Apple does test their phones in real-world circumstances! Unfortunately, the new phones tend to get lost, stolen, forgotten in a bar in Redwood City and then put up on Gizmodo, that sort of thing. So, I wouldn't be surprised if Apple has cut down on real-world testing, which would be unfortunate, as extensive real-world testing is one of the hallmarks that Apple is known for: it's part of their "secret sauce" that makes their products easy to use and appealing to the mass market consumer.
The iPhone *did* have a feature to accomplish this, along with many more other useful tools. It replaced the signal bar icon with a numeric dB readout. Google for "iphone field test mode". Unfortunately, Apple revoked this feature as of iOS 4! The old code no longer works, and nobody has yet leaked a new code.
Those corporations, the lucky owners of the grandfathered "Class A" address space, realize their value. They're holding out, as the price can only go up over time, as demand pressure increases. The auctions, in a few years, should provide considerable income. When IANA finally announces they're completely out of IPv4 addresses, you know which stocks to buy!
Better yet, why not just repurpose some of those useless "Class D" and "Class E" allocations?
Add that to the 16 "future use" networks of Class E, to get a total of 28/8 networks that are ripe for the taking. That's more address space than in your entire list above!
There's some stellar examples of AT&T employees (formerly SBC, formerly Pacific Bell) at the dslreports.com site. The "SBC Direct" forum is for bypassing tech support and speaking directly to somebody who knows what they are doing.
Very nice! Both of these have saved me a lot of time, trying to fight my way through a large corporation, trying to reach a person who had both the knowledge and the power to fix the situations I found myself in.
Whenever I have to fly on an airplane, I carry anything valuable as a carryon.
The few times I've ever used a checked bag, it was for clothes and other bulky items. If TSA wants to steal my dirty underwear, they're welcome to it!
Three obvious things:
1) If a rogue TSA employee can take things from your luggage at will, then they can also ADD things to your luggage. Big security risk there.
2) TSA employees are already scanned as they enter. They should also be scanned as they leave! It's common practice in casinos, department stores, and other places that have loose valuables and untrustworthy employees.
3) At a minimum, cameras should be covering every location within the baggage system. This would create a video trail, starting with the timestamp that is created when the luggage is first submitted to the airline. Then, upon report of theft, the video could be watched.
It's a sad day when the TSA is more worrisome than the terrorists.
Wouldn't it be great if airplanes were like trains, in that you personally carry all baggage with you, and that your baggage never leaves your possession?
Measure G passed in 2000. This is a county ordinance that allows for people to grow up to 25 female marijuana plants for personal use (i.e. not for sale) legally.
As for wind turbine crashes, here's some fascinating videos from a storm in Europe during which the speed-limiting system failed on one of the turbines.
Scary! Still, as for having to be near a wind turbine that crashes or a nuclear power plant that melts down, I know which one I'd rather take my chances with....
Won't work anymore. At least in California, one of the subtle things the old AT&T Wireless did before before selling to Comcast was to do away with local franchises.
Now it's statewide franchises.
Complaints now have to be registered at the state level. Good luck trying to effect a change there!
Designing a website so that it can be properly used by the blind also helps design it well for cellphones. Many cellphones have trouble displaying images, or the connection speed is so slow that many people choose to disable images.
Clickable images are often useless on a cellphone, which scales down the image to the point of being unreadable, and also lacks a mouse pointer with which to click on the image.
Flash, and the more complicated parts of JavaScript, are often not supported. AJAX probably won't work.
And finally, many cellphone users are paying by the KB for their downloads! I certainly don't want that charge to be wasted on a useless Flash animation that only serves as a gatekeeper to the real content I'm trying to get at.
Designing a website for the blind isn't profitable. However, designing for cellphones is!
Maybe setting a browser to spoof the User-Agent setting, to appear to be coming from a cellphone, might help?
And, this article coming right after the article showing how the major oilfield in North Dakota might just be viable.
There's hope for this country yet!
I wish there were some underwater photos showing what the subway cars are like after spending several years underwater. (The CGI animation doesn't count.)
1) Easy drop-in of existing TCP, UDP, ICMP, etc. protocols. Just widen the IP address field and leave it at that. Don't try to reinvent 25 years of protocol engineering overnight. There should have been "IPv5", which would have just widened the various IP address fields in all the protocols from 4 to 16 bits, and left it at that. Existing semantics, roles, security models, etc. would be left unchanged.
Even better, this would enable trivial IPv4-to-IPv5 mapping: simply zero-extend all IPv4 address fields and you have a perfect IPv5 packet. It would be an easy 1-to-1 mapping that could be cheaply implemented in hardware. The migration to IPv5 from IPv4 could be nearly seamless. After that, new IPv6 features could be adopted whenever there is demand, so IPv5 would eventually become IPv6.
2) To encourage use, give every existing public IPv4 address its own/64 space of IPv6 address space behind it. For example, the public IPv4 address 129.65.2.119 would get the public IPv6 address block 0.0.0.0.129.65.2.119/64 for "free". Which brings me to my next point.
3) IPv6 should have kept existing syntax. Applications and protocols break, with IPv6's goofy colon-based syntax. The colon has many other longstanding meanings in software, such as port number. An easy and obvious migration path would be to keep the numbers-and-dots notation. Just add more "octets".
SMTP already uses this notation, so it will not be unfamiliar to existing IPv4 sysadmins.
4) Integrate IPv6 addressing with NAT forwarding. Each layer of NAT would correspond to simply filling in 32 more bits of IPv6's 128-bit address. The IPv6 address would simply become a "NAT path", describing how to reach machines behind multiple layers of NAT, just like directories in a filesystem. The/64 subnet of IPv6 would give enough room for 2 layers of IPv4 NAT, at 32 bits each.
Let's say the public IPv4 address 129.65.2.119 is a NAT router, and it has the IPv6 address of 0.0.0.0.129.65.2.119.0.0.0.0.0.0.0.0. Behind the NAT is an interior router at the private address of 192.168.1.100. Filling in 32 more bits gives us the IPv6 address of 0.0.0.0.129.65.2.119.192.168.1.100.0.0.0.0. Let's say that router at 192.168.1.100 is also a NAT. Behind that NAT is a computer with the address 169.254.1.1. Now, we have an IPv6 address of 0.0.0.0.129.65.2.119.192.168.1.100.169.254.1.1. This computer, even though it's behind 2 layers of NAT, can still address any other computer in the world that is also behind 2 layers of NAT. With some creative bit-shifting and address rewriting, an IPv6-aware IPv4 NAT could scale this up to even more layers.
5) Finally, for hardware that just can't be upgraded to process IPv6 packets, have a standard way to encapsulate an IPv6 packet in the payload portion of an IPv4 UDP packet. Perhaps reserve a "special" IPv4 address, such as 255.255.255.254, and use it for the destination, when embedding an IPv6 packet inside of an IPv4 UDP packet in this way.
This gives an upstream router the opportunity to intercept the packet and apply further processing, similar to what is now done for IPv4 SSM multicast. The upstream router could notice the packet being sent to that special nonexistent address, and strip out the IPv6 packet from the payload, and send the IPv6 packet on its way. The process would also work in reverse, so that replies could be received.
I would appreciate responses on why these ideas wouldn't work. Or, would they indeed work?
My favorite feature of Eudora, that I miss most in Thunderbird, is the support for simultaneous multiple tasks.
In the lower left corner of the screen, was a square containing multiple progress bars for each task the program was handling at the moment. A good example would be for checking incoming mail on all of your email accounts.
Eudora would check each of them in parallel. You could monitor the progress of each server's transactions, by watching the progress bars. There was one progress bar for each server. Need to cancel a server that is acting slowly? Right-click on the progress bar and this was done, without disrupting the other servers at all.
In Thunderbird, to contrast, each server is checked one at a time. If a server has a problem, it will block all servers following in sequence. If you cancel, you lose all servers next in sequence as well. The UI of Thunderbird isn't really made to deal with multiple tasks. The best you can see is in the status bar at the very bottom of the window, and this gives no indication of progress, and often goes blank, even when more work is yet to be done.
This was really important back in the days of dialup, when everything was slow, not just email. You would frequently have to wait a long time for email to finish being received. So, having a good UI for this was important. Eudora made the wait a lot less frustrating, by keeping the user fully informed during all operations, and providing plenty of opportunity for the user to cancel a task without disrupting other tasks.
This would be very handy to extend to the newer features expected of modern email clients. Spam Bayesian sorting, filtering of email into appropriate folders, scanning for viruses, and so on, could all be handled as individual tasks using Eudora's interface.
This also becomes important when doing folder operations via IMAP. Thunderbird can get confused when a lengthy folder operation is disrupted, such as moving or copying all messages in a folder. If the user attempts to do something else during this time, the UI in Thunderbird often "forgets" the ongoing task entirely, and the folder will be left in a partial state. The user would then need to clean it up manually later. This isn't a problem with Eudora, thanks to this ability to smoothly handle multiple tasks at once, working each in the background until all tasks are cleanly finished.
I really hope they can bring this feature back, in the new Eudora! I was very disappointed when this didn't make it into Thunderbird 2.0, even though it was a fairly popular suggestion logged in the Bugzilla.
As others have said, 32 bits/sec should still be doable on current hardware.
1) Just oversample, and look for the steep changes in the waveform, to recover the original clock edges of the signal.
It's what analog modems and soundcard modems have been doing for some time. The landline phone system is only 8000 Hz, but sampling is done at the highest multiple of that frequency that the soundcard can do, usually 48000 Hz, for maximum clarity. Software can then analyze the 48000 Hz recorded signal and recover the original 8000 Hz wave. This also helps compensate for clock drift.
2) A classic PC serial port, of which there are many still in use on modern machines, can still be programmed that low.
The original way of programming the PC serial port's baud rate was (115200 divided by X), where X was a 16-bit register, giving a range from 1 to 65536. To achieve 32 baud, simply program this register with a value of 3600. It divides evenly, too. 115200 divided by 3600 is 32. This way of programming the baud rate was also the reason for the classic PC serial port's historic "speed limit" of 115200 baud, since the minimum value for the register was 1.
So, in summary, I don't think slowing down to 32 bits/sec is a problem. As others have said, I think the critical component is the very delicate analog radio components that have been very carefully tuned to isolate and recover this very weak signal from the Voyager probes. As for those, I'd be afraid of them breaking down. But, as for transferring the signal into the computer once the signal has been recovered, no problem.
I believe this issue was settled some time ago, there might have been a lawsuit, but I couldn't find any information.
To summarize, "unlimited" is an old term from the days of dialup modems, and refers to the maximum amount of time you are allowed to stay dialed in and connected: minutes per session, hours per month, and so on. With today's modern broadband connections, kept always-on and connected 24/7, referring to them as "unlimited" is correct. The definition, unfortunately, is old.
However, this says nothing about the bandwidth you are allowed to use. This is today's top issue. We really need another definition to describe this.
With dialup modems, few people really cared about bandwidth consumption, as they were so slow that they didn't make much of an impact, even when continually ran at top speed. With today's fast broadband connections, you can consume a lot of bandwidth in a hurry, and to be affordable at residential prices, they are deliberately oversold.
There's a reason a T1 line still costs $600+/month. You're allowed to run anything and everything over it, no filtering, no capping, and to keep it maxed out at full wire speed, both upload and download, 24/7. Bandwidth to the Internet backbone, unfortunately, is still expensive. I wish it weren't true, but it is. I guess somebody has to pay for all that copper, fiber, and electricity....
I have both Comcast cable and AT&T DSL. I'm really hesitant to use the Comcast cable for much of anything, because of this cap. It is great for games and Web browsing, because it is indeed very fast and responsive. However, for bulk downloads, I would steer clear of it, and BitTorrent is right out.
DSL is slower, but I've never heard of a monthly bandwidth limit. I believe that the slower throughput speed of DSL is self-policing. DSL is also individually wired to each customer, unlike cable, as cable's bandwidth is shared throughout entire neighborhoods. So, the only one you hurt by maxing out the bandwidth of DSL is yourself, and with a packet shaper, this becomes less of a problem.
It varies from area to area, but it appears the "secret" Comcast limit has been determined to be roughly 100 gigabytes per month. I believe this is a cumulative total of both upload and download.
This has been going on for some time, and the good people at broadbandreports.com have much to say about it....
Do this with a USB key and it will have Linux on it, ready to be booted up. Works on any PC, needing no installation, and leaving no traces behind (unlike SanDisk's lousy "U3" software). Because it's Knoppix, it's all self-contained, and can autodetect enough hardware to be useful.
Because Knoppix was intended to be ran from CD, it doesn't write anything back to the USB key. This is a nice fringe benefit, as it will make your USB key last longer (no write cycles).
The only unfortunate thing is that few BIOS's can cleanly boot from USB keys. It's a shame that motherboard makers don't really consider this important, and so don't test for it.
I often think that IPv6 would see greater adoption if it were made easier for end users to get IPv6 addresses assigned to them. I am surprised they just don't map each IPv4 address to and from a corresponding block of IPv6 addresses, to leverage the existing IPv4 address allocation and routing infrastructure already in place.
Also, people might like IPv6 better if its addresses were more readable, using traditional IPv4 notation, but just with more octets. 0.0.0.0.1.2.3.4/64
The idea is that for each publicly routable IPv4 address existing in the world, a corresponding/64 block of IPv6 addresses would also exist, ready to be used, and easily routed to the same destination, by simply following the routing rules already in place for the original IPv4 address.
The upper 32 bits of the IPv6 address would be a well-known constant. I'm just using all zeroes in this example.
The next 32 bits are for the original public IPv4 address.
The rightmost 64 bits are for other devices "behind" that public IPv4 address, such as additional computers behind an IPv4 NAT. Now, IPv6 can address each of them directly, even though IPv4 can't!
You could have up to two layers of NAT be represented in a single IPv6 address, for example:
Public IPv4 address 1.2.3.4 ISP's NAT translates this to 10.0.0.1 End user's NAT translates this to 192.168.1.1
The complete IPv6 address is: 0000:0000:0102:0304:0A00:0001:C0A8:0001 Or, for readability, 0.0.0.0.1.2.3.4.10.0.0.1.192.168.1.1
I've just reached a device behind two layers of NAT, effortlessly, using IPv6! It's a shame this simple idea isn't more widely deployed.
Is it just me, or are humans getting better and better at science as time progresses?...
Everything seems to be a big puzzle, and we seem to be getting faster and more accurate with putting these puzzles together....
Has science always been this inexorable in it's progress? Yes. The rate of change is increasing. Each new invention makes it slightly easier to invent the next invention, and so on. It's exponential, one of those scary J-curves. Nobody knows where it will all end up spiking upward to, but many people have thought of this before.
As somebody else pointed out, the phrase you're looking for is "technological singularity".
Google this for many fun hours reading. Read some papers by Ray Kurzweil. Buy a book or two by Vernor Vinge.
That's one of the best things about United Airlines. Their in-flight audio system usually has the cockpit audio chatter available as a channel you can listen to (usually channel 9). There is a kill switch for this audio in the cockpit, of course. Unfortunately sometimes pilots hit the switch and then forget to take it back off, so this channel ends up going silent for the remainder of the flight.
I keep hoping that one of the domains will have tiered pricing based on desirability of the name.
By this, I mean charging a premium price for:
Short domain names Dictionary words Other desirable factors (no numbers, no punctuation, etc.?)
I often wonder why the powers that be do not put a premium price on the "good" domains. You don't need to learn combinatorial theory in order to see that the very short domain names (3 letters, etc.) are few and far between, and in high demand.
For instance:
1-letter names = Not allowed, by stupid administrative policy, although a few are grandfathered (q.com) 2-letter names = Not allowed, by stupid administrative policy, although a few are grandfathered (xe.com) 3-letter names = $10,000/year 4-letter names = $500/year 5-letter names = $25/year Longer names that are dictionary words = $15/year Longer names that are not dictionary words = $5/year
It's a growing problem with many Linux software packages these days: they depend on having a valid hostname in order to function.
The GNOME desktop is especially notorious for this.
I'm sad to see the Ubuntu installer also having this flaw.
Many people are in situations like yours, where they don't have access to valid hostnames, but a perfectly good IP address to use.
It's gotten so bad in some situations that I've taken to setting up a "hostname wall", similar to a DNS wall: basically a huge table of dummy hostnames like ip-10-1-1-1 and ip-192-168-1-1, resolving to exactly what you'd expect. I'm seriously thinking of writing my own glibc libnss plugin to do this automatically. Has anybody else done this already? It would save me the trouble:)
I also have a problem with apps stealing focus... not in Windows as you would expect, but in GNOME (distro is RHEL4).
Gaim steals focus.
KDE has the handy "Focus stealing prevention level" slider, that you can crank up to make popups less invasive. Unfortunately, GNOME does not have this. And, for various reasons, I need to use GNOME, not KDE.
I've played around with various settings, and the best I can find is to turn off alerts entirely in Gaim. This isn't the best solution, because it prevents me from easily seeing new messages, and it applies only to Gaim. Other apps also steal focus, and I'd love a more general solution.
Do newer versions of GNOME have something like KDE's "Focus stealing prevention level"?
Slashdot Mirror
Apple does test their phones in real-world circumstances! Unfortunately, the new phones tend to get lost, stolen, forgotten in a bar in Redwood City and then put up on Gizmodo, that sort of thing. So, I wouldn't be surprised if Apple has cut down on real-world testing, which would be unfortunate, as extensive real-world testing is one of the hallmarks that Apple is known for: it's part of their "secret sauce" that makes their products easy to use and appealing to the mass market consumer.
The iPhone *did* have a feature to accomplish this, along with many more other useful tools. It replaced the signal bar icon with a numeric dB readout. Google for "iphone field test mode". Unfortunately, Apple revoked this feature as of iOS 4! The old code no longer works, and nobody has yet leaked a new code.
Those corporations, the lucky owners of the grandfathered "Class A" address space, realize their value. They're holding out, as the price can only go up over time, as demand pressure increases. The auctions, in a few years, should provide considerable income. When IANA finally announces they're completely out of IPv4 addresses, you know which stocks to buy!
Better yet, why not just repurpose some of those useless "Class D" and "Class E" allocations?
Multicast is from 224 to 239.
According to http://www.iana.org/assignments/multicast-addresses/ only these have been allocated: 224, 232, 233, 239
This leaves 12 unused /8 networks: 225, 226, 227, 228, 229, 230, 231, 234, 235, 236, 237, 238
Add that to the 16 "future use" networks of Class E, to get a total of 28 /8 networks that are ripe for the taking. That's more address space than in your entire list above!
There's some stellar examples of AT&T employees (formerly SBC, formerly Pacific Bell) at the dslreports.com site. The "SBC Direct" forum is for bypassing tech support and speaking directly to somebody who knows what they are doing.
http://www.dslreports.com/forum/sbcdirect
Also, for Comcast, there's "comcastcares" on Twitter. There was a recent article written about this person recently.
http://twitter.com/comcastcares
http://www.businessweek.com/managing/content/jan2009/ca20090113_373506.htm
Very nice! Both of these have saved me a lot of time, trying to fight my way through a large corporation, trying to reach a person who had both the knowledge and the power to fix the situations I found myself in.
Whenever I have to fly on an airplane, I carry anything valuable as a carryon.
The few times I've ever used a checked bag, it was for clothes and other bulky items. If TSA wants to steal my dirty underwear, they're welcome to it!
Three obvious things:
1) If a rogue TSA employee can take things from your luggage at will, then they can also ADD things to your luggage. Big security risk there.
2) TSA employees are already scanned as they enter. They should also be scanned as they leave! It's common practice in casinos, department stores, and other places that have loose valuables and untrustworthy employees.
3) At a minimum, cameras should be covering every location within the baggage system. This would create a video trail, starting with the timestamp that is created when the luggage is first submitted to the airline. Then, upon report of theft, the video could be watched.
It's a sad day when the TSA is more worrisome than the terrorists.
Wouldn't it be great if airplanes were like trains, in that you personally carry all baggage with you, and that your baggage never leaves your possession?
Read the text of the law again. The sentence you quoted only applies to marijuana "for sale".
To clarify, Measure G does allow personal growing and use, but not sale, of up to 25 plants.
I grew up in Mendocino County.
Measure G passed in 2000. This is a county ordinance that allows for people to grow up to 25 female marijuana plants for personal use (i.e. not for sale) legally.
http://www.canorml.org/news/mendorelse.html
http://stopthedrugwar.org/in_the_trenches/2007/apr/25/amma_press_release_victory_mendo
I'm not sure, but this ordinance seems compatible with the statewide Proposition 215. It is still illegal under federal law, of course.
As for wind turbine crashes, here's some fascinating videos from a storm in Europe during which the speed-limiting system failed on one of the turbines.
http://www.youtube.com/watch?v=u14tBwO5QVQ
http://www.youtube.com/watch?v=7nSB1SdVHqQ
Scary! Still, as for having to be near a wind turbine that crashes or a nuclear power plant that melts down, I know which one I'd rather take my chances with....
Won't work anymore. At least in California, one of the subtle things the old AT&T Wireless did before before selling to Comcast was to do away with local franchises.
Now it's statewide franchises.
Complaints now have to be registered at the state level. Good luck trying to effect a change there!
I wonder why the Chinese just don't take over the Russian GLONASS system.
Does the world really need *four* parallel GPS systems? Three should be more than enough.
The Russians and Chinese are close enough, that you'd think the Russians would welcome the help to complete their GLONASS system.
Designing a website so that it can be properly used by the blind also helps design it well for cellphones. Many cellphones have trouble displaying images, or the connection speed is so slow that many people choose to disable images.
Clickable images are often useless on a cellphone, which scales down the image to the point of being unreadable, and also lacks a mouse pointer with which to click on the image.
Flash, and the more complicated parts of JavaScript, are often not supported. AJAX probably won't work.
And finally, many cellphone users are paying by the KB for their downloads! I certainly don't want that charge to be wasted on a useless Flash animation that only serves as a gatekeeper to the real content I'm trying to get at.
Designing a website for the blind isn't profitable. However, designing for cellphones is!
Maybe setting a browser to spoof the User-Agent setting, to appear to be coming from a cellphone, might help?
What a great idea!
And, this article coming right after the article showing how the major oilfield in North Dakota might just be viable.
There's hope for this country yet!
I wish there were some underwater photos showing what the subway cars are like after spending several years underwater. (The CGI animation doesn't count.)
IPv6 should have had this:
/64 space of IPv6 address space behind it. For example, the public IPv4 address 129.65.2.119 would get the public IPv6 address block 0.0.0.0.129.65.2.119/64 for "free". Which brings me to my next point.
/64 subnet of IPv6 would give enough room for 2 layers of IPv4 NAT, at 32 bits each.
1) Easy drop-in of existing TCP, UDP, ICMP, etc. protocols. Just widen the IP address field and leave it at that. Don't try to reinvent 25 years of protocol engineering overnight. There should have been "IPv5", which would have just widened the various IP address fields in all the protocols from 4 to 16 bits, and left it at that. Existing semantics, roles, security models, etc. would be left unchanged.
Even better, this would enable trivial IPv4-to-IPv5 mapping: simply zero-extend all IPv4 address fields and you have a perfect IPv5 packet. It would be an easy 1-to-1 mapping that could be cheaply implemented in hardware. The migration to IPv5 from IPv4 could be nearly seamless. After that, new IPv6 features could be adopted whenever there is demand, so IPv5 would eventually become IPv6.
2) To encourage use, give every existing public IPv4 address its own
3) IPv6 should have kept existing syntax. Applications and protocols break, with IPv6's goofy colon-based syntax. The colon has many other longstanding meanings in software, such as port number. An easy and obvious migration path would be to keep the numbers-and-dots notation. Just add more "octets".
SMTP already uses this notation, so it will not be unfamiliar to existing IPv4 sysadmins.
4) Integrate IPv6 addressing with NAT forwarding. Each layer of NAT would correspond to simply filling in 32 more bits of IPv6's 128-bit address. The IPv6 address would simply become a "NAT path", describing how to reach machines behind multiple layers of NAT, just like directories in a filesystem. The
Let's say the public IPv4 address 129.65.2.119 is a NAT router, and it has the IPv6 address of 0.0.0.0.129.65.2.119.0.0.0.0.0.0.0.0. Behind the NAT is an interior router at the private address of 192.168.1.100. Filling in 32 more bits gives us the IPv6 address of 0.0.0.0.129.65.2.119.192.168.1.100.0.0.0.0. Let's say that router at 192.168.1.100 is also a NAT. Behind that NAT is a computer with the address 169.254.1.1. Now, we have an IPv6 address of 0.0.0.0.129.65.2.119.192.168.1.100.169.254.1.1. This computer, even though it's behind 2 layers of NAT, can still address any other computer in the world that is also behind 2 layers of NAT. With some creative bit-shifting and address rewriting, an IPv6-aware IPv4 NAT could scale this up to even more layers.
5) Finally, for hardware that just can't be upgraded to process IPv6 packets, have a standard way to encapsulate an IPv6 packet in the payload portion of an IPv4 UDP packet. Perhaps reserve a "special" IPv4 address, such as 255.255.255.254, and use it for the destination, when embedding an IPv6 packet inside of an IPv4 UDP packet in this way.
This gives an upstream router the opportunity to intercept the packet and apply further processing, similar to what is now done for IPv4 SSM multicast. The upstream router could notice the packet being sent to that special nonexistent address, and strip out the IPv6 packet from the payload, and send the IPv6 packet on its way. The process would also work in reverse, so that replies could be received.
I would appreciate responses on why these ideas wouldn't work. Or, would they indeed work?
I'm glad to see Eudora coming back.
My favorite feature of Eudora, that I miss most in Thunderbird, is the support for simultaneous multiple tasks.
In the lower left corner of the screen, was a square containing multiple progress bars for each task the program was handling at the moment. A good example would be for checking incoming mail on all of your email accounts.
Eudora would check each of them in parallel. You could monitor the progress of each server's transactions, by watching the progress bars. There was one progress bar for each server. Need to cancel a server that is acting slowly? Right-click on the progress bar and this was done, without disrupting the other servers at all.
In Thunderbird, to contrast, each server is checked one at a time. If a server has a problem, it will block all servers following in sequence. If you cancel, you lose all servers next in sequence as well. The UI of Thunderbird isn't really made to deal with multiple tasks. The best you can see is in the status bar at the very bottom of the window, and this gives no indication of progress, and often goes blank, even when more work is yet to be done.
This was really important back in the days of dialup, when everything was slow, not just email. You would frequently have to wait a long time for email to finish being received. So, having a good UI for this was important. Eudora made the wait a lot less frustrating, by keeping the user fully informed during all operations, and providing plenty of opportunity for the user to cancel a task without disrupting other tasks.
This would be very handy to extend to the newer features expected of modern email clients. Spam Bayesian sorting, filtering of email into appropriate folders, scanning for viruses, and so on, could all be handled as individual tasks using Eudora's interface.
This also becomes important when doing folder operations via IMAP. Thunderbird can get confused when a lengthy folder operation is disrupted, such as moving or copying all messages in a folder. If the user attempts to do something else during this time, the UI in Thunderbird often "forgets" the ongoing task entirely, and the folder will be left in a partial state. The user would then need to clean it up manually later. This isn't a problem with Eudora, thanks to this ability to smoothly handle multiple tasks at once, working each in the background until all tasks are cleanly finished.
I really hope they can bring this feature back, in the new Eudora! I was very disappointed when this didn't make it into Thunderbird 2.0, even though it was a fairly popular suggestion logged in the Bugzilla.
As others have said, 32 bits/sec should still be doable on current hardware.
1) Just oversample, and look for the steep changes in the waveform, to recover the original clock edges of the signal.
It's what analog modems and soundcard modems have been doing for some time. The landline phone system is only 8000 Hz, but sampling is done at the highest multiple of that frequency that the soundcard can do, usually 48000 Hz, for maximum clarity. Software can then analyze the 48000 Hz recorded signal and recover the original 8000 Hz wave. This also helps compensate for clock drift.
2) A classic PC serial port, of which there are many still in use on modern machines, can still be programmed that low.
The original way of programming the PC serial port's baud rate was (115200 divided by X), where X was a 16-bit register, giving a range from 1 to 65536. To achieve 32 baud, simply program this register with a value of 3600. It divides evenly, too. 115200 divided by 3600 is 32. This way of programming the baud rate was also the reason for the classic PC serial port's historic "speed limit" of 115200 baud, since the minimum value for the register was 1.
So, in summary, I don't think slowing down to 32 bits/sec is a problem. As others have said, I think the critical component is the very delicate analog radio components that have been very carefully tuned to isolate and recover this very weak signal from the Voyager probes. As for those, I'd be afraid of them breaking down. But, as for transferring the signal into the computer once the signal has been recovered, no problem.
You're probably right, the cap is most likely 200GB/month total traffic. So, 100GB upload and 100GB download would total enough to reach the cap.
I guessed at my numbers the same way you did, by doing various searches. Saw what numbers came up when people discussed getting "The Letter".
Until Comcast comes clean and discloses the cap criteria, nobody really knows for sure. It's all hearsay.
I believe this issue was settled some time ago, there might have been a lawsuit, but I couldn't find any information.
To summarize, "unlimited" is an old term from the days of dialup modems, and refers to the maximum amount of time you are allowed to stay dialed in and connected: minutes per session, hours per month, and so on. With today's modern broadband connections, kept always-on and connected 24/7, referring to them as "unlimited" is correct. The definition, unfortunately, is old.
However, this says nothing about the bandwidth you are allowed to use. This is today's top issue. We really need another definition to describe this.
With dialup modems, few people really cared about bandwidth consumption, as they were so slow that they didn't make much of an impact, even when continually ran at top speed. With today's fast broadband connections, you can consume a lot of bandwidth in a hurry, and to be affordable at residential prices, they are deliberately oversold.
There's a reason a T1 line still costs $600+/month. You're allowed to run anything and everything over it, no filtering, no capping, and to keep it maxed out at full wire speed, both upload and download, 24/7. Bandwidth to the Internet backbone, unfortunately, is still expensive. I wish it weren't true, but it is. I guess somebody has to pay for all that copper, fiber, and electricity....
I have both Comcast cable and AT&T DSL. I'm really hesitant to use the Comcast cable for much of anything, because of this cap. It is great for games and Web browsing, because it is indeed very fast and responsive. However, for bulk downloads, I would steer clear of it, and BitTorrent is right out.
DSL is slower, but I've never heard of a monthly bandwidth limit. I believe that the slower throughput speed of DSL is self-policing. DSL is also individually wired to each customer, unlike cable, as cable's bandwidth is shared throughout entire neighborhoods. So, the only one you hurt by maxing out the bandwidth of DSL is yourself, and with a packet shaper, this becomes less of a problem.
It varies from area to area, but it appears the "secret" Comcast limit has been determined to be roughly 100 gigabytes per month. I believe this is a cumulative total of both upload and download.
This has been going on for some time, and the good people at broadbandreports.com have much to say about it....
You can already boot Linux up over USB just fine, using Knoppix.
Here's a guide I wrote:
http://www.knoppix.net/wiki/Bootable_USB_Key
Do this with a USB key and it will have Linux on it, ready to be booted up. Works on any PC, needing no installation, and leaving no traces behind (unlike SanDisk's lousy "U3" software). Because it's Knoppix, it's all self-contained, and can autodetect enough hardware to be useful.
Because Knoppix was intended to be ran from CD, it doesn't write anything back to the USB key. This is a nice fringe benefit, as it will make your USB key last longer (no write cycles).
The only unfortunate thing is that few BIOS's can cleanly boot from USB keys. It's a shame that motherboard makers don't really consider this important, and so don't test for it.
I often think that IPv6 would see greater adoption if it were made easier for end users to get IPv6 addresses assigned to them. I am surprised they just don't map each IPv4 address to and from a corresponding block of IPv6 addresses, to leverage the existing IPv4 address allocation and routing infrastructure already in place.
/64 block of IPv6 addresses would also exist, ready to be used, and easily routed to the same destination, by simply following the routing rules already in place for the original IPv4 address.
For example:
IPv4 address 1.2.3.4
IPv6 address block 0000:0000:0102:0304/64
Also, people might like IPv6 better if its addresses were more readable, using traditional IPv4 notation, but just with more octets.
0.0.0.0.1.2.3.4/64
The idea is that for each publicly routable IPv4 address existing in the world, a corresponding
The upper 32 bits of the IPv6 address would be a well-known constant. I'm just using all zeroes in this example.
The next 32 bits are for the original public IPv4 address.
The rightmost 64 bits are for other devices "behind" that public IPv4 address, such as additional computers behind an IPv4 NAT. Now, IPv6 can address each of them directly, even though IPv4 can't!
You could have up to two layers of NAT be represented in a single IPv6 address, for example:
Public IPv4 address 1.2.3.4
ISP's NAT translates this to 10.0.0.1
End user's NAT translates this to 192.168.1.1
The complete IPv6 address is: 0000:0000:0102:0304:0A00:0001:C0A8:0001
Or, for readability, 0.0.0.0.1.2.3.4.10.0.0.1.192.168.1.1
I've just reached a device behind two layers of NAT, effortlessly, using IPv6! It's a shame this simple idea isn't more widely deployed.
Everything seems to be a big puzzle, and we seem to be getting faster and more accurate with putting these puzzles together.
Has science always been this inexorable in it's progress? Yes. The rate of change is increasing. Each new invention makes it slightly easier to invent the next invention, and so on. It's exponential, one of those scary J-curves. Nobody knows where it will all end up spiking upward to, but many people have thought of this before.
As somebody else pointed out, the phrase you're looking for is "technological singularity".
Google this for many fun hours reading. Read some papers by Ray Kurzweil. Buy a book or two by Vernor Vinge.
http://en.wikipedia.org/wiki/Technological_singul
http://www.google.com/search?q=technological+sing
That's one of the best things about United Airlines. Their in-flight audio system usually has the cockpit audio chatter available as a channel you can listen to (usually channel 9). There is a kill switch for this audio in the cockpit, of course. Unfortunately sometimes pilots hit the switch and then forget to take it back off, so this channel ends up going silent for the remainder of the flight.
I keep hoping that one of the domains will have tiered pricing based on desirability of the name.
By this, I mean charging a premium price for:
Short domain names
Dictionary words
Other desirable factors (no numbers, no punctuation, etc.?)
I often wonder why the powers that be do not put a premium price on the "good" domains. You don't need to learn combinatorial theory in order to see that the very short domain names (3 letters, etc.) are few and far between, and in high demand.
For instance:
1-letter names = Not allowed, by stupid administrative policy, although a few are grandfathered (q.com)
2-letter names = Not allowed, by stupid administrative policy, although a few are grandfathered (xe.com)
3-letter names = $10,000/year
4-letter names = $500/year
5-letter names = $25/year
Longer names that are dictionary words = $15/year
Longer names that are not dictionary words = $5/year
It's a growing problem with many Linux software packages these days: they depend on having a valid hostname in order to function.
:)
The GNOME desktop is especially notorious for this.
I'm sad to see the Ubuntu installer also having this flaw.
Many people are in situations like yours, where they don't have access to valid hostnames, but a perfectly good IP address to use.
It's gotten so bad in some situations that I've taken to setting up a "hostname wall", similar to a DNS wall: basically a huge table of dummy hostnames like ip-10-1-1-1 and ip-192-168-1-1, resolving to exactly what you'd expect. I'm seriously thinking of writing my own glibc libnss plugin to do this automatically. Has anybody else done this already? It would save me the trouble
I also have a problem with apps stealing focus... not in Windows as you would expect, but in GNOME (distro is RHEL4).
Gaim steals focus.
KDE has the handy "Focus stealing prevention level" slider, that you can crank up to make popups less invasive. Unfortunately, GNOME does not have this. And, for various reasons, I need to use GNOME, not KDE.
I've played around with various settings, and the best I can find is to turn off alerts entirely in Gaim. This isn't the best solution, because it prevents me from easily seeing new messages, and it applies only to Gaim. Other apps also steal focus, and I'd love a more general solution.
Do newer versions of GNOME have something like KDE's "Focus stealing prevention level"?