saying.
One can't control "leakage" like people popping pictures of you and tagging, but what doesn't go to a public forum doesn't get indexed, so just keeping vacation photos private is the best thing. Want to share them with friends? There are means to do it with others privately (well away from mass indexers), as opposed to tossing them onto a social networking site.
AFAIK, in California, they do that for older cars, newer ones get a device that plugs into the OBD II port and they log from that.
Texas is similar. If the vehicle has an OBD II port, they plug in their reader, pull the values from that and call it done. The only time exhaust checking might be actually used is if it is obvious someone did a custom tune (the OBD II port showing clean on a coal roller truck, for example.)
I remember one company that did a dog/pony show for an all-SSD SAN product that did this, although I forgot the name of the company.
Their SAN had fast Intel SSD for the landing zone where the data had one pass at being deduplicated. Then, there was a background task that deduplicated the data a lot better (but took more CPU power) and moved the data to slower, but higher capacity solid-state drives. Both the faster Intel SSD and the slower (but bigger in capacity) Samsungs would definitely have a place in something like this.
Only thing that made me scratch my head was the fact that (IIRC) they didn't factor in for failure on the landing zone SSD. If that failed, you would have to completely down that respective SAN controller to replaced the dead modules. At least they used some redundancy on the main storage array.
I can see another use for these fast SSD units. Windows Server 2016 offers Storage Spaces Direct, which is similar in functionality to VMWare's Virtual SAN, where it presents all the compute nodes' hard disks as one backing storage (in MS's case, a CSV replacement.) Having a SSD like this on all the nodes in a cluster will wind up being useful, just to handle the ton of random I/O that virtualization requires.
3D printers remind me of the beige box PC industry in the 1990s, bicycle parts makers in the 1990s (with everything CNC machined and anodized), and the inexpensive MP3 player market.
What I see is that a bunch of little guys are going to fight amongst themselves, and as soon as there are a few big players, some big company will swoop in, buy them out, and own the playing field, either a single company, or 2-3 firms (just like how paper printing is now, with just a relatively few companies offering models.)
One can be creative with 3D parts, but there is a limit that the plastic from the current generation can handle. At best, it is something to make to hone an injection mold from so "real" parts can be manufactured. Plus, the parts are rough, so they need sanded and coated with something like Smooth-On's epoxy if using them directly for a task.
Done right, it can be useful. Things like manufacturer MAC blocking and having one MAC per port is a way to ensure someone doesn't attach a switch or Wi-Fi AP to an internal network.
On some networks like POS networks, it is one extra security measure, just because someone can't unplug a cash register, plug in their laptop and go at that segment. Not foolproof (as one can figure out the POS's MAC and spoof it), but it does stop the guy who wants to plug into a network jack because the public store wireless is too slow for his video streaming. For networks that have more machines, MAC locking isn't worth the time, but for static networks, it can be a help, similar to a lock on the bank doors leading to the vault.
My sentiments exactly. There is a lot more someone can do who has physical access, is willing to face felony level malicious mischief charges and is willing to end any chance of a meaningful career in IT (heck, a meaningful career anywhere, for that matter.) Just walking up to a rack and yanking all drives out will bring a data center to its knees. Yes, some data centers actually take the time to use the locks on the equipment, but most don't bother since the locks tend to be engineered to hold plastic bezels in place and provide nuisance protection as opposed to actual physical attack resistance.
This is absolutely nothing new. Back in the early 1990s, I worked with a guy who had "adapters" which were 120VAC to coax Ethernet, 120VAC to serial, 120VAC to thicknet, and 120VAC to SCSI.
One place I worked at had someone use customized surge suppressors on Ethernet drops that went from a public area to a private area, because they were afraid of this.
This is nothing new... This is in the same category of stuff like sticking blobs of Superglue into the locks on a building as part of a "denial of service" attack.
These days, the fix is easy... if really worried and wireless isn't an option, go with single mode fiber if concerned that someone is going to use a network drop for an attack. If someone blows out the NIC on the other end with a 100+ laser, it will only blow out the SFP.
What Tesla brought to the table was making electric cars good looking, cool, fast, and fun to drive. Before that, we did have electric cars. However, they looked like the ZAP Sparkee, cute, little, underpowered bubble things with a range of footsteps. Tesla brought interstate travel to the table for electric cars. They also got places to install electric charging stations, and legitimized people plugging into the wall at stores and such [1], which was considered theft previously.
Tesla definitely doesn't sell cars like Toyota... but for what they offer sets a standard for other automakers to follow. Things like vehicle updates that add features, even for vehicles several years old, decent service (even in areas where they are forbidden to sell vehicles), a very good safety record, and excellent customer service. Plus, when you pop the hood of a Tesla Model S, it is awesome -- another place to toss suitcases and other items.
[1]: Well, except for Alaska where stores and other places have outlets to plug into to keep vehicle heaters going.
The trick is to tag an expiration date on all info. John Doe tells the robot about his drink preferences, and the robot will retain those preferences either until the drinks are served and the tab closed, or until there is a certain point in time, where the drink preference info is flagged to expire. Every so often, a garbage collector task runs, purges all robot preferences that are expired and not flagged for retention [1].
In general, expiration timestamps might be something to have in a database row, because when combined with a garbage collection task, it ensures that data will be tossed without having to actively go and delete it. Backup systems do this already, where if I don't flag a backup snapshot, after a certain time, snapshots expire, and eventually get overwritten.
[1]: A transaction could be flagged for retention if the dining parties decide to checkwalk, for example. Of course, this can be abused by setting the threshold for retaining transactions extremely low, but it should be in place if need be.
The problem is that on Marshmallow, the dev has to have a manifest allowing for the switched permissions to happen included with the APK file, otherwise, it will default to its traditional all or nothing.
As mentioned elsewhere, I will plug xPrivacy. If an app wants everything under the sun, it can have it. The camera will happily give black frames, the app can slurp all the video from /dev/null it can handle, the user is always located in the same spot, and the IMEI/UDID is always different each time it is asked.
On iOS, I recommend PMP, but that requires a jailbreak.
We also have long since passed peak coal. Most coal plants are burning lignite coal, which is a small step up from peat moss, which is the dirtiest, most impure coal one can get.
A battery makes sense. It may take some cost to build, but plates and contents are recyclable, and all the amount of pollutants and CO2 are costs which can't be measured.
I applaud any steps in the direction of less CO2.
Here in Austin, I see more Teslas than I do Leafs, however I am starting to see those BMW i3s start popping up.
The one thing I wish Texas has, and Toyota doesn't sell them here are plug-in Priuses.
I wouldn't be surprised to see things go forward and vehicles like pickup trucks start having hybrid drivetrains. For a one ton truck, electric motors will be very useful, especially because their torque is at peak at 0 RPM, and that is where it is needed for towing, and higher RPMs, the gasser or diesel engine can take it from there. Of course, there are practical uses too. Stick a heavy duty inverter that feeds from a truck's batteries, and now you don't need a generator or a PTO driven generator head for construction work in a rural area, or use the batteries for power (even A/C) as opposed to using a a generator.
Or take a Volt and extend that concept, where the drivetrain and engine are electric, but there is a generator that can be of whatever fuel type the customer specifies which will fire on when the batteries get low. This way, only a relatively small subsystem needs changed when fuel needs change.
There are still a ton of Teslas on the roads here (well, namely Austin.) People just buy out of state and bring them here, and Austin has a repair depot for them, so one isn't SOL when they need maintenance.
I do agree, Texas needs to get with the times... if someone is going to buy a Tesla, they will, and the TT&L money is going to go to Texas or another state.
Windows has had the Desktops application and the PowerToys suite before that, which allowed for virtual desktops for decades now (since the NT 4.0 days.)
It is front and center in Windows 10, but it isn't really anything that wasn't able to be fetched before.
As for what I do, I use virtualization a lot, so instead of virtual desktops, I use the column selector to pick the VM I want to use, and go with that. Yes, there is definitely the performance hit (mainly I/O, which can be mitigated by a good amount of RAM and a SSD because multiple operating systems do lots of random reads/writes), but the advantage is separation and security. The VM holding Facebook, if that gets compromised, isn't going to affect the VM with Quicken/Quickbooks/TurboTax.
The ideal fuel wouldn't be hydrogen, but something like propane that is relatively easy to store, is not a greenhouse gas if it leaks, and takes a proper oxygen/fuel ratio to ignite as opposed to being set off by virtually anything.
The ideal would be ethanol. It isn't toxic like methanol, has a decent energy per unit volume (not as good as gasoline or diesel, but not horri-bad.) Alcohol is somewhat corrosive, but nothing that can't be engineered around, and in Brazil, this is quite a solved problem.
Part of the definition should be software that sends or alters data and the machine configuration from a user's machine without explicit authorization and without a direct, primary purpose beneficial to the user.
Something like VirusTotal where a user scans a file against a good amount of AV programs passes these two tests. It has a primary reason to grab and upload a file, and the user explicitly uploaded it.
Browser fingerprinting software, update "services", loopback tunneling services to MITM SSL, and many other items do not benefit the user, nor are they often even authorized, so they are thus malware.
The gray area are processes like Blizzard's Warden and Valve's VAC which scan and upload data to see if a user is cheating.