Increased with respect to what? Certainly modern design has increased design costs with respect to lifting capacity, production costs, testing costs, safety, and/or some other area -- why else would it still be in use?
Or you could just use out-of-band confirmations for critical data, and retransmit at an appropriate interval when you don't see the confirmation. And even that only applies to unsolicited transmissions -- for interactive traffic the remote host already expects a reply, and can simply retransmit the request if it doesn't get one.
For example, it would be trivial to use this model for loose interactivity without wasting more than a few bytes of bandwidth: 1. Earth host transmits new instructions with sequence number 1 2. Remote host never gets the instructions for whatever reason 3. Earth hosts transmits new instructions with sequence number 2 4. Remote host gets them, notices that the sequence number is 2, and asks for a retransmit of sequence 1 before applying sequence 2
You're right that it's not possible to lose data in the middle of the network with pure TCP. But if the original source gets hit by a meteor in the middle of (re)transmitting, you'd still lose the connection.
Plus the worst-case scenario you describe here is this: 1. I request data from a remote resource 2. The remote resource transmits my data to some intermediate node 3. Data is cached at some intermediate node, and the source is notified of the competed transfer 4. The intermediate node is destroyed 5. I notice that my data never arrived and request it again
That worst-case is essentially indifferent from the worst-case for pure TCP, it's just that the timeout in step 5 has to be much longer. Transmission logs on the source server cannot be considered a reliable indicator of delivery, but that's pretty minor in the grand scheme of things, and really easy to work around.
Generally the proxy on a SOHO router runs as a forward-only cache (or even just a simple proxy) to your ISP's DNS. As such it's really your ISP's DNS that is or isn't vulnerable, because you aren't ever going to see records from anyone else, nor will anyone else know you're asking for them.
The test listed above -- http://entropy.dns-oarc.net/test/ -- will let you know what the rest of the world sees as your DNS source address, and whether or not that source is sufficiently randomized.
Can't I just get the engine that gives me 60-80 MPG, and forget about the batteries, thereby further improving fuel efficiency by reducing vehicle weight? Or is that 60-80 MPG more like ~30 MPG actual efficiency coupled with an assumption that you'll be stopping and idling regularly? Because if it's the later you don't get 60-80 MPG on long trips, you get 30 MPG, which I can already have today.
As a commuter vehicle I think batteries are great. Let's just stop pretending that adding hundreds of pounds of batteries to a car improves the efficiency outside of commuter use.
That said furnaces are more efficient for heat production than electricity ever is.
I don't know about your furnace, but mine is somewhat less than 100% efficient. Electric heat, on the other hand, generally is 100% efficient -- there's no exhaust, and even "waste" energy from fans, thermostats, etc. eventually becomes heat.
Agreed. The article is about charging stations, but the Volt and competitors will charge just fine on 220V in your garage overnight. Given the number of cars GM is talking about - up to hundreds of thousands, no grid upgrades are needed, especially since charging will likely be mostly at night.
Which is great if you live someplace where you've got or could get 220V service near your parking space. But for anyone who lives in a apartment having suitable electric service installed at their parking space seems unlikely.
As for putting the power company out of business, I'm all for it.
There are good reasons for making the transmission infrastructure publicly owned, because that's a natural monopoly -- we want universal service, and we'd like to avoid useless redundancy in infrastructure.
But I don't see why power generation needs to be publicly owned. If the lines are public property, why can't anyone with a generator hook up and sell as much power as they care to generate at whatever rates the market will bear? In addition to keeping prices low and encouraging diversity in generation sources it would let people willing to pay more for their power invest directly in wind or solar generation by selecting a power provider that used only such technologies.
Most of the FAT-32 limitations are not relevant on flash drives. The 4 GB per-file limit could be annoying depending on what you're moving around and how big your media is, but other than that I don't see the problem. There are very few flash drives that exceed the 128 GB volume limit of FAT32. Even for USB drives four 128 GB volumes, while perhaps slightly annoying, would let you access the entire drive. And file permissions and ownership are completely irrelevant on removable media.
It would be nice if NTFS support were better, or if Windows supported at least one fairly modern non-MS file system, but honestly most of the whining about the limitations of FAT32 is misplaced. I think a far superior solution to improving NTFS support would be to make portable network drives the replacement for any removable storage bigger than say 32 GB -- just plug it in to the USB or Ethernet port and mount the CIFS volume from whatever OS you like. Then you'd only be dealing with NTFS on dual-boot systems, where it is entirely reasonable to install drivers in Windows.
Right, no one will get the patch because distros never package patches with their packages. Except for Debian. And RedHat. And CentOS. And Gentoo. And Ubuntu. And Mandrake. And every other distro in existence, including from-source distros like LFS and Zinux (the later of which includes DJBDNS and the IPv6 patch).
Now, distros may or may not decide to package the IPv6 patch, but it's a little silly to pretend that most people run the stock software from the source site.
I understand that stock qmail doesn't compare to modern MTAs like Postfix. But keep in mind that stock qmail was last updated a couple of years before Postfix 1.0 *existed*, and the Internet was a different place back then. And note that there are versions that are much more up-to-date and are perfectly reasonable choices for an MTA, such as qmail-ldap.
That being said, I have switched most of my installations to postfix. I wanted GSSAPI support, and the qmail-ldap author has explicitly refused to support SASL or other such authentication mechanisms.
I'd actually be surprised if your lease didn't say you weren't allowed to post signs without approval, as that's a pretty standard clause is leases. Moreover I suspect that your landlord is allowed to post signs (or at least certain kinds of signs) on the property without your consent.
As for your right to campaign, your landlord can and probably does place reasonable restrictions on that. For example, you wouldn't be allowed to run a campaign headquarters that admitted the general public, employees, or large numbers of volunteers. And you probably can't post signs. But your landlord's rights only extend with respect to the property and its use, and clauses to forbid you from running a calling campaign from your home, or from posting signs on other property would be unenforceable.
The government has only worked to counteract (or enforce, depending on your point of view) discrimination on a very specific set of conditions defined by recent statues, and specifically not the constitution or its amendments. And even in that respect the reach of the government is limited to places that claim to be open to the general public -- requiring registration and refusing to take government money is enough to make you a "private club" and circumvent most government interference.
It really has nothing to do with who made the chips, it's all about quality control, most of which happens somewhere other than where the actual ICs were fabricated. I'm not saying that Apple or their suppliers do anything special in that respect (or that they don't), just that "better" RAM is very rarely manufactured any differently, it's just held to higher standards after the fact.
Apple does overcharge for upgrades, and it's kind of shame, because Dell also overcharges for updates (more on their servers than on their desktops), just not as much. But let's not pretend that adding RAM is some mystical process that requires 3 days fasting and a master's degree.
You are allowed to install RAM without voiding the warranty. There are even illustrated instructions packaged with the machine detailing exactly how to do it, so that people with very limited technical knowledge could reasonably install more RAM without assistance.
On machines where the WiFI card is not standard you'll similarly find illustrated instructions for installing it. Even the hard drive is user replaceable. The towers again include instructions, and while not included Apple publishes "DIY" guides for other case types.
My guess is they match the tracker exchanges to flag your IP, as most trackers are not encrypted even if the transfer traffic itself is. Then they send resets for any connection to/from your IP that they can't identify at "allowed" for the next 15 minutes or so. If I'm right it's not quite as bad as just resetting streams willy-nilly, but it's pretty close, particular if you run non-standard network applications.
If you expanded your campaign to include not must Muslism sympathizers but also Muslism-hating bigots you could take care of the entire problem in one fell swoop.
You can use more than one default route. For reliability you'll want keep some sort of connection-specific check script and reset to a dedicated route if one of the connections goes do. But while both of them are up you can use both connections together so long as you have multiple data streams:
ip route add default table "${MULTI_TABLE}" \
nexthop via "${T1_GW}" dev "${T1_DEV}" weight 1 \
nexthop via "${DSL_GW}" dev "${DSL_DEV}" weight 3
There's a bit more to it than that, but the above example is the heart of a routing policy that splits traffic 3:1 between the DSL and T1. Google should be able to show you the rest.
It's not clear to me how GPL licensing creates free compilers but BSD licensing creates $400 compilers. If some company takes a $0, BSD-licensed compiler, changes two lines of code, and re-sells it under a non-BSD license with no improvements for $400, why would I pay them for it rather than use the $0 original?
If the reseller makes improvements, isn't it reasonable to be able to choose between the lesser, $0 version and the better, $400 version? And what's to stop me from reverse-engineering their improvements, applying those changes to the $0 BSD version, and releasing the updated version under a BSD license?
If you want to force your code to remain open-source, and/or don't want people to be able to integrate your code with non-open-source code, you're absolutely welcome to do so, and the GPL is a great choice. But let's not pretend that having a company re-publish your BSD code under another license somehow removes the utility, availability or openness of the original code.
Or you could buy a printer that supports PostScript. I know it's an evil Adobe abomination, but it's really easy to print to, commonly supported in both network and local drivers, and has a standard printer-definition format to allow selection of hardware-specific options without the need for a hardware-specific driver.
Honestly, in a day and age when even non-tech families have a home network it seems silly to use USB connections and hardware-specific drivers for printers -- just spend the extra $50 and get a printer that can operate with direct interaction from a host CPU.
Hear hear! I think snapshots, mirrors, stripes, encryption, compression and resizing are all very useful things. But I'd like my file system to stick to managing files and use the volume and block layers to provide those features under any file system.
The same people that let remote users enter arbitrary data into an SQL query, or who use non-parameterized queries in the first place. Or who set a "logged_in=1" cookie after authentication and check only that value for future verification.
Slashdot Mirror
Increased with respect to what? Certainly modern design has increased design costs with respect to lifting capacity, production costs, testing costs, safety, and/or some other area -- why else would it still be in use?
Or you could just use out-of-band confirmations for critical data, and retransmit at an appropriate interval when you don't see the confirmation. And even that only applies to unsolicited transmissions -- for interactive traffic the remote host already expects a reply, and can simply retransmit the request if it doesn't get one.
For example, it would be trivial to use this model for loose interactivity without wasting more than a few bytes of bandwidth:
1. Earth host transmits new instructions with sequence number 1
2. Remote host never gets the instructions for whatever reason
3. Earth hosts transmits new instructions with sequence number 2
4. Remote host gets them, notices that the sequence number is 2, and asks for a retransmit of sequence 1 before applying sequence 2
You're right that it's not possible to lose data in the middle of the network with pure TCP. But if the original source gets hit by a meteor in the middle of (re)transmitting, you'd still lose the connection.
Plus the worst-case scenario you describe here is this:
1. I request data from a remote resource
2. The remote resource transmits my data to some intermediate node
3. Data is cached at some intermediate node, and the source is notified of the competed transfer
4. The intermediate node is destroyed
5. I notice that my data never arrived and request it again
That worst-case is essentially indifferent from the worst-case for pure TCP, it's just that the timeout in step 5 has to be much longer. Transmission logs on the source server cannot be considered a reliable indicator of delivery, but that's pretty minor in the grand scheme of things, and really easy to work around.
Generally the proxy on a SOHO router runs as a forward-only cache (or even just a simple proxy) to your ISP's DNS. As such it's really your ISP's DNS that is or isn't vulnerable, because you aren't ever going to see records from anyone else, nor will anyone else know you're asking for them.
The test listed above -- http://entropy.dns-oarc.net/test/ -- will let you know what the rest of the world sees as your DNS source address, and whether or not that source is sufficiently randomized.
Can't I just get the engine that gives me 60-80 MPG, and forget about the batteries, thereby further improving fuel efficiency by reducing vehicle weight? Or is that 60-80 MPG more like ~30 MPG actual efficiency coupled with an assumption that you'll be stopping and idling regularly? Because if it's the later you don't get 60-80 MPG on long trips, you get 30 MPG, which I can already have today.
As a commuter vehicle I think batteries are great. Let's just stop pretending that adding hundreds of pounds of batteries to a car improves the efficiency outside of commuter use.
That said furnaces are more efficient for heat production than electricity ever is.
I don't know about your furnace, but mine is somewhat less than 100% efficient. Electric heat, on the other hand, generally is 100% efficient -- there's no exhaust, and even "waste" energy from fans, thermostats, etc. eventually becomes heat.
Agreed. The article is about charging stations, but the Volt and competitors will charge just fine on 220V in your garage overnight. Given the number of cars GM is talking about - up to hundreds of thousands, no grid upgrades are needed, especially since charging will likely be mostly at night.
Which is great if you live someplace where you've got or could get 220V service near your parking space. But for anyone who lives in a apartment having suitable electric service installed at their parking space seems unlikely.
As for putting the power company out of business, I'm all for it.
There are good reasons for making the transmission infrastructure publicly owned, because that's a natural monopoly -- we want universal service, and we'd like to avoid useless redundancy in infrastructure.
But I don't see why power generation needs to be publicly owned. If the lines are public property, why can't anyone with a generator hook up and sell as much power as they care to generate at whatever rates the market will bear? In addition to keeping prices low and encouraging diversity in generation sources it would let people willing to pay more for their power invest directly in wind or solar generation by selecting a power provider that used only such technologies.
Most of the FAT-32 limitations are not relevant on flash drives. The 4 GB per-file limit could be annoying depending on what you're moving around and how big your media is, but other than that I don't see the problem. There are very few flash drives that exceed the 128 GB volume limit of FAT32. Even for USB drives four 128 GB volumes, while perhaps slightly annoying, would let you access the entire drive. And file permissions and ownership are completely irrelevant on removable media.
It would be nice if NTFS support were better, or if Windows supported at least one fairly modern non-MS file system, but honestly most of the whining about the limitations of FAT32 is misplaced. I think a far superior solution to improving NTFS support would be to make portable network drives the replacement for any removable storage bigger than say 32 GB -- just plug it in to the USB or Ethernet port and mount the CIFS volume from whatever OS you like. Then you'd only be dealing with NTFS on dual-boot systems, where it is entirely reasonable to install drivers in Windows.
DJBDJS is now in the public domain. You can ship it however you want.
You'd rather that it just ignored your request and did something different instead?
Right, no one will get the patch because distros never package patches with their packages. Except for Debian. And RedHat. And CentOS. And Gentoo. And Ubuntu. And Mandrake. And every other distro in existence, including from-source distros like LFS and Zinux (the later of which includes DJBDNS and the IPv6 patch).
Now, distros may or may not decide to package the IPv6 patch, but it's a little silly to pretend that most people run the stock software from the source site.
I understand that stock qmail doesn't compare to modern MTAs like Postfix. But keep in mind that stock qmail was last updated a couple of years before Postfix 1.0 *existed*, and the Internet was a different place back then. And note that there are versions that are much more up-to-date and are perfectly reasonable choices for an MTA, such as qmail-ldap.
That being said, I have switched most of my installations to postfix. I wanted GSSAPI support, and the qmail-ldap author has explicitly refused to support SASL or other such authentication mechanisms.
I'd actually be surprised if your lease didn't say you weren't allowed to post signs without approval, as that's a pretty standard clause is leases. Moreover I suspect that your landlord is allowed to post signs (or at least certain kinds of signs) on the property without your consent.
As for your right to campaign, your landlord can and probably does place reasonable restrictions on that. For example, you wouldn't be allowed to run a campaign headquarters that admitted the general public, employees, or large numbers of volunteers. And you probably can't post signs. But your landlord's rights only extend with respect to the property and its use, and clauses to forbid you from running a calling campaign from your home, or from posting signs on other property would be unenforceable.
The government has only worked to counteract (or enforce, depending on your point of view) discrimination on a very specific set of conditions defined by recent statues, and specifically not the constitution or its amendments. And even in that respect the reach of the government is limited to places that claim to be open to the general public -- requiring registration and refusing to take government money is enough to make you a "private club" and circumvent most government interference.
If programmers wrote programs the way builders built buildings, it would still be cheaper to do accounting on paper.
It really has nothing to do with who made the chips, it's all about quality control, most of which happens somewhere other than where the actual ICs were fabricated. I'm not saying that Apple or their suppliers do anything special in that respect (or that they don't), just that "better" RAM is very rarely manufactured any differently, it's just held to higher standards after the fact.
Apple does overcharge for upgrades, and it's kind of shame, because Dell also overcharges for updates (more on their servers than on their desktops), just not as much. But let's not pretend that adding RAM is some mystical process that requires 3 days fasting and a master's degree.
You are allowed to install RAM without voiding the warranty. There are even illustrated instructions packaged with the machine detailing exactly how to do it, so that people with very limited technical knowledge could reasonably install more RAM without assistance.
On machines where the WiFI card is not standard you'll similarly find illustrated instructions for installing it. Even the hard drive is user replaceable. The towers again include instructions, and while not included Apple publishes "DIY" guides for other case types.
My guess is they match the tracker exchanges to flag your IP, as most trackers are not encrypted even if the transfer traffic itself is. Then they send resets for any connection to/from your IP that they can't identify at "allowed" for the next 15 minutes or so. If I'm right it's not quite as bad as just resetting streams willy-nilly, but it's pretty close, particular if you run non-standard network applications.
If you expanded your campaign to include not must Muslism sympathizers but also Muslism-hating bigots you could take care of the entire problem in one fell swoop.
You can use more than one default route. For reliability you'll want keep some sort of connection-specific check script and reset to a dedicated route if one of the connections goes do. But while both of them are up you can use both connections together so long as you have multiple data streams:
ip route add default table "${MULTI_TABLE}" \
nexthop via "${T1_GW}" dev "${T1_DEV}" weight 1 \
nexthop via "${DSL_GW}" dev "${DSL_DEV}" weight 3
There's a bit more to it than that, but the above example is the heart of a routing policy that splits traffic 3:1 between the DSL and T1. Google should be able to show you the rest.
Sweden apparently has more than one DSLAM provider at the CO too, which is not an option in many places. (Oh how I long for a CLEC)
It's not clear to me how GPL licensing creates free compilers but BSD licensing creates $400 compilers. If some company takes a $0, BSD-licensed compiler, changes two lines of code, and re-sells it under a non-BSD license with no improvements for $400, why would I pay them for it rather than use the $0 original?
If the reseller makes improvements, isn't it reasonable to be able to choose between the lesser, $0 version and the better, $400 version? And what's to stop me from reverse-engineering their improvements, applying those changes to the $0 BSD version, and releasing the updated version under a BSD license?
If you want to force your code to remain open-source, and/or don't want people to be able to integrate your code with non-open-source code, you're absolutely welcome to do so, and the GPL is a great choice. But let's not pretend that having a company re-publish your BSD code under another license somehow removes the utility, availability or openness of the original code.
Or you could buy a printer that supports PostScript. I know it's an evil Adobe abomination, but it's really easy to print to, commonly supported in both network and local drivers, and has a standard printer-definition format to allow selection of hardware-specific options without the need for a hardware-specific driver.
Honestly, in a day and age when even non-tech families have a home network it seems silly to use USB connections and hardware-specific drivers for printers -- just spend the extra $50 and get a printer that can operate with direct interaction from a host CPU.
Hear hear! I think snapshots, mirrors, stripes, encryption, compression and resizing are all very useful things. But I'd like my file system to stick to managing files and use the volume and block layers to provide those features under any file system.
The same people that let remote users enter arbitrary data into an SQL query, or who use non-parameterized queries in the first place. Or who set a "logged_in=1" cookie after authentication and check only that value for future verification.