> I doubt very much the company is going to survive this
I'm sick of hearing this stuff. I heard it about virtually every Trump-related news story for the last year, I heard it about New Orleans, I heard it about BP. I heard it about Volkswagen.
Look, the market already priced this disaster, based on what data is available. They lost about 15% of their value. They have solid fundamentals that aren't changing in any sort of way that pose an existential threat to the company; unless there's a lot more of this than we can see.
It doesn't look like anything spectacular to me, but the tree placement, giant warehouse, and that it's on the side of a hill make me wonder if it's just a disposal site of some sort.
If it's anything of importance, I'll bet it's a nuclear or toxic waste disposal site.
I always get frustrated when I have to dig deep and manually read over scanned non-OCR'd PDFs.
Honestly, all I want is for all the documents and laws to be available. I would think that SVN with a web component would be great to be able to see the current laws, including the ability to 'go back in time', and it would work for posting minutes and agendas as well.
I'm not overly concerned with -presentation-, having the data available in a digital form alone is a major step forward, and you can build out from there.
The difference is that most network admins shirk from the task of responsibly implementing QoS, but they'd gladly pay a hefty licensing fee to their wireless vendors for a product with a name like WiFox that 'boosts performance' by clobbering the network instead of cleverly balancing it to perform well.
And a reason for poor uptake of Vista over 7 was immature image deployment and customization tools. Half the options weren't documented, and the process was very unclear. By the time 7 came out, there was enough backfill on the documentation and examples online for people like me to actually work on customizing corporate images.
This is actually something that's woefully inadequate in the Linux world, real-world examples of the cool stuff you can do with HOWTOs for building a server and clients that are running off of shared volumes and centralized authentication. All the parts are there, and they work fantastically, but you sort of have to figure it out yourself how to put it all together.
This isn't why Windows 7 is 6.1, or why Windows 8 is 6.2.
The reason is that Windows 7 actually is just a minor revision on Vista, and 8 is a minor revision from that. Under the hood, the big changes were between NT 4.0 and Windows 2000 (Windows NT 5), then between 2000 and Vista (Windows 6). The changes from 5.0 to 5.1 (2000 to XP) or from 6.0 to 6.1 to 6.2 (Vista, 7, and 8) were incremental in nature as far as the inner workings of the OS are concerned.
The real reason 7 felt so much faster than Vista: When they made Vista, they planned on you booting up very infrequently, so they scheduled a lot of junk to happen at boot and login, thinking that users would just 'sleep' instead of rebooting. Windows 7 (And Vista SP2) backs off a bit and does the housekeeping when you're not using the computer. Vista actually wasn't really 'slow', it's just 'irrationally busy' doing stuff with the I/O (indexing, precaching, defragmenting, etc.) while you're just trying to get to your gosh-darned desktop.
So your de-dupe ran for a week before you cut it out? On a modern CPU, the de-dupe is limited not by the CPU speed (since deduplication basically just checksums blocks of storage), but by the speed of the drives.
What you need to do is put all this data onto a single RAID10 array with high IO performance. 5TB of data, plus room to grow on a RAID10 with decent IOPS would probably be something like 6 3TB SATA drives on a new array controller. Set up the array with a large stripe size to prioritize reads (writes are going to be 'fast enough' on a RAID10, trust me). Once you have that hooked-up with your files copied onto it, you want to connect the drive to an OS that can natively deduplicate, like Windows Server 2012. If you must, you can set this box up as a storage server (with a low-end CPU, an old 'Core 2' should be able to keep up with 180MB/sec I/O), and keep your workstation separate. Reading this entire array (when full) through the CPU -should- take about 6-10 hours, deduplication will take slightly longer.
If you don't want to do deduplication at the block level, and you want to actually only have one copy of each duplicated file, you'll need to write scripts that do something like this:
1. Run through the data store and checksum each file (except for those ending in ".mychecksum" with AES128. 2. For each file, create an empty file named.."mychecksum" next to it. This will create the 'index' using the filesystem, which will be MUCH faster than having to read the data from inside each file. 3. Search through the store and concatenate all the ".mychecksum" files into a single CSV. 4. Run sed+unique on the file to see what will be nixed (i.e. Get a report) 5. Create another script that actually takes the output from step 4 and deletes ONE of the duplicate files. You can test by -renaming- ONE of the files to.deleteme and then deleting all those files after you confirm that it worked. 6. Repeat as necessary, possibly with a scheduled job.
I'm in the northeast, a very low-end power user (bottom quartile), and the math still doesn't work out for me on PV. What DOES work is solar thermal to warm up a tank in the basement that sits before the hot water heaters (preheating water) and pumping heat into the living space via baseboard radiators. Unfortunately, those systems are not as cookie-cutter, so getting someone to put them in is almost impossible.
Unfortunately, all the energy saving stuff I see seems geared for newer homes or homes in sunny and hot areas. Where I am, people generally don't even have Air Conditioning, they have 110 year-old homes that aren't well insulated (and often can't be without $15K of asbestos remediation and $5K of rewiring, neither is subsidized). I have yet to meet a contractor who understands that I want windows on the south and that ALLOW lots of infrared in.
I've been the 'young buck' in IT at every job I've had, but I'm starting to get a bit older now. I'm actually the guy who builds images for a university. What you have is important, but an IT department can't just be filled with young folks. I rely on the same kind of people you work with to *actually deploy the work* that I do, and fix the computers when they break, and file tickets when my stuff has bugs in it. It's frustrating having to deal with people who are always working around bugs or limitations that no longer exist, but I can't take on all the responsibility myself.
What I would suggest in your situation is to ask your manager or their manager to give you some time to build newer images, using the tools Microsoft gives people like us to work with modern hardware. Make it a project to modernize imagine deployment. Learn some Powershell so you can write some kick-ass 'firstrun' and management scripts. Investigate how to do things like app virtualization, user data redirection, and remote BIOS management that will make things better for everyone. And most of all, DOCUMENT the processes for the old-timers so they're able to do the things that you figure out.
If you want, you can contact me off-list and I can help you get started with Windows Imaging and some handy scripts we use.
I have my PC (a Lenovo ThinkServer TS130, 16GB RAM, dual HDD, Quadro 600 video card, Sandy Bridge i7) plugged-in via a kill-a-watt. It idles at about 50W. When it's crunching away (compressing video on 8 threads, virtualizing three computers, and playing a movie), it uses about 220w.
The fact is that gamers tend to buy big PSUs because they use them as a pissing-contest. Few but the true monster rigs use more than 300W. Even high-end workstations from the Pentium 4 era came with 375W PSUs, and they had plenty of overhead for fully-decking the system with SLI video cards and RAID arrays.
I was actually shopping for a 160W PSU for a Core 2 workstation that I'm looking to give away recently, I never managed to get the thing to use more than 140W, and it seemed silly to have a 400W PSU sitting there being inefficient all the time. I ended up using a PSU from mini-box that uses an AC->DC brick.
People need to understand and respect just how awesome water is as a coolant. The specific heat of the stuff (basically, how much heat you can 'sink' into a gram of it) and its benign, well-understood nature, and the fact that its density only changes a little bit between freezing and boiling points make it quite awesome.
I live in a city with a river through it. I really don't know why they aren't doing cooling via air-to-water heat pumps. It's really absurd to blow fans all day when the river could carry away 100X the heat without too many ill effects.
Actually, there is. Microsoft is prohibited from doing this on x86-based stuff by the antitrust agreement they're in. They can't prohibit vendors from allowing other operating systems and locking-in 'wintel' hardware/software. The agreement does NOT cover ARM.
> the kernel changes too damned fast and the key would be tied to a particular kernel
That's not true, and it's not how it would work.
Your distribution will be able to package signed bootloaders (GRUB) and kernels. The kernels from your distribution should work fine as long as they Play The Game (which I suspect some, like Debian, will not).
As for the 'kernel changing too damned fast': First, you should know that Windows kernels are ALSO updated between service packs, the Windows kernel components change fairly frequently. Second, the kinds of distributions that get used in managed environments (like RedHat) don't get changed often at all, maybe twice a year. RHEL is still running patched revisions of 2.6.32 from January 2010.
I suspect that in the future, we'll have to 'root' our computers just like we have to root our phones today to load alternate firmware.
> Do you really want to connect all the timing devices in a hospital to an outside public server?
Yes. Because NTP as a client is totally secure. Really. It is.
I'd rather take the one-in-a-billion chance that there's a way to somehow do something bad to a client device over the internet than definitely have my devices run in a messed-up state all the time.
From a risk-management perspective, using NTP to a trusted set of external hosts is a clear win.
Just so you know, Windows Vista and newer have solved this problem. You can push updates AND drivers into offline images. You can put a base image on a server and just load drivers from a database. We cut our number of images from 26 to 1, and it never needs more than three months of Windows Updates.
And it's not just for corporate use, the tools are in the Windows ADK, you just need to script up a recursive driver add after you imagex the machine.
I had written up scripts that automatically took an XP install all the way to where I wanted it to be, and these new tools just blow mine away.
I've had much better luck finding out where they hang out (bars, usually) after a session, then I bring them a small check, made out to their campaign. Once that happens, they usually give you their -real- email address or phone number.
I bought a state rep about $50 in drinks one night, cut him a check the next day, and my ideas on Net Metering made it into the next revision of the bill. I did the same for a city councilman, who is now using a few of my ideas to save money.
The great thing about contribution limits, which are usually under $1,000 per-contributor, is that you get a lot of bang-for-your-buck for a $100 or $200 contribution.
I live in a state that's had 80 years of solid Democrat rule. The district I live in has a student:teacher ratio of 12:1 and we spend over $15K per student annually. Yet despite this, the buildings are literally falling apart, classes are fully-packed, supplies, computers, and books are hard to get.
The problem isn't spending. We spend plenty. The problem is that there's no incentive to produce better outputs. We tried raising salaries, spending more in general. It didn't help.
"I'd prefer it if they left the decision between me and my doctor...."
Just to be coy, that translates to me as:
"I'd prefer it if they left the decision between someone who won't have to pay for something and the guy who wants to get paid...."
Arguably, the insurance company (or the government, in single-payer) is the ONLY party involved with enough information to make the proper decision about what level of care you receive.
I've been saying this all along. The answer for these companies is not to cap or throttle, it's to behave like a good citizen on the internet and either peer with or colocate the data customers want.
Now imagine if Google, Apple, Amazon, and Netflix could host a few boxes inside the Comcast network. Everyone wins. Unfortunately, that's just not how higher-ups in most organizations think.
I have a fair number of pictures, but those are small. This guy has a multi-terabyte thing going; between that and his username, I assume he has a LOT of video. In the case of video, it's time to say 'I won't need all this raw footage in DV format again'. It's OK to compress the video of the birthday party down to 1080p h.264 at medium quality, toss the project files, and let history remember you that way.
I think it's time to admit that you're a hoarder. What exactly -is- your personal data that's so precious? I run a server just to keep my skill set up and run my side business, but I've only managed to accumulate around 600GB of data, only about 35GB of it is 'mine', the rest is client backups.
So first admit that you're a hoarder, then decide if you wan to address that issue or indulge it. If you choose to indulge it, you're going to want to build a small home server. Something with a low-end 64-bit CPU (i3?), a gigabit LAN port, and lots SATA ports and 3.5" drive bays. Buy a bunch of high-quality (WD RE4?) matching drives that fit your data needs times two (you're RAIDing space away). Once you have that, install Linux on it, build a software RAID-1 or 0+1 array (don't do RAID-5 unless you can handle days of rebuild time), and format it with something accessible (read: in the kernel, like EXT4). Create a share on the array with Samba and happily access it from all your machines (don't bother with Netatalk or NFS; CIFS is great on all platforms). As your data needs grow, you can add drives in pairs or replace drives with larger ones and grow the volume. If you need backup, you'll want another array, preferably on another low-end box (an enclosure on your desktop?) but it can be built on a RAID-0 or JBOD to save money.
Mac OS X programs mostly are 'fat binaries'. If you right-click and 'view contents' of an application you'll see that it's really just a directory filled with files. You'll often see that there are x86, x64, and PPC binaries in the same package, sharing the same 'resource' files (which are now actually files instead of HFS magic).
It would be trivial to add more architectures to Mac OS X. Basically all you need is a compiler (which already exists) and for developers to actually target it.
I'm not actually sure that would be of much use. A trip from the northeast to California would be at least 15 hours, and a trip from New York to Florida would be 8 hours. There is no price point that makes a trip that long 'worth it' to do.
Regional high-speed rail would be awesome, though. It would be awesome to be able to easily commute to/from New York City on a daily basis from Boston or D.C.. That would have very real positive economic consequences for the entire eastern seaboard.
Similar story if you could link San Francisco and Los Angeles in under two hours.
You basically have to beat planes on time AND price to make it work. And you obviously have to beat the pants off of cars, because people WILL drive if it's not much faster to take the train.
I have a friend who works at a facility that runs those very exercises. You make it sound like that was the end of the story. In reality, we re-ran the exercises again because you don't just walk away from expensive simulations like that; after a loss, you create a team to re-tool for a few years in the future that prevents that kind of loss again.
Today, that same strategy wouldn't work. We upgraded the anti-boat and anti-missile technologies. We have been running simulations against massive numbers of speedboats for a decade now, successfully.
I'm not saying we wouldn't take losses or that they wouldn't be significant, but I don't think the US Military would stand by and allow us to fall into utter complacency, there are a LOT of checks and balances preventing that. 'Safety' seems to be the one thing our government can do correctly, probably because we've established that it has an unlimited price ceiling.
Slashdot Mirror
> I doubt very much the company is going to survive this
I'm sick of hearing this stuff. I heard it about virtually every Trump-related news story for the last year, I heard it about New Orleans, I heard it about BP. I heard it about Volkswagen.
Look, the market already priced this disaster, based on what data is available. They lost about 15% of their value. They have solid fundamentals that aren't changing in any sort of way that pose an existential threat to the company; unless there's a lot more of this than we can see.
It doesn't look like anything spectacular to me, but the tree placement, giant warehouse, and that it's on the side of a hill make me wonder if it's just a disposal site of some sort.
If it's anything of importance, I'll bet it's a nuclear or toxic waste disposal site.
I always get frustrated when I have to dig deep and manually read over scanned non-OCR'd PDFs.
Honestly, all I want is for all the documents and laws to be available. I would think that SVN with a web component would be great to be able to see the current laws, including the ability to 'go back in time', and it would work for posting minutes and agendas as well.
I'm not overly concerned with -presentation-, having the data available in a digital form alone is a major step forward, and you can build out from there.
The difference is that most network admins shirk from the task of responsibly implementing QoS, but they'd gladly pay a hefty licensing fee to their wireless vendors for a product with a name like WiFox that 'boosts performance' by clobbering the network instead of cleverly balancing it to perform well.
And a reason for poor uptake of Vista over 7 was immature image deployment and customization tools. Half the options weren't documented, and the process was very unclear. By the time 7 came out, there was enough backfill on the documentation and examples online for people like me to actually work on customizing corporate images.
This is actually something that's woefully inadequate in the Linux world, real-world examples of the cool stuff you can do with HOWTOs for building a server and clients that are running off of shared volumes and centralized authentication. All the parts are there, and they work fantastically, but you sort of have to figure it out yourself how to put it all together.
This isn't why Windows 7 is 6.1, or why Windows 8 is 6.2.
The reason is that Windows 7 actually is just a minor revision on Vista, and 8 is a minor revision from that. Under the hood, the big changes were between NT 4.0 and Windows 2000 (Windows NT 5), then between 2000 and Vista (Windows 6). The changes from 5.0 to 5.1 (2000 to XP) or from 6.0 to 6.1 to 6.2 (Vista, 7, and 8) were incremental in nature as far as the inner workings of the OS are concerned.
The real reason 7 felt so much faster than Vista: When they made Vista, they planned on you booting up very infrequently, so they scheduled a lot of junk to happen at boot and login, thinking that users would just 'sleep' instead of rebooting. Windows 7 (And Vista SP2) backs off a bit and does the housekeeping when you're not using the computer. Vista actually wasn't really 'slow', it's just 'irrationally busy' doing stuff with the I/O (indexing, precaching, defragmenting, etc.) while you're just trying to get to your gosh-darned desktop.
So your de-dupe ran for a week before you cut it out? On a modern CPU, the de-dupe is limited not by the CPU speed (since deduplication basically just checksums blocks of storage), but by the speed of the drives.
What you need to do is put all this data onto a single RAID10 array with high IO performance. 5TB of data, plus room to grow on a RAID10 with decent IOPS would probably be something like 6 3TB SATA drives on a new array controller. Set up the array with a large stripe size to prioritize reads (writes are going to be 'fast enough' on a RAID10, trust me). Once you have that hooked-up with your files copied onto it, you want to connect the drive to an OS that can natively deduplicate, like Windows Server 2012. If you must, you can set this box up as a storage server (with a low-end CPU, an old 'Core 2' should be able to keep up with 180MB/sec I/O), and keep your workstation separate. Reading this entire array (when full) through the CPU -should- take about 6-10 hours, deduplication will take slightly longer.
If you don't want to do deduplication at the block level, and you want to actually only have one copy of each duplicated file, you'll need to write scripts that do something like this:
1. Run through the data store and checksum each file (except for those ending in ".mychecksum" with AES128. .."mychecksum" next to it. This will create the 'index' using the filesystem, which will be MUCH faster than having to read the data from inside each file. .deleteme and then deleting all those files after you confirm that it worked.
2. For each file, create an empty file named
3. Search through the store and concatenate all the ".mychecksum" files into a single CSV.
4. Run sed+unique on the file to see what will be nixed (i.e. Get a report)
5. Create another script that actually takes the output from step 4 and deletes ONE of the duplicate files. You can test by -renaming- ONE of the files to
6. Repeat as necessary, possibly with a scheduled job.
I'm in the northeast, a very low-end power user (bottom quartile), and the math still doesn't work out for me on PV. What DOES work is solar thermal to warm up a tank in the basement that sits before the hot water heaters (preheating water) and pumping heat into the living space via baseboard radiators. Unfortunately, those systems are not as cookie-cutter, so getting someone to put them in is almost impossible.
Unfortunately, all the energy saving stuff I see seems geared for newer homes or homes in sunny and hot areas. Where I am, people generally don't even have Air Conditioning, they have 110 year-old homes that aren't well insulated (and often can't be without $15K of asbestos remediation and $5K of rewiring, neither is subsidized). I have yet to meet a contractor who understands that I want windows on the south and that ALLOW lots of infrared in.
I've been the 'young buck' in IT at every job I've had, but I'm starting to get a bit older now. I'm actually the guy who builds images for a university. What you have is important, but an IT department can't just be filled with young folks. I rely on the same kind of people you work with to *actually deploy the work* that I do, and fix the computers when they break, and file tickets when my stuff has bugs in it. It's frustrating having to deal with people who are always working around bugs or limitations that no longer exist, but I can't take on all the responsibility myself.
What I would suggest in your situation is to ask your manager or their manager to give you some time to build newer images, using the tools Microsoft gives people like us to work with modern hardware. Make it a project to modernize imagine deployment. Learn some Powershell so you can write some kick-ass 'firstrun' and management scripts. Investigate how to do things like app virtualization, user data redirection, and remote BIOS management that will make things better for everyone. And most of all, DOCUMENT the processes for the old-timers so they're able to do the things that you figure out.
If you want, you can contact me off-list and I can help you get started with Windows Imaging and some handy scripts we use.
I have my PC (a Lenovo ThinkServer TS130, 16GB RAM, dual HDD, Quadro 600 video card, Sandy Bridge i7) plugged-in via a kill-a-watt. It idles at about 50W. When it's crunching away (compressing video on 8 threads, virtualizing three computers, and playing a movie), it uses about 220w.
The fact is that gamers tend to buy big PSUs because they use them as a pissing-contest. Few but the true monster rigs use more than 300W. Even high-end workstations from the Pentium 4 era came with 375W PSUs, and they had plenty of overhead for fully-decking the system with SLI video cards and RAID arrays.
I was actually shopping for a 160W PSU for a Core 2 workstation that I'm looking to give away recently, I never managed to get the thing to use more than 140W, and it seemed silly to have a 400W PSU sitting there being inefficient all the time. I ended up using a PSU from mini-box that uses an AC->DC brick.
People need to understand and respect just how awesome water is as a coolant. The specific heat of the stuff (basically, how much heat you can 'sink' into a gram of it) and its benign, well-understood nature, and the fact that its density only changes a little bit between freezing and boiling points make it quite awesome.
I live in a city with a river through it. I really don't know why they aren't doing cooling via air-to-water heat pumps. It's really absurd to blow fans all day when the river could carry away 100X the heat without too many ill effects.
Actually, there is. Microsoft is prohibited from doing this on x86-based stuff by the antitrust agreement they're in. They can't prohibit vendors from allowing other operating systems and locking-in 'wintel' hardware/software. The agreement does NOT cover ARM.
> the kernel changes too damned fast and the key would be tied to a particular kernel
That's not true, and it's not how it would work.
Your distribution will be able to package signed bootloaders (GRUB) and kernels. The kernels from your distribution should work fine as long as they Play The Game (which I suspect some, like Debian, will not).
As for the 'kernel changing too damned fast': First, you should know that Windows kernels are ALSO updated between service packs, the Windows kernel components change fairly frequently. Second, the kinds of distributions that get used in managed environments (like RedHat) don't get changed often at all, maybe twice a year. RHEL is still running patched revisions of 2.6.32 from January 2010.
I suspect that in the future, we'll have to 'root' our computers just like we have to root our phones today to load alternate firmware.
> Do you really want to connect all the timing devices in a hospital to an outside public server?
Yes. Because NTP as a client is totally secure. Really. It is.
I'd rather take the one-in-a-billion chance that there's a way to somehow do something bad to a client device over the internet than definitely have my devices run in a messed-up state all the time.
From a risk-management perspective, using NTP to a trusted set of external hosts is a clear win.
Just so you know, Windows Vista and newer have solved this problem. You can push updates AND drivers into offline images. You can put a base image on a server and just load drivers from a database. We cut our number of images from 26 to 1, and it never needs more than three months of Windows Updates.
And it's not just for corporate use, the tools are in the Windows ADK, you just need to script up a recursive driver add after you imagex the machine.
I had written up scripts that automatically took an XP install all the way to where I wanted it to be, and these new tools just blow mine away.
I've had much better luck finding out where they hang out (bars, usually) after a session, then I bring them a small check, made out to their campaign. Once that happens, they usually give you their -real- email address or phone number.
I bought a state rep about $50 in drinks one night, cut him a check the next day, and my ideas on Net Metering made it into the next revision of the bill. I did the same for a city councilman, who is now using a few of my ideas to save money.
The great thing about contribution limits, which are usually under $1,000 per-contributor, is that you get a lot of bang-for-your-buck for a $100 or $200 contribution.
I live in a state that's had 80 years of solid Democrat rule. The district I live in has a student:teacher ratio of 12:1 and we spend over $15K per student annually. Yet despite this, the buildings are literally falling apart, classes are fully-packed, supplies, computers, and books are hard to get.
The problem isn't spending. We spend plenty. The problem is that there's no incentive to produce better outputs. We tried raising salaries, spending more in general. It didn't help.
"I'd prefer it if they left the decision between me and my doctor...."
Just to be coy, that translates to me as:
"I'd prefer it if they left the decision between someone who won't have to pay for something and the guy who wants to get paid...."
Arguably, the insurance company (or the government, in single-payer) is the ONLY party involved with enough information to make the proper decision about what level of care you receive.
I've been saying this all along. The answer for these companies is not to cap or throttle, it's to behave like a good citizen on the internet and either peer with or colocate the data customers want.
Now imagine if Google, Apple, Amazon, and Netflix could host a few boxes inside the Comcast network. Everyone wins. Unfortunately, that's just not how higher-ups in most organizations think.
I have a fair number of pictures, but those are small. This guy has a multi-terabyte thing going; between that and his username, I assume he has a LOT of video. In the case of video, it's time to say 'I won't need all this raw footage in DV format again'. It's OK to compress the video of the birthday party down to 1080p h.264 at medium quality, toss the project files, and let history remember you that way.
I think it's time to admit that you're a hoarder. What exactly -is- your personal data that's so precious? I run a server just to keep my skill set up and run my side business, but I've only managed to accumulate around 600GB of data, only about 35GB of it is 'mine', the rest is client backups.
So first admit that you're a hoarder, then decide if you wan to address that issue or indulge it. If you choose to indulge it, you're going to want to build a small home server. Something with a low-end 64-bit CPU (i3?), a gigabit LAN port, and lots SATA ports and 3.5" drive bays. Buy a bunch of high-quality (WD RE4?) matching drives that fit your data needs times two (you're RAIDing space away). Once you have that, install Linux on it, build a software RAID-1 or 0+1 array (don't do RAID-5 unless you can handle days of rebuild time), and format it with something accessible (read: in the kernel, like EXT4). Create a share on the array with Samba and happily access it from all your machines (don't bother with Netatalk or NFS; CIFS is great on all platforms). As your data needs grow, you can add drives in pairs or replace drives with larger ones and grow the volume. If you need backup, you'll want another array, preferably on another low-end box (an enclosure on your desktop?) but it can be built on a RAID-0 or JBOD to save money.
Mac OS X programs mostly are 'fat binaries'. If you right-click and 'view contents' of an application you'll see that it's really just a directory filled with files. You'll often see that there are x86, x64, and PPC binaries in the same package, sharing the same 'resource' files (which are now actually files instead of HFS magic).
It would be trivial to add more architectures to Mac OS X. Basically all you need is a compiler (which already exists) and for developers to actually target it.
> a national network of bullet trains?
I'm not actually sure that would be of much use. A trip from the northeast to California would be at least 15 hours, and a trip from New York to Florida would be 8 hours. There is no price point that makes a trip that long 'worth it' to do.
Regional high-speed rail would be awesome, though. It would be awesome to be able to easily commute to/from New York City on a daily basis from Boston or D.C.. That would have very real positive economic consequences for the entire eastern seaboard.
Similar story if you could link San Francisco and Los Angeles in under two hours.
You basically have to beat planes on time AND price to make it work. And you obviously have to beat the pants off of cars, because people WILL drive if it's not much faster to take the train.
I have a friend who works at a facility that runs those very exercises. You make it sound like that was the end of the story. In reality, we re-ran the exercises again because you don't just walk away from expensive simulations like that; after a loss, you create a team to re-tool for a few years in the future that prevents that kind of loss again.
Today, that same strategy wouldn't work. We upgraded the anti-boat and anti-missile technologies. We have been running simulations against massive numbers of speedboats for a decade now, successfully.
I'm not saying we wouldn't take losses or that they wouldn't be significant, but I don't think the US Military would stand by and allow us to fall into utter complacency, there are a LOT of checks and balances preventing that. 'Safety' seems to be the one thing our government can do correctly, probably because we've established that it has an unlimited price ceiling.
Hey, give him a break. Most people who's job is to admin Windows systems don't even understand Windows.