It's the nontrivial cases that get you. Where you have to learn how to hammer the package manager into doing what you want. Once you have learnt that for one system it's painful to re-learn it for another.
Last time I used fedora (which admittedly was several releases ago) dependency changes in updates lead to me accidently removing gdm. I installed it again but afterwards it refused to start for no obvious reason. Did I do something wrong? was there a fault in the packages? I don't know but either way it seriously put me off fedora.
Having cards with only an ID number has disadvantages. Either you require all payment accepters to be continuously online* or you lose the ability to tell people their balance in realtime and refuse people who have run out of credit.
The best compromise is probably to store the balance on both the card and in a central database. Then cross-check those values frequently to check for foul play.
PI is irrational, 1/3rd isn't. 1/3 could be represented perfectly if the implementation had a "repeating" bit. AFAIK,
You'd need more than one extra bit to represent reccuring binary fractions because you need to store the point at which the pattern repeats. And you would still only be able to store a subset of rational numbers exactly because you would still have a limited number of bits.
It's true that there is an automatic flow of packages from debian to ubuntu and no automatic flow in the opposite direction but it is also true that a lot of the more radical stuff (multiarch, "modern" init system) happens in ubuntu first because it takes debian so long to argue about the details. It's also true that a fair number of core packages in debian are maintained by canonical employees.
I think ubuntu would certainly be seriously weakened without debian and would likely have to either kick out a large part of universe or accept that it would remain unmaintained but I doubt the dissapearance of debian (not that it's likely to happen) would make them "go away in 6 months".
Nowadays, we have smartphones and tablets that are just as portable as those "pods" I thought of so long ago and much more powerful. Add a bluetooth keyboard and mouse as well as a monitor (via a HDMI Out port) and you could turn your smartphone/tablet into a full fledged keyboard that would rival any desktop computer from 2000.
So why don't people do that?
I think a large part of it is software. Android was simply not designed as a desktop OS. It was designed for devices with small touchscreens and it shows both in the interface of the OS itself (no multiwindow for example) and in the design of applications untended to run on it. MS went to all the effort of porting windows to arm but then refused to let people actually run third party desktop apps on the thing*.
The surface pro line has promise if MS can shed the stigma the surface line has built up and get the weight down.
* Yes I know there is a hack to bypass this but that's not the point.
I suspect that many governments really want to get rid of cash but know they can't get away with it just yet.
So they tighten the screws on large scale cash users by doing things like requiring the banks to report large cash transactions and limiting the size of the largest banknotes printed. Oh and they don't link this to inflation so the "real value" of the largest transaction you can easilly make with cash gradually declines.
You're right that it only takes one compromised Certificate Authority - whether it's compromised by the NSA, hackers, or a corrupt owner - to issue what would be considered legitimate certificates for hundreds of websites like Microsoft or a bank.
Or just by their own sloppy but widely accepted practices.
For example many SSL CAs will issue a cert to anyone who can show they can receive mail for one of a list of "likely administrative" email addresses on the domain. This creates two problems.
1: If you can insert yourself between the server and the internet then you can intercept the mail and reigster yourself a SSL cert. 2: Sometimes the site admins idea of "likely administative" addresses may not line up exactly with the CAs allowing a normal user to register themselves one of those addresses.
The cert itself is not actually the biggest problem with https for the little guy. I'm aware of at least one CA that is recognised by major browsers and hands out basic certs for free and many others that hand them out for far less than the price you give.
The bigger issue is IPs. Right now webhosts are practically limited* to one cert per IP and that means you need a dedicated IP which your hosting provider will almost certainly charge you extra for (partly because it increases their costs slightly, partly because they can)
* Hopefully in a few years time this will no longer be the case as browser/os combinations that don't support SNI die off.
The problem with that argument is that it assumes that passive monitoring of unencrypted data and MITM of encrypted connections have equal cost to the attacker. In practice passive listening has essentially zero risk of getting caught and low CPU cost (just the cost of running whatever filter the attacker wants to try and pull out interesting information). MITM has significant CPU cost and significant chance of getting caught (since the attacker does not know if you are verifying the certificates or not).
The problem is that the web uses a page by page model. Each page request is logically* a seperate "conversation" between client and server. Forms may or may not be submitted to the same hostname they are received from. Links that cross hostnames may contain sensitive information in the url.
By the time the UI elements are displayed it's TOO LATE. The request containing potentially private information has been sent. Therefore the expected security level needs to be encoded as part of the URL so that it is known to the browser when the user follows a link or clicks the submit button on a form.
* Connections may be reused to improve performance but the user has no control or expectations over that.
If the chance of any result being found by chance is 20%, it stands to reason that about 20% of all results were found by chance and, therefore, not to be expected to be reproducible.
Statistical significant levels only tell us about the chance of a study producing a false positive, they say nothing about the chance of a study producing a true positive.
So if the chance of a true positive is low then the false positives could easilly outnumber the true positives.
They should have maintained the same partial pressure of O2
Transitioning from an atmospheric pressure oxygen/nitrogen mix to low pressure pure oxygen is going to be a lot more complex than transition from atmospheric pressure pure oxygen to low pressure pure oxygen. Screw up that transition process and you could end up with the PPO2 dropping to deadly levels.
If I want to give someone (or possiblly copies to a group of people) a piece of media without any expectation of getting it back then optical discs are still an order of magnitude cheaper than USB sticks (about 20p for a DVD-R about $5 for a usb stick . If I buy software then it will most likely come on optical media. If I buy a peice of hardware then the drivers will probablly come on a CD (yes you can often download them too but that is extra hassle, especially if the available internet connection is slow).
It's certainly less of an issue than it used to be with more software available by download and broadband connections delivered over Ethernet/wifi reducing the chicken and egg problem when setting up a new computer but I still wouldn't want my main computer to be without an optical drive even if I only use it occasionally.
According to the article these FTTdp boxes will be "reverse powered" from the customer premisis equipment eliminating the need to provision power supplies for them. Making small boxes of electronics that can live up poles or in underground chambers is hardly a new thing.
Ultimately I guess it comes down to how much do the FTTdp boxes cost. If they are $50K each then there is no way it will be viable but I don't see any reason for them to be that expensive.
It is all in the infrastructure and that is basically fixed whether you use one 1 byte or 10 terabytes.
It depends.
What you have in reality is a situation of stepped costs. Extra usage costs nothing UNTIL it reaches the point that it leaves a link in the provider's network congested. The provider than has a few choices.
1: Try and reroute the traffic, this may help initially but it's likely just going to put the problem off into the future. 2: Try and push the heaviest users to cut back either through throttling measures or by introducing charges for data use. 3: Spend money on upgrading the network either by adding more links or by replacing equipment to improve the speed of.existing ones, both these things cost money. 4: Leave the link congested, build a reputation for low quality service from light and heavy users alike.
The problem is internet service in many places is provided by cable TV providers.
Those providers have a motive to set prices not based on the actual cost of providing the bandwidth but based on making internet TV services (which for most users are by far the biggest bandwidth consumers) less competitive with their own non-internet TC services.
most subscribers will not be within 250 meters of their DSLAM anyway
The whole point of "FTTdp" is to move the DSLAM closer to the end user.
Copper cables run from the telephone exchange to what is known officially a "primary connection point" or more loosely a "cabinet". This is usually a green box by the side of the road. The primary connection point is basically a massive patch panel allowing any line from the exchange to be connected to any customer line. Cables from the PCP then run to distribution points, these may be at the top of poles, on the side of buildings or underground. Finally a thin cable with only a couple of pairs runs to your house.
DSL was traditionally deployed at the telephone exchange, so it had to go through all those cables to get to your house. Recent BT openreach introduced FTTC where a second cabinet is installed alongside the PCP cabinet containing the DSLAM. That brought the DSLAM closer to the customer allowing them to offer higher speeds. They also introduced "Fiber to the premisis on demand" which lets you get higher speeds if FTTC is not enough but installation charges are fairly high.
Now they are trialling (and of course this may or may not pan out) the possibility of having DSLAMs at the distribution point powered by the customer premisis equipment (thereby avoiding the need to get a mains supply to every DP).
and the upload speed will always be just a fraction of 1Gbps.
Of course that is the way BT like it:(. Affordable symmetric "broadband" connections would cannibalise their leased line revenues even more than current broadband services do. Even the FTTP services have relatively lousy upstream bandwidth:(
On the plus side BT openreach's fiber to the cabinet does at least offer better upstream than virgin media cable.
The problem is if you do something once people will assume you will do it again (no matter how much you yell about it being a "one off") when a similar situation arrises in the future. So by "forgiving" current illegal immigrants you encourage more illegal immigration in the future.
Umm.I don't think this is about removing fiber, indeed by the sounds of it it's about installing fiber closer to the customer.
Many telcos are reluctant to do fiber to the premisis because it means sending a fiber tech into the home to liase with the customer about locations and then route and splice the new fiber, whereas apparently with this new tech they can achive gigabit speeds while only having to route the fiber to within 250m of a cluster of premises (e.g. to the top of the pole serving those premisis).
The question I would have is whether there is enough demand in the required speed range to make it worthwhile doing this. Does it make sense to put boxes on the top of poles that only end up serving one house each because everyone else is happy with their existing FTTC (or even regular ADSL) service or does it make more sense just to run fiber directly to those few houses who want something more than FTTC can offer?
I've always found the iptables tutorial from frozentux to be reasonablly comprehensive, maybe it's missing some really fancy stuff but the important stuff about the theory of operation and what the targets do is all there.
Sync has always been originated by the graphics card so no special assistance from the monitor would be needed to lock the framerates and timings of multiple monitors together.
The problem is that traditionally monitors don't just use the sync signals to sync the start of a frame/line, they also use them as part of the process of working out what geometry and timings the graphics card is sending. Furthermore some monitors will only "lock" successfully if the timings are roughly what they expect. So you can't use weird-ass framerates and you can't vary the framerate dynamically without causing glitches.
Displayport unlike other display interfaces is packet based so afaict it shouldn't suffer this limitation but I guess for compatibility reasons they don't want to do anything weird unless they know the monitor at the other end will support it.
I don't see anything about them selling chips to end users, just stuff about them selling upgrade modules. I guess each module will be specific to one make/model of monitor and will require cooperation of the monitor manufacturer to produce.
AIUI first generation thunderbolt is basically equivilent to PCIe 2.0 x4 while second generation thunderbolt is basically equvilent to PCIe 3.0 x4. Afaict that is tolerable but suboptimal for running an external GPU.
The trouble is that there is little to no difference in feel between a plug that is completely the wrong way round and a plug that is merely slightly misaligned or just a bit stiff. So even if you have it the right way round there is a good chance you will think you have it the wrong way round.
Some ports are certainly worse than others. I have an old PCI USB card that usually takes me about 5-6 attempts to plug into blind. I'm not sure if it's the stiffness of the sockets or the fact they protrude slightly from the panel rather than being recessed.
I highly doubt they would be prosecuted just for coming forward. Going after someone for taping a show for their own use would be massively bad PR. If they started selling copies I'm pretty sure the BBCs lawyers would be onto them very quickly.
Not sure what would happen if they put it on bittorrent or similar.
Slashdot Mirror
Sure in the trivial cases it's the same.
It's the nontrivial cases that get you. Where you have to learn how to hammer the package manager into doing what you want. Once you have learnt that for one system it's painful to re-learn it for another.
Last time I used fedora (which admittedly was several releases ago) dependency changes in updates lead to me accidently removing gdm. I installed it again but afterwards it refused to start for no obvious reason. Did I do something wrong? was there a fault in the packages? I don't know but either way it seriously put me off fedora.
Having cards with only an ID number has disadvantages. Either you require all payment accepters to be continuously online* or you lose the ability to tell people their balance in realtime and refuse people who have run out of credit.
The best compromise is probably to store the balance on both the card and in a central database. Then cross-check those values frequently to check for foul play.
* Which is problematic for a transport system.
PI is irrational, 1/3rd isn't. 1/3 could be represented perfectly if the implementation had a "repeating" bit. AFAIK,
You'd need more than one extra bit to represent reccuring binary fractions because you need to store the point at which the pattern repeats. And you would still only be able to store a subset of rational numbers exactly because you would still have a limited number of bits.
It's a little more complex than that.
It's true that there is an automatic flow of packages from debian to ubuntu and no automatic flow in the opposite direction but it is also true that a lot of the more radical stuff (multiarch, "modern" init system) happens in ubuntu first because it takes debian so long to argue about the details. It's also true that a fair number of core packages in debian are maintained by canonical employees.
I think ubuntu would certainly be seriously weakened without debian and would likely have to either kick out a large part of universe or accept that it would remain unmaintained but I doubt the dissapearance of debian (not that it's likely to happen) would make them "go away in 6 months".
Nowadays, we have smartphones and tablets that are just as portable as those "pods" I thought of so long ago and much more powerful. Add a bluetooth keyboard and mouse as well as a monitor (via a HDMI Out port) and you could turn your smartphone/tablet into a full fledged keyboard that would rival any desktop computer from 2000.
So why don't people do that?
I think a large part of it is software. Android was simply not designed as a desktop OS. It was designed for devices with small touchscreens and it shows both in the interface of the OS itself (no multiwindow for example) and in the design of applications untended to run on it. MS went to all the effort of porting windows to arm but then refused to let people actually run third party desktop apps on the thing*.
The surface pro line has promise if MS can shed the stigma the surface line has built up and get the weight down.
* Yes I know there is a hack to bypass this but that's not the point.
I suspect that many governments really want to get rid of cash but know they can't get away with it just yet.
So they tighten the screws on large scale cash users by doing things like requiring the banks to report large cash transactions and limiting the size of the largest banknotes printed. Oh and they don't link this to inflation so the "real value" of the largest transaction you can easilly make with cash gradually declines.
You're right that it only takes one compromised Certificate Authority - whether it's compromised by the NSA, hackers, or a corrupt owner - to issue what would be considered legitimate certificates for hundreds of websites like Microsoft or a bank.
Or just by their own sloppy but widely accepted practices.
For example many SSL CAs will issue a cert to anyone who can show they can receive mail for one of a list of "likely administrative" email addresses on the domain. This creates two problems.
1: If you can insert yourself between the server and the internet then you can intercept the mail and reigster yourself a SSL cert.
2: Sometimes the site admins idea of "likely administative" addresses may not line up exactly with the CAs allowing a normal user to register themselves one of those addresses.
The cert itself is not actually the biggest problem with https for the little guy. I'm aware of at least one CA that is recognised by major browsers and hands out basic certs for free and many others that hand them out for far less than the price you give.
The bigger issue is IPs. Right now webhosts are practically limited* to one cert per IP and that means you need a dedicated IP which your hosting provider will almost certainly charge you extra for (partly because it increases their costs slightly, partly because they can)
* Hopefully in a few years time this will no longer be the case as browser/os combinations that don't support SNI die off.
The problem with that argument is that it assumes that passive monitoring of unencrypted data and MITM of encrypted connections have equal cost to the attacker. In practice passive listening has essentially zero risk of getting caught and low CPU cost (just the cost of running whatever filter the attacker wants to try and pull out interesting information). MITM has significant CPU cost and significant chance of getting caught (since the attacker does not know if you are verifying the certificates or not).
It's not that simple.
The problem is that the web uses a page by page model. Each page request is logically* a seperate "conversation" between client and server. Forms may or may not be submitted to the same hostname they are received from. Links that cross hostnames may contain sensitive information in the url.
By the time the UI elements are displayed it's TOO LATE. The request containing potentially private information has been sent. Therefore the expected security level needs to be encoded as part of the URL so that it is known to the browser when the user follows a link or clicks the submit button on a form.
* Connections may be reused to improve performance but the user has no control or expectations over that.
If the chance of any result being found by chance is 20%, it stands to reason that about 20% of all results were found by chance and, therefore, not to be expected to be reproducible.
Statistical significant levels only tell us about the chance of a study producing a false positive, they say nothing about the chance of a study producing a true positive.
So if the chance of a true positive is low then the false positives could easilly outnumber the true positives.
They should have maintained the same partial pressure of O2
Transitioning from an atmospheric pressure oxygen/nitrogen mix to low pressure pure oxygen is going to be a lot more complex than transition from atmospheric pressure pure oxygen to low pressure pure oxygen. Screw up that transition process and you could end up with the PPO2 dropping to deadly levels.
If I want to give someone (or possiblly copies to a group of people) a piece of media without any expectation of getting it back then optical discs are still an order of magnitude cheaper than USB sticks (about 20p for a DVD-R about $5 for a usb stick . If I buy software then it will most likely come on optical media. If I buy a peice of hardware then the drivers will probablly come on a CD (yes you can often download them too but that is extra hassle, especially if the available internet connection is slow).
It's certainly less of an issue than it used to be with more software available by download and broadband connections delivered over Ethernet/wifi reducing the chicken and egg problem when setting up a new computer but I still wouldn't want my main computer to be without an optical drive even if I only use it occasionally.
According to the article these FTTdp boxes will be "reverse powered" from the customer premisis equipment eliminating the need to provision power supplies for them. Making small boxes of electronics that can live up poles or in underground chambers is hardly a new thing.
Ultimately I guess it comes down to how much do the FTTdp boxes cost. If they are $50K each then there is no way it will be viable but I don't see any reason for them to be that expensive.
It is all in the infrastructure and that is basically fixed whether you use one 1 byte or 10 terabytes.
It depends.
What you have in reality is a situation of stepped costs. Extra usage costs nothing UNTIL it reaches the point that it leaves a link in the provider's network congested. The provider than has a few choices.
1: Try and reroute the traffic, this may help initially but it's likely just going to put the problem off into the future. .existing ones, both these things cost money.
2: Try and push the heaviest users to cut back either through throttling measures or by introducing charges for data use.
3: Spend money on upgrading the network either by adding more links or by replacing equipment to improve the speed of
4: Leave the link congested, build a reputation for low quality service from light and heavy users alike.
The problem is internet service in many places is provided by cable TV providers.
Those providers have a motive to set prices not based on the actual cost of providing the bandwidth but based on making internet TV services (which for most users are by far the biggest bandwidth consumers) less competitive with their own non-internet TC services.
most subscribers will not be within 250 meters of their DSLAM anyway
The whole point of "FTTdp" is to move the DSLAM closer to the end user.
Copper cables run from the telephone exchange to what is known officially a "primary connection point" or more loosely a "cabinet". This is usually a green box by the side of the road. The primary connection point is basically a massive patch panel allowing any line from the exchange to be connected to any customer line. Cables from the PCP then run to distribution points, these may be at the top of poles, on the side of buildings or underground. Finally a thin cable with only a couple of pairs runs to your house.
DSL was traditionally deployed at the telephone exchange, so it had to go through all those cables to get to your house. Recent BT openreach introduced FTTC where a second cabinet is installed alongside the PCP cabinet containing the DSLAM. That brought the DSLAM closer to the customer allowing them to offer higher speeds. They also introduced "Fiber to the premisis on demand" which lets you get higher speeds if FTTC is not enough but installation charges are fairly high.
Now they are trialling (and of course this may or may not pan out) the possibility of having DSLAMs at the distribution point powered by the customer premisis equipment (thereby avoiding the need to get a mains supply to every DP).
and the upload speed will always be just a fraction of 1Gbps.
Of course that is the way BT like it :(. Affordable symmetric "broadband" connections would cannibalise their leased line revenues even more than current broadband services do. Even the FTTP services have relatively lousy upstream bandwidth :(
On the plus side BT openreach's fiber to the cabinet does at least offer better upstream than virgin media cable.
The problem is if you do something once people will assume you will do it again (no matter how much you yell about it being a "one off") when a similar situation arrises in the future. So by "forgiving" current illegal immigrants you encourage more illegal immigration in the future.
Umm.I don't think this is about removing fiber, indeed by the sounds of it it's about installing fiber closer to the customer.
Many telcos are reluctant to do fiber to the premisis because it means sending a fiber tech into the home to liase with the customer about locations and then route and splice the new fiber, whereas apparently with this new tech they can achive gigabit speeds while only having to route the fiber to within 250m of a cluster of premises (e.g. to the top of the pole serving those premisis).
The question I would have is whether there is enough demand in the required speed range to make it worthwhile doing this. Does it make sense to put boxes on the top of poles that only end up serving one house each because everyone else is happy with their existing FTTC (or even regular ADSL) service or does it make more sense just to run fiber directly to those few houses who want something more than FTTC can offer?
I've always found the iptables tutorial from frozentux to be reasonablly comprehensive, maybe it's missing some really fancy stuff but the important stuff about the theory of operation and what the targets do is all there.
i.e. sync originated by the graphics card
Sync has always been originated by the graphics card so no special assistance from the monitor would be needed to lock the framerates and timings of multiple monitors together.
The problem is that traditionally monitors don't just use the sync signals to sync the start of a frame/line, they also use them as part of the process of working out what geometry and timings the graphics card is sending. Furthermore some monitors will only "lock" successfully if the timings are roughly what they expect. So you can't use weird-ass framerates and you can't vary the framerate dynamically without causing glitches.
Displayport unlike other display interfaces is packet based so afaict it shouldn't suffer this limitation but I guess for compatibility reasons they don't want to do anything weird unless they know the monitor at the other end will support it.
I don't see anything about them selling chips to end users, just stuff about them selling upgrade modules. I guess each module will be specific to one make/model of monitor and will require cooperation of the monitor manufacturer to produce.
AIUI first generation thunderbolt is basically equivilent to PCIe 2.0 x4 while second generation thunderbolt is basically equvilent to PCIe 3.0 x4. Afaict that is tolerable but suboptimal for running an external GPU.
mmm.
The trouble is that there is little to no difference in feel between a plug that is completely the wrong way round and a plug that is merely slightly misaligned or just a bit stiff. So even if you have it the right way round there is a good chance you will think you have it the wrong way round.
Some ports are certainly worse than others. I have an old PCI USB card that usually takes me about 5-6 attempts to plug into blind. I'm not sure if it's the stiffness of the sockets or the fact they protrude slightly from the panel rather than being recessed.
I highly doubt they would be prosecuted just for coming forward. Going after someone for taping a show for their own use would be massively bad PR. If they started selling copies I'm pretty sure the BBCs lawyers would be onto them very quickly.
Not sure what would happen if they put it on bittorrent or similar.