The western end is very recognisable, the borders may have moved a little but the countries are in pretty much the same places as they are today.
What has mostly changed afaict is that the east end of germany and the west end of russia have been pushed back and there are now a load of small countries in the gap.
I agree that they probablly won't be illegal to own or to use on your own private land. The real question is where will it be legal to use them. Will they be treated like electric wheelchairs and slow electric scooters used by old folks and allowed nearly everywhere? will thier use be restricted to locations with explicit permission from the landowner? etc
The thing is exoskeletons will allow not just heavier armor but also heavier weapons.
So unless we get some magic armor that can prevent all sizes and speeds of projectile from killing/incapacitating thier targets the infantry with the exoskeletons will have the advantage.
With powered armor like that you'll see heavier weapons too So powered infantry will be at a big advantage both offensively and defensively over conventional infantry and probablly light vehicles too...
Of course we probably won't see 100% powered infantry deployment in any army soon. Probablly, I would guess they would be introduced as a special unit type for jobs like raiding buildings which are dangerous for conventional infantry but not really practical for vehicles.
Nothing a cup of water wont put a stop to. I'm sure the military will have the sense to test that thier suits can work in the presense of water. It's not that big a deal to design stuff to work when wet though it does require some compromises.
Sometimes, a real live human needs to walk out to the machine and toggle the power. Layer upon layer of clever hardware abstraction can help (usually at the cost of performance), but at the end of the day, you still need someone to go flip the switch occasionally. Sure occasionally they do, the trick is to 1: reduce it to a level that it is rarely a problem 2: seperate the tasks of dealing with a hardware issue (which may require physical access even in a virtualised environment) and rebooting/reimaging because you fucked things up (which in a virtualised environment or a physical one with good lights-out management should not)
What happens when the network connected power button fails? You call the datacenter staff and get them to either replace it or move your server to a working one.
Realistically, machines don't just randomly go down hard for no reason. Realistically machines mostly go down because of either flaky hardware, flaky software (both of which can be either lived-with or fixed) or because some admin made a mistake. Occasionally a part fails. An actual fire is pretty unlikely with modern hardware due to the use of flame-retardant materials.
your solution needs to have the ability to deal with everything from pressing the reset button to resolving a worst-case scenario. Ultimately you reach a point where the cost of solving an issue is not worth it given the very low probability and/or the other implications that would likely span from the same base issue (e.g. if there is global thermonuclear war you probablly won't care much about your servers anymore).
Seriously if there is a major fire at the datacenter the only way you are going to get things back up in anything like a timely manner is to do it elsewhere (either using a ready-made backup system or arranging things there and then). Being near the data center won't help you if the fire brigade won't let you in! Have things replicated at sufficiant backup sites far enough apart and you have pretty much done all you can.
The stuff mentioned in the paragraph above is probablly overkill for dev machines though, it all comes back to cost-benifit and risk management.
Assuming the machines are only used for development purposes then one option is to get MSDN subscriptions for all the developers who will use the group of dev servers. MSDN subscriptions are licensed by user not by install.
Sounds like it would cause a lot of lost work as those breakers on the edge of tripping decided they were going to trip (due to some small surge or whatever)
Also in many places computers aren't the only things that need to be plugged in.
Personally I think development servers are fine under the devs desks. Easy for the devs to deal with if they fuck something up. Low density so no real cooling problem and if the power goes out in the office the dev servers wouldn't be much use anyway.
The shuttle was designed as a compromise between a load of different requirements and ultimately ended up with a number of major flaws including
* while it was reusable most of the advantages of reusability were lost because of big refurb requirements every flight. Furthermore the reusability made incremental development harder. So IMO we ended up with the worst of both worlds there. * The shuttle is essentially a mini space-station that goes up and down every time. Great for standalone work in space but very wasteful when working with a proper space-station. * the side-mount "stack" is fundamentally dangerous because it means if something goes wrong with the stack it is far more likely to damage the crew compartment than with a traditional stack. The foam that took down columbia would have been a non-issue with a traditional stack and even an incident like the challenger one would probablly have been more survivable with a traditional stack.
Suppose you enjoy music so you form/join a band. Noone knows who you are so the only venues you can get to play at are pubs and similar. You may put some samples of your music online and offer copies for sale but few people know you exist so there isn't much income from there either.
Then suddenly someone comes along offering you a huge advance and enough promotion to let you play big venues. Sure you may not make much (if any) money off the record sales but the package is still very tempting when the alternative is most likely to stay obscure forever.
No but according to the GGP the missing segment was london to paris.
That is a journey that could be made in a whole range of ways, it could have been flown but not shown up in the record because of the fact it was completely outside the US. It could also have been done by train as the GP mentioned or it could even have been done by road (with a short peice on a train or boat to get accross the channel)
And that is probably the reason for this change. Desipite the inflamatory headline there is no change here. I don't think either MS or sony has ever included a HDMI cable with thier consoles. Afaict at least here in the UK the PS3 has always come with a composite cable (plus a composite to scart adaptor) and the XBOX 360 has always come with a component cable.
At least sony and MS are fitting standard HDMI ports so I don't have to buy an expensive console specific cable for digital (though I do if I want to put my PS3 on component).
The trouble is there are so many things which affect cancer rates that unless something happens which drives them sky high (e.g. chernobyl) any man-made environmental factors are lost in the noise.
Now I don't know what incident you are referring to The one where amazon remotely deleted copies of 1984 due to licensing issues.
and not in anyway a problem with the concept of a virtual library itself. The inherent problem is any digital library without drm will be used as a source to download stuff from and keep it forever (yes some people make copies of physical books but it's enough of a PITA and produces a result of sufficiantly low quality that only the desperate tend to do it). I highly doubt most people would buy a book when they can quickly and easilly download it for free from a legit source.
So the publishers are almost certain to insist on DRM to ensure books are returned and that any loan limitations (whether in the form of copy counts or user limitations) are complied with. Afaict most of the major publishers won't even let thier books be SOLD without drm. It is IMO very likely that the publishers will insist that said DRM includes a revocation feature.
So until/unless the major commercial publishers go away or the governemnt forces the issue (possible but I doubt it at least in the west) any library is going to have to either use drm (with all the badness that implies) to implement the publishers demands or stick to paper for thier collections of major publishers works.
I believe there are a few reasons people buy books rather than borrowing them from the library.
1: conviniance. If I want to borrow a book I have to find a library with it in stock (often much harder than finding a bookstore with it in stock), either go there and borrow it (if i'm allowed to, I may not be depending on what library it is in) or pay for an interlibrary loan request and wait ages (vs with purchasing I can just get it shipped in a couple of days from amazon). Then when I do get the book I have to remember to return/renew it under the threat of (often quite steep) fines and it may be recalled at the least conviniant time. If i'm travelling I probablly won't be anywhere near any library i'm allowed to use and even if I am I still have the problem of returning the books.
2: wanting to keep a book long term. Yes some library systems will let you renew indefinitely but not all and even when yours does you still have to remember to do it, have the threat of recalls, have to remain a member of the library system (e.g. my undergrad card has expired and I don't yet have either my alumni card or my postgrad card) and are limited in the number of books you can have out at once.
3: marketing, bookstores do far far more marketing of thier services
If you had a non-drm digital library reason 2 would dissapear and reason 1 would likely either dissapear or be seriously reduced. I doubt marketing alone would be enough to keep book sales alive.
Among which is that the number of students that can access any specific work is limited only by the numbers of computers and readers availible; and not by the number of copies in stock. The natural limits imposed by real books (only one person can borrow each copy at a time and photocopying too much of it is a PITA) are almost certain to be a replaced by artificial limits imposed by the digitial library system. At least at the uni i'm at most electronic resources are restricted to current students and staff due to licensing issues whereas the physical resources in the library are accessible to anyone who can get a library card.
And what happens when there is something like the kindle incident we saw recently? Books dissapearing in that way would be just as worrying (yes physical books can get stolen but as long as there is a copy in an academic library somewhere in the country researchers who need a copy can get it, when books get so rare that is no longer the case libraries start taking MUCH better care of them) if it was happening to an academic library system than when it is happening to individuals.
My understanding is that the annotations are stored seperately and could indeed be accessed after the book dissapeared. The trouble is without the context provided by the exact version of the book they are meant to go with the annotations lose a lot of thier meaning.
So if amazon has restored the exact version of the book they killed then I don't see the annotations regaining thier context as too serious.
Afaict the biggest issue for webhosts nowadays is dynamic content. It's pretty easy for poorly optimised dynamic content to cause a lot of load and worse (unless the provider has a very fancy setup) that load is focussed on the one machine that hosts your site. That means anyone who shares a machine with you gets a massive performance hit (possiblly to the point of being unusable) when your unoptimised dynamic content site gets a big burst of load.
The good thing about network bandwidth is that there is usually far more bandwidth from a web server to the facility backbone than that server normally needs so if a site causes a burst of bandwidth it's nowhere near as big a deal as if a site starts hogging the CPU or disk.
The big thing that a properly run datacenter gives you is highly reliable power, network links and cooling (granted the last one may not be an issue if it's just the odd box and your local climate isn't too hot).
FTTH will give you more bandwidth (though not nessacerally much more) to your communication providers most local node but that is all it's likely to get you. It's still likely to both be badly congested and not particulally reliable/quickly fixed.
As I see it there are two approaches to designing a big deployment.
One is to try and make all the hardware as reliable as possible by using high quality parts and by demanding of features that allow problems to be detected early and redundancy to be restored quickly with no downtime and minimal risk. The trouble with this approach is that even the best nodes will still die from time to time so you still need redundancy at higher levels for anything important.
The other (the google approach) is to not worry too much about the reliability of individual nodes but instead design your system to tolerate node failure.
Um..and what do you plan on running these disks with?.......The HD's are cheap compared to the other parts that create a storage system. That is the whole point of TFA, they have managed to make a machine to hold/power/access a shitload of drives on the cheap such that the cost of thier system is only 50% more than the raw drives.
And what is more it's mostly off the shelf kit, afaict the only custom bits are the case and the PSU wiring harnesses.
Granted it's not pretty and the higher level system will have to be able to tolerate storage pod downtime (but TBH you have to deal with that anyway, no server is perfectly reliable)
See, your problem is that you run drives on the 12v rail. They run theirs on 5v. Most drives need both, I don't thing theese are any exception (in TFA they claim to use "4 pin molex" connectors to connect the backplanes to the power supply)
Slashdot Mirror
Here is a map of europe in 1900 http://images.google.com/imgres?imgurl=http://www1.aucegypt.edu/faculty/sedgwick/Map/EU1914.jpg&imgrefurl=http://www1.aucegypt.edu/faculty/sedgwick/Map/MapEU1900.html&usg=__I3nXzblc67kO1q7FyiGeG2wkQlY=&h=572&w=900&sz=174&hl=en&start=54&um=1&tbnid=CA8Y22lbGiNirM:&tbnh=93&tbnw=146&prev=/images%3Fq%3Dmap%2Bof%2Beurope%2B1914%26ndsp%3D18%26hl%3Den%26client%3Diceweasel-a%26rls%3Dorg.debian:en-GB:unofficial%26sa%3DN%26start%3D36%26um%3D1
The western end is very recognisable, the borders may have moved a little but the countries are in pretty much the same places as they are today.
What has mostly changed afaict is that the east end of germany and the west end of russia have been pushed back and there are now a load of small countries in the gap.
I agree that they probablly won't be illegal to own or to use on your own private land. The real question is where will it be legal to use them. Will they be treated like electric wheelchairs and slow electric scooters used by old folks and allowed nearly everywhere? will thier use be restricted to locations with explicit permission from the landowner? etc
The thing is exoskeletons will allow not just heavier armor but also heavier weapons.
So unless we get some magic armor that can prevent all sizes and speeds of projectile from killing/incapacitating thier targets the infantry with the exoskeletons will have the advantage.
With powered armor like that you'll see heavier weapons too
So powered infantry will be at a big advantage both offensively and defensively over conventional infantry and probablly light vehicles too...
Of course we probably won't see 100% powered infantry deployment in any army soon.
Probablly, I would guess they would be introduced as a special unit type for jobs like raiding buildings which are dangerous for conventional infantry but not really practical for vehicles.
Nothing a cup of water wont put a stop to.
I'm sure the military will have the sense to test that thier suits can work in the presense of water. It's not that big a deal to design stuff to work when wet though it does require some compromises.
Sometimes, a real live human needs to walk out to the machine and toggle the power. Layer upon layer of clever hardware abstraction can help (usually at the cost of performance), but at the end of the day, you still need someone to go flip the switch occasionally.
Sure occasionally they do, the trick is to
1: reduce it to a level that it is rarely a problem
2: seperate the tasks of dealing with a hardware issue (which may require physical access even in a virtualised environment) and rebooting/reimaging because you fucked things up (which in a virtualised environment or a physical one with good lights-out management should not)
What happens when the network connected power button fails?
You call the datacenter staff and get them to either replace it or move your server to a working one.
Realistically, machines don't just randomly go down hard for no reason.
Realistically machines mostly go down because of either flaky hardware, flaky software (both of which can be either lived-with or fixed) or because some admin made a mistake. Occasionally a part fails. An actual fire is pretty unlikely with modern hardware due to the use of flame-retardant materials.
your solution needs to have the ability to deal with everything from pressing the reset button to resolving a worst-case scenario.
Ultimately you reach a point where the cost of solving an issue is not worth it given the very low probability and/or the other implications that would likely span from the same base issue (e.g. if there is global thermonuclear war you probablly won't care much about your servers anymore).
Seriously if there is a major fire at the datacenter the only way you are going to get things back up in anything like a timely manner is to do it elsewhere (either using a ready-made backup system or arranging things there and then). Being near the data center won't help you if the fire brigade won't let you in! Have things replicated at sufficiant backup sites far enough apart and you have pretty much done all you can.
The stuff mentioned in the paragraph above is probablly overkill for dev machines though, it all comes back to cost-benifit and risk management.
Assuming the machines are only used for development purposes then one option is to get MSDN subscriptions for all the developers who will use the group of dev servers. MSDN subscriptions are licensed by user not by install.
Sounds like it would cause a lot of lost work as those breakers on the edge of tripping decided they were going to trip (due to some small surge or whatever)
Also in many places computers aren't the only things that need to be plugged in.
Personally I think development servers are fine under the devs desks. Easy for the devs to deal with if they fuck something up. Low density so no real cooling problem and if the power goes out in the office the dev servers wouldn't be much use anyway.
The shuttle was designed as a compromise between a load of different requirements and ultimately ended up with a number of major flaws including
* while it was reusable most of the advantages of reusability were lost because of big refurb requirements every flight. Furthermore the reusability made incremental development harder. So IMO we ended up with the worst of both worlds there.
* The shuttle is essentially a mini space-station that goes up and down every time. Great for standalone work in space but very wasteful when working with a proper space-station.
* the side-mount "stack" is fundamentally dangerous because it means if something goes wrong with the stack it is far more likely to damage the crew compartment than with a traditional stack. The foam that took down columbia would have been a non-issue with a traditional stack and even an incident like the challenger one would probablly have been more survivable with a traditional stack.
Suppose you enjoy music so you form/join a band. Noone knows who you are so the only venues you can get to play at are pubs and similar. You may put some samples of your music online and offer copies for sale but few people know you exist so there isn't much income from there either.
Then suddenly someone comes along offering you a huge advance and enough promotion to let you play big venues. Sure you may not make much (if any) money off the record sales but the package is still very tempting when the alternative is most likely to stay obscure forever.
I don't see this changing anytime soon.
No but according to the GGP the missing segment was london to paris.
That is a journey that could be made in a whole range of ways, it could have been flown but not shown up in the record because of the fact it was completely outside the US. It could also have been done by train as the GP mentioned or it could even have been done by road (with a short peice on a train or boat to get accross the channel)
their POS controller server
Can you clarify whether you meant "point of sale", "peice of shit" or both?
Which makes it even more dangerous.
And that is probably the reason for this change.
Desipite the inflamatory headline there is no change here. I don't think either MS or sony has ever included a HDMI cable with thier consoles. Afaict at least here in the UK the PS3 has always come with a composite cable (plus a composite to scart adaptor) and the XBOX 360 has always come with a component cable.
At least sony and MS are fitting standard HDMI ports so I don't have to buy an expensive console specific cable for digital (though I do if I want to put my PS3 on component).
The trouble is there are so many things which affect cancer rates that unless something happens which drives them sky high (e.g. chernobyl) any man-made environmental factors are lost in the noise.
Now I don't know what incident you are referring to
The one where amazon remotely deleted copies of 1984 due to licensing issues.
and not in anyway a problem with the concept of a virtual library itself.
The inherent problem is any digital library without drm will be used as a source to download stuff from and keep it forever (yes some people make copies of physical books but it's enough of a PITA and produces a result of sufficiantly low quality that only the desperate tend to do it). I highly doubt most people would buy a book when they can quickly and easilly download it for free from a legit source.
So the publishers are almost certain to insist on DRM to ensure books are returned and that any loan limitations (whether in the form of copy counts or user limitations) are complied with. Afaict most of the major publishers won't even let thier books be SOLD without drm. It is IMO very likely that the publishers will insist that said DRM includes a revocation feature.
So until/unless the major commercial publishers go away or the governemnt forces the issue (possible but I doubt it at least in the west) any library is going to have to either use drm (with all the badness that implies) to implement the publishers demands or stick to paper for thier collections of major publishers works.
I believe there are a few reasons people buy books rather than borrowing them from the library.
1: conviniance. If I want to borrow a book I have to find a library with it in stock (often much harder than finding a bookstore with it in stock), either go there and borrow it (if i'm allowed to, I may not be depending on what library it is in) or pay for an interlibrary loan request and wait ages (vs with purchasing I can just get it shipped in a couple of days from amazon). Then when I do get the book I have to remember to return/renew it under the threat of (often quite steep) fines and it may be recalled at the least conviniant time. If i'm travelling I probablly won't be anywhere near any library i'm allowed to use and even if I am I still have the problem of returning the books.
2: wanting to keep a book long term. Yes some library systems will let you renew indefinitely but not all and even when yours does you still have to remember to do it, have the threat of recalls, have to remain a member of the library system (e.g. my undergrad card has expired and I don't yet have either my alumni card or my postgrad card) and are limited in the number of books you can have out at once.
3: marketing, bookstores do far far more marketing of thier services
If you had a non-drm digital library reason 2 would dissapear and reason 1 would likely either dissapear or be seriously reduced. I doubt marketing alone would be enough to keep book sales alive.
Among which is that the number of students that can access any specific work is limited only by the numbers of computers and readers availible; and not by the number of copies in stock.
The natural limits imposed by real books (only one person can borrow each copy at a time and photocopying too much of it is a PITA) are almost certain to be a replaced by artificial limits imposed by the digitial library system. At least at the uni i'm at most electronic resources are restricted to current students and staff due to licensing issues whereas the physical resources in the library are accessible to anyone who can get a library card.
And what happens when there is something like the kindle incident we saw recently? Books dissapearing in that way would be just as worrying (yes physical books can get stolen but as long as there is a copy in an academic library somewhere in the country researchers who need a copy can get it, when books get so rare that is no longer the case libraries start taking MUCH better care of them) if it was happening to an academic library system than when it is happening to individuals.
My understanding is that the annotations are stored seperately and could indeed be accessed after the book dissapeared. The trouble is without the context provided by the exact version of the book they are meant to go with the annotations lose a lot of thier meaning.
So if amazon has restored the exact version of the book they killed then I don't see the annotations regaining thier context as too serious.
Afaict the biggest issue for webhosts nowadays is dynamic content. It's pretty easy for poorly optimised dynamic content to cause a lot of load and worse (unless the provider has a very fancy setup) that load is focussed on the one machine that hosts your site. That means anyone who shares a machine with you gets a massive performance hit (possiblly to the point of being unusable) when your unoptimised dynamic content site gets a big burst of load.
The good thing about network bandwidth is that there is usually far more bandwidth from a web server to the facility backbone than that server normally needs so if a site causes a burst of bandwidth it's nowhere near as big a deal as if a site starts hogging the CPU or disk.
The big thing that a properly run datacenter gives you is highly reliable power, network links and cooling (granted the last one may not be an issue if it's just the odd box and your local climate isn't too hot).
FTTH will give you more bandwidth (though not nessacerally much more) to your communication providers most local node but that is all it's likely to get you. It's still likely to both be badly congested and not particulally reliable/quickly fixed.
Sounds like you need to bug virgin until they replace the box.
As I see it there are two approaches to designing a big deployment.
One is to try and make all the hardware as reliable as possible by using high quality parts and by demanding of features that allow problems to be detected early and redundancy to be restored quickly with no downtime and minimal risk. The trouble with this approach is that even the best nodes will still die from time to time so you still need redundancy at higher levels for anything important.
The other (the google approach) is to not worry too much about the reliability of individual nodes but instead design your system to tolerate node failure.
Um..and what do you plan on running these disks with?.......The HD's are cheap compared to the other parts that create a storage system.
That is the whole point of TFA, they have managed to make a machine to hold/power/access a shitload of drives on the cheap such that the cost of thier system is only 50% more than the raw drives.
And what is more it's mostly off the shelf kit, afaict the only custom bits are the case and the PSU wiring harnesses.
Granted it's not pretty and the higher level system will have to be able to tolerate storage pod downtime (but TBH you have to deal with that anyway, no server is perfectly reliable)
See, your problem is that you run drives on the 12v rail. They run theirs on 5v.
Most drives need both, I don't thing theese are any exception (in TFA they claim to use "4 pin molex" connectors to connect the backplanes to the power supply)