You can't realistically expect users to know how to configure this stuff and it doesn't actually cost the company anything extra.
Actually, I can expect that. And I can even show you a pretty graph that indicates folks are doing an increasingly better job with encrypting their wireless networks.
As an anecdote, my own experiences with wardriving in small-town Ohio have been interesting to me. Some towns and neighborhoods are full of wide-open networks. Some are almost completely locked-down. Some people will have two SSIDs for their house, like a WPA-protected network called "Jones" and a second non-encrypted "Jones Guest".
And there's plenty of savvy people out there who even give different family members their own encrypted WLANs, judging from the SSIDs that I see.
Generally speaking, I've seen folks make good progress over the past few years. Gone are the days when I could just open my laptop in any old neighborhood, pick one of several "linksys" APs, and get Internet access.
It's only a matter of time before we can engineer plants or processes which mimic the photosynthetic process or tweak it to minimize the amount of afterprocessing for biofuels.
What makes you think the company licensing the patent is going to allow [alternative] to be produced for less than the price of oil?
Patents. You said so yourself.
Their inclination would be to keep the price of plant-based fuels to be above oil, if and as long as it is profitable to do so. Once the patents expire, they'll have to compete with other things (like oil, coal, and other companies using the exact same plant-based processes which used to have patent protection) on a more realistic playing field.
That might be a solution with 12,500 workstations, but at my (much, much smaller) company it ain't gonna work:
According to VMware, it costs $5,000 to get started with 50 client licenses. And additional licenses are $39.
For those prices, it's more profitable to stick with XP for as long as possible, and then run things in virtualized XP Mode once the hardware catches up with the overhead (in the natural course of attrition and replacement).
Easy: HFCS would go up in price. Goods made from HFCS would also go up in price.
This trend would continue until it would became cheaper to use sugar instead, and then the price of sugar would begin to edge up and balance things out.
The bigger question is this: If junk food were more expensive (as described above), would people buy less of it?
likely from a 220v source like your dryer and stove use
Not to nitpick but if you are referring to the US and/or Canada it should be 240V, not 220V. The standard residential connection in the US is a split-phase connection where each leg aims for 120V. Cross the two and you get 240V.
Not to nitpick but if you are referring to (pick a country), then anything between about 220V and 240V is "220." Or "240," depending on local parlance.
The two are synonymous. Much as "110," "115," and "120" are.
Power, once distributed, isn't very exact -- no matter what the nationality. Here in the States, I've seen unladen "120V" circuits ranging between 112V to 125V. And if we're doing the center-tapped-transformer thing, then this equates to "224 to 250V".
It's quite normal for this sort of variation to occur in the real world, and is going to remain that way until we all see auto-switching multitap transformers by the street, or personal motor-generators, or something similar (none of which is likely to ever happen).
likely from a 220v source like your dryer and stove use
Not to nitpick but if you are referring to the US and/or Canada it should be 240V, not 220V. The standard residential connection in the US is a split-phase connection where each leg aims for 120V. Cross the two and you get 240V.
Not to nitpick but if you are referring to (pick a country), then anything between about 220V and 240V is "220." Or "240," depending on local parlance.
The two are synonymous. Much as "110," "115," and "120" are.
Power, once distributed, isn't very exact -- no matter what the nationality. Here in the States, I've seen unladen "120V" circuits ranging between 112V to 128V. And if we're doing the center-tapped-transformer thing, as common in the US, then this equates to "224V to 256V" when using the entire secondary winding.
It's quite normal for this sort of variation to occur in the real world, and is going to remain that way until we all see auto-switching multitap transformers by the street, or personal motor-generators, or some similar expensive thing (none of which are ever likely to happen).
I think the most disturbing thing to come out of your comment is that I hadn't realised that 1 pound of gasoline has the same energy as 10 pounds of TNT. That doesn't seem right.
It's not what you've got -- it's how you use it.
According to Wikipedia:
Gas is 44.4 megaJoules per kilogram.
TNT is 4.184 megaJoules per kilogram.
TNT goes snap. Gasoline, by comparison, goes foop.
I'm really not very surprised by these numbers. Gasoline might make a fine explosive, if it were easier to make it go bang instead of just burn. And TNT might be somewhat useful for fueling cars, if it were easier to make it burn instead of explode.
I'm not a kid. I'm not going to school. I won't be going to school. I haven't seen a classroom proper in 17 years.
But I was a kid once; an atypical kid like many here on Slashdot, but a kid nonetheless. I remember being a kid.
And as a kid, I had real problems in school. I hated duplication of effort. I was terrible and slow at writing. I used to be admonished by my teachers:
"You can't use a computer to do that work."
"But are my answers correct?"
"Well yes, they are. But you can't rely on a computer, because when there's a problem to solve, there won't always be a computer around to help you figure it out."
Which, of course, was bullshit. Not long after I gave up on school altogether, computers were crawling out of the woodwork. By the time I became an adult and started making real money doing real things for real people, they were ubiquitous.
Nowadays, I carry a computer in my pants pocket that does things which were unimaginable when I was a kid. I use it all the time. And I keep a laptop nearby. These are tools that I use to help me in my professional career, which involves solving real problems in the real world.
Keeping computers out of a classroom is the same as depriving a mechanic the use of a wrench while insisting that they figure out some more archaic fashion in which to adjust a bolt. It's a useful tool now, it will continue to be useful later, and kids might as well familiarize themselves with using the tools available to them to solve problems as early as possible.
Put a Linux box on your network with any IP in the same subnet as the other machines. It doesn't even need two NICs. Just turn on IP forwarding, make the default gateway on the Linux box the AT&T gateway and make the default gateway on everything else the Linux box. Then you can use egress throttling or whatever you want, because all traffic to the internet is going into and then back out of the Linux box.
Cool.
This not only breaks DHCP, thus making wireless access non-trivial, but by extension ruins local DNS resolution. And without DMZ Plus, it gives me twice as many configurations to fuck with every time I need a port opened from the outside world to some widget or other.
Thanks, but no thanks. I gave up on this level of configuration insanity over a decade ago, and I'm not going back.
It didn't seem like much of a problem with a handful (or even a bunch) of TCP connections, as you seem to report.
It blew itself apart when there were hundreds, or thousands of connections to different hosts*.
You may think of my torrenting practices what you will, but they didn't work with DMZ Plus and a second router. However, they worked fine with neither.
I may try again. I am not hopeful of positive results.
*: Initially, I was using the same OpenWRT router that I had, just the day before, been doing the exact same thing with, only with Roadrunner and no NAT madness, with very similar network speeds and connection counts. It worked fine. UVerse showed up, I configured DMZ Plus, and things went to shit. The experimentation and basic troubleshooting came later, and consistently returned DMZ Plus as the source of problems.
Random TCP connection hangs, killing torrent throughput. Peers would just sit there and do nothing for a time. Partial blocks were common. It was a mess to watch.
Tried with a variety of WRT54G/GL/GS routers, using a variety of popular third-party firmware, and not doing anything fancy. Eventually, I got sick of playing whack-a-mole with an invisible mallet, and realized that the mole itself was the devil.
The various permutations of WRT-based routers I was trying were things that always worked fine on network connections of similar speed which don't do any sort of NAT: Bridged ADSL, PPPOE, DOCSIS...pretty much anything dumb like that.
And, the various permutations of WRT-based routers I was trying were also things that always even worked fine with two in series, each performing NAT. I want to blame the WRTs, but I can't.
And, in all cases, removing the extraneous router and turning off DMZ Plus has always rectified the problem, doing the same workload.
So, yeah: DMZ Plus is broken. Unless it has magically healed itself in the past 1.5 years or so.
Phone cabling, between you and the CO, is always twisted or otherwise arranged in a very specific fashion to accomplish much of the same thing, to keep crosstalk down. You'd think that it wasn't important Way Back When, but Way Back When when these guys were spending a fortune in time and materials installing 100-pair lead-sheathed cable, they tried to get it right so they wouldn't have to go back and do it over again.
The problem with this is not that it doesn't work, or that it is old, but that they didn't see any purpose in making the wire behave outside of the audio range. Meanwhile, modern DSL variants extend up to a few MHz: That it works at all is a testament to the amount of thought that went into designing these old cable plants.
Details in "A manual of telephony", dating to 1905. (Electrical engineering isn't exactly a new science...)
That said: I've got a pretty solid 19.2mbps pipe over VDSL, which passes through a couple of thousand feet of old lead-sheathed cable and a thousand or so feet of more modern stuff. The wire from the modem to the pole is all new Cat5e, but that's just a drop in the bucket.
The fact is the Android platform isn't nearly ready for mainstream consumption (no way I recommend it to non-techies)
Which, you know, is funny. I used to think the same thing.
And then, I started asking non-techie folks about how they like their shiny new Droid. They all seemed to like it just fine as a widget that makes phone calls, does text messages, browses Teh Intarwebs, and plays a cheesy game, and were satisfied with their limited use.
The Compact Disc chose 44050 Hz sampling rate because our ears generally cannot hear anything over 22000 Hz.
Wrong, wrong, and mostly correct.
The sampling rate is 44100 Hz. If Nyquist were the only reason for that, it'd have been a more round number: 40KHz, 44KHz, or maybe 48KHz.
Why, then, was it chosen to be exactly 44.1KHz? Because that's the datarate supported by the video recorders of the time.
Way back then, we didn't have magic storage devices which could deal with hundreds of megabytes of audio easily. So, video recorders were used instead. Specifically, they used professional PAL recorders with PCM adapters to convert the bits to/from video signals.
It just turns out that the rate of (44.1KHz * 16 bits * 2 channels) 1411.2kbps is exactly what can be neatly and reliably encoded on these PAL machines, and that 44.1KHz was reasonable with Nyquist and humans.
This, simply, is how professional digital audio of the time was recorded, before CDs. It only made sense to carry that forward.
Quoth Wikipedia: It is quite interesting to see that the number 44.100 has its roots back in the old German TV system from the 1940s, which used 441 lines at 50 Hz field frequency, resulting in a line frequency of 11.025 kHz, which is exactly 1/4 of the sampling frequency used in Audio CDs today.
And, obviously, had the existing standard been different, they'd have chosen different tradeoffs: Either the bit-depth or the sampling rate could easily have been moved up or down*. 44.1KHz, 16-bit PCM stereo was simply convenient.
*: NTSC PCM adapters show this point neatly, by having a slightly different sampling rate of 44.056KHz.
I keep a (quite old, by this point) USB docking station at work.
I just set the laptop down on my desk, plug in a single USB cable, and the following things come up and start working:
10/100 Ethernet RS-232 serial Parallel PS/2 mouse/keyboard ports 2 more USB 2.0 ports. One of these connects a flatbed scanner, while the other is used for audio.
It works fine (especially since the office LAN is still 100mbps). The Ethernet is roughly as fast as that which is built into my laptop, and handily beats the snot out of 802.11-anything.
An unpowered hub is not just a ball of wirenuts -- there's logic in there that needs power, too.
So, imagine it if you will:
Assume that the drive wants 500mA to spin up, and works fine when plugged into a PC, or into a powered hub.
Now, put a non-powered hub in series with it. Assume that the powered hub itself wants 100mA just for itself.
And, voila! 600mA worth of current draw, from a port specified to only supply 500mA.
Now, since in practice USB is all over the map with regards to available current, it might work, or it might not work.
You're simply experiencing the "might not work" end of this uncertainty.
I further question the merit of using an unpowered hub with a hard disk, anyway. The drive itself already wants all of the power that can be sucked out of a USB port, leaving none left over for even a simple mouse or keyboard.
Quoth Wikipedia:
USB current is allocated in units of 100 mA up to a maximum total of 500 mA per port. Therefore a compliant bus powered hub can have no more than four downstream ports and cannot offer more than four 100 mA units of current in total to downstream devices (since one unit is needed for the hub itself). If more units of current are required by a device than can be supplied by the port it is plugged into, the operating system usually reports this to the user.
Slashdot Mirror
Actually, I can expect that. And I can even show you a pretty graph that indicates folks are doing an increasingly better job with encrypting their wireless networks.
As an anecdote, my own experiences with wardriving in small-town Ohio have been interesting to me. Some towns and neighborhoods are full of wide-open networks. Some are almost completely locked-down. Some people will have two SSIDs for their house, like a WPA-protected network called "Jones" and a second non-encrypted "Jones Guest".
And there's plenty of savvy people out there who even give different family members their own encrypted WLANs, judging from the SSIDs that I see.
Generally speaking, I've seen folks make good progress over the past few years. Gone are the days when I could just open my laptop in any old neighborhood, pick one of several "linksys" APs, and get Internet access.
Patents. You said so yourself.
Their inclination would be to keep the price of plant-based fuels to be above oil, if and as long as it is profitable to do so. Once the patents expire, they'll have to compete with other things (like oil, coal, and other companies using the exact same plant-based processes which used to have patent protection) on a more realistic playing field.
That might be a solution with 12,500 workstations, but at my (much, much smaller) company it ain't gonna work:
According to VMware, it costs $5,000 to get started with 50 client licenses. And additional licenses are $39.
For those prices, it's more profitable to stick with XP for as long as possible, and then run things in virtualized XP Mode once the hardware catches up with the overhead (in the natural course of attrition and replacement).
I agree.
So, then: Should subsidies on HFCS be eliminated? If so: What is it, exactly, that keeps us from removing the subsidies that have an effect on HFCS?
My favorite coffee shop has RJ45 ports at the tables on a switched network.
Still sniffable, obviously, but at least not passively: One must do some amount of ARP poisoning or MAC overflow in order to get much meaningful data.
I've heard of cooking with C4. I think I even have an Army field manual describing it, somewhere...
But cooking with C4 is rather a very different sort of thing than trying to cook with TNT. :)
Easy: HFCS would go up in price. Goods made from HFCS would also go up in price.
This trend would continue until it would became cheaper to use sugar instead, and then the price of sugar would begin to edge up and balance things out.
The bigger question is this: If junk food were more expensive (as described above), would people buy less of it?
Not to nitpick but if you are referring to (pick a country), then anything between about 220V and 240V is "220." Or "240," depending on local parlance.
The two are synonymous. Much as "110," "115," and "120" are.
Power, once distributed, isn't very exact -- no matter what the nationality. Here in the States, I've seen unladen "120V" circuits ranging between 112V to 125V. And if we're doing the center-tapped-transformer thing, then this equates to "224 to 250V".
It's quite normal for this sort of variation to occur in the real world, and is going to remain that way until we all see auto-switching multitap transformers by the street, or personal motor-generators, or something similar (none of which is likely to ever happen).
Quit being a pedant where it isn't required.
Not to nitpick but if you are referring to (pick a country), then anything between about 220V and 240V is "220." Or "240," depending on local parlance.
The two are synonymous. Much as "110," "115," and "120" are.
Power, once distributed, isn't very exact -- no matter what the nationality. Here in the States, I've seen unladen "120V" circuits ranging between 112V to 128V. And if we're doing the center-tapped-transformer thing, as common in the US, then this equates to "224V to 256V" when using the entire secondary winding.
It's quite normal for this sort of variation to occur in the real world, and is going to remain that way until we all see auto-switching multitap transformers by the street, or personal motor-generators, or some similar expensive thing (none of which are ever likely to happen).
Quit being a pedant where it isn't required.
It's not what you've got -- it's how you use it.
According to Wikipedia:
Gas is 44.4 megaJoules per kilogram.
TNT is 4.184 megaJoules per kilogram.
TNT goes snap. Gasoline, by comparison, goes foop.
I'm really not very surprised by these numbers. Gasoline might make a fine explosive, if it were easier to make it go bang instead of just burn. And TNT might be somewhat useful for fueling cars, if it were easier to make it burn instead of explode.
I'm not a kid. I'm not going to school. I won't be going to school. I haven't seen a classroom proper in 17 years.
But I was a kid once; an atypical kid like many here on Slashdot, but a kid nonetheless. I remember being a kid.
And as a kid, I had real problems in school. I hated duplication of effort. I was terrible and slow at writing. I used to be admonished by my teachers:
"You can't use a computer to do that work."
"But are my answers correct?"
"Well yes, they are. But you can't rely on a computer, because when there's a problem to solve, there won't always be a computer around to help you figure it out."
Which, of course, was bullshit. Not long after I gave up on school altogether, computers were crawling out of the woodwork. By the time I became an adult and started making real money doing real things for real people, they were ubiquitous.
Nowadays, I carry a computer in my pants pocket that does things which were unimaginable when I was a kid. I use it all the time. And I keep a laptop nearby. These are tools that I use to help me in my professional career, which involves solving real problems in the real world.
Keeping computers out of a classroom is the same as depriving a mechanic the use of a wrench while insisting that they figure out some more archaic fashion in which to adjust a bolt. It's a useful tool now, it will continue to be useful later, and kids might as well familiarize themselves with using the tools available to them to solve problems as early as possible.
Cool.
This not only breaks DHCP, thus making wireless access non-trivial, but by extension ruins local DNS resolution. And without DMZ Plus, it gives me twice as many configurations to fuck with every time I need a port opened from the outside world to some widget or other.
Thanks, but no thanks. I gave up on this level of configuration insanity over a decade ago, and I'm not going back.
It didn't seem like much of a problem with a handful (or even a bunch) of TCP connections, as you seem to report.
It blew itself apart when there were hundreds, or thousands of connections to different hosts*.
You may think of my torrenting practices what you will, but they didn't work with DMZ Plus and a second router. However, they worked fine with neither.
I may try again. I am not hopeful of positive results.
*: Initially, I was using the same OpenWRT router that I had, just the day before, been doing the exact same thing with, only with Roadrunner and no NAT madness, with very similar network speeds and connection counts. It worked fine. UVerse showed up, I configured DMZ Plus, and things went to shit. The experimentation and basic troubleshooting came later, and consistently returned DMZ Plus as the source of problems.
Random TCP connection hangs, killing torrent throughput. Peers would just sit there and do nothing for a time. Partial blocks were common. It was a mess to watch.
Tried with a variety of WRT54G/GL/GS routers, using a variety of popular third-party firmware, and not doing anything fancy. Eventually, I got sick of playing whack-a-mole with an invisible mallet, and realized that the mole itself was the devil.
The various permutations of WRT-based routers I was trying were things that always worked fine on network connections of similar speed which don't do any sort of NAT: Bridged ADSL, PPPOE, DOCSIS...pretty much anything dumb like that.
And, the various permutations of WRT-based routers I was trying were also things that always even worked fine with two in series, each performing NAT. I want to blame the WRTs, but I can't.
And, in all cases, removing the extraneous router and turning off DMZ Plus has always rectified the problem, doing the same workload.
So, yeah: DMZ Plus is broken. Unless it has magically healed itself in the past 1.5 years or so.
I'd love to hear otherwise.
Phone cabling, between you and the CO, is always twisted or otherwise arranged in a very specific fashion to accomplish much of the same thing, to keep crosstalk down. You'd think that it wasn't important Way Back When, but Way Back When when these guys were spending a fortune in time and materials installing 100-pair lead-sheathed cable, they tried to get it right so they wouldn't have to go back and do it over again.
The problem with this is not that it doesn't work, or that it is old, but that they didn't see any purpose in making the wire behave outside of the audio range. Meanwhile, modern DSL variants extend up to a few MHz: That it works at all is a testament to the amount of thought that went into designing these old cable plants.
Details in "A manual of telephony", dating to 1905. (Electrical engineering isn't exactly a new science...)
That said: I've got a pretty solid 19.2mbps pipe over VDSL, which passes through a couple of thousand feet of old lead-sheathed cable and a thousand or so feet of more modern stuff. The wire from the modem to the pole is all new Cat5e, but that's just a drop in the bucket.
Please describe a method for prioritizing acks using U-Verse with the ubiquitous and irreplaceable 2-wire "residential gateway" which is supplied.
(And, hint: If it involves egress throttling on an external router and DMZ Plus on the gateway, you fail. DMZ Plus is broken.)
If your coffee is bitter, you're doing it wrong.
Which, you know, is funny. I used to think the same thing.
And then, I started asking non-techie folks about how they like their shiny new Droid. They all seemed to like it just fine as a widget that makes phone calls, does text messages, browses Teh Intarwebs, and plays a cheesy game, and were satisfied with their limited use.
Wrong, wrong, and mostly correct.
The sampling rate is 44100 Hz. If Nyquist were the only reason for that, it'd have been a more round number: 40KHz, 44KHz, or maybe 48KHz.
Why, then, was it chosen to be exactly 44.1KHz? Because that's the datarate supported by the video recorders of the time.
Way back then, we didn't have magic storage devices which could deal with hundreds of megabytes of audio easily. So, video recorders were used instead. Specifically, they used professional PAL recorders with PCM adapters to convert the bits to/from video signals.
It just turns out that the rate of (44.1KHz * 16 bits * 2 channels) 1411.2kbps is exactly what can be neatly and reliably encoded on these PAL machines, and that 44.1KHz was reasonable with Nyquist and humans.
This, simply, is how professional digital audio of the time was recorded, before CDs. It only made sense to carry that forward.
Quoth Wikipedia: It is quite interesting to see that the number 44.100 has its roots back in the old German TV system from the 1940s, which used 441 lines at 50 Hz field frequency, resulting in a line frequency of 11.025 kHz, which is exactly 1/4 of the sampling frequency used in Audio CDs today.
And, obviously, had the existing standard been different, they'd have chosen different tradeoffs: Either the bit-depth or the sampling rate could easily have been moved up or down*. 44.1KHz, 16-bit PCM stereo was simply convenient.
*: NTSC PCM adapters show this point neatly, by having a slightly different sampling rate of 44.056KHz.
Bless you.
I keep a (quite old, by this point) USB docking station at work.
I just set the laptop down on my desk, plug in a single USB cable, and the following things come up and start working:
10/100 Ethernet
RS-232 serial
Parallel
PS/2 mouse/keyboard ports
2 more USB 2.0 ports. One of these connects a flatbed scanner, while the other is used for audio.
It works fine (especially since the office LAN is still 100mbps). The Ethernet is roughly as fast as that which is built into my laptop, and handily beats the snot out of 802.11-anything.
If Gig-E is all you want, then just buy that by itself.
It ain't as snazzy as something that sits on a PCI Express bus, but sometimes you gotta deal with what you got.
*shrug*
Oh. Cool.
So it works sort of like this, except for that small detail about how it doesn't exist yet.
Wake me up when it becomes cheaper than doing it the old fashioned way.
*yawn*
Wrong assumptions.
An unpowered hub is not just a ball of wirenuts -- there's logic in there that needs power, too.
So, imagine it if you will:
Assume that the drive wants 500mA to spin up, and works fine when plugged into a PC, or into a powered hub.
Now, put a non-powered hub in series with it. Assume that the powered hub itself wants 100mA just for itself.
And, voila! 600mA worth of current draw, from a port specified to only supply 500mA.
Now, since in practice USB is all over the map with regards to available current, it might work, or it might not work.
You're simply experiencing the "might not work" end of this uncertainty.
I further question the merit of using an unpowered hub with a hard disk, anyway. The drive itself already wants all of the power that can be sucked out of a USB port, leaving none left over for even a simple mouse or keyboard.
Quoth Wikipedia:
USB current is allocated in units of 100 mA up to a maximum total of 500 mA per port. Therefore a compliant bus powered hub can have no more than four downstream ports and cannot offer more than four 100 mA units of current in total to downstream devices (since one unit is needed for the hub itself). If more units of current are required by a device than can be supplied by the port it is plugged into, the operating system usually reports this to the user.
I've had a 1TB "green" WD drive spinning in a Linux-based file-serverish box at home for 3 years.
It works fine, producing never a bad bit, a hiccup, nor any other errant thing.
I've never done any special tweaking with hdparm, or anything else on that drive, since it seemed quite happy by default.
YMMV.
It worked fine for me right after Slashdot posted it.
Now when I try, Firefox wants me to download some sort of zero-byte file named "view" instead of show me a web page.
Which is cool: I didn't think it'd be possible to slashdot Google, but it seems to have happened.
Good job, folks! Keep up the good work.