There is a safety aspect to this also. By distributing power only in the amount, manner and moment that a device requests it, we wouldn't need to have 120V outlets active throughout our buildings at all times. Electrocution could almost be eliminated if power was intelligently distributed and monitored for noise, over-current, and ground faults.
- The full 120V (or whatever the device needs) is only sent when the device actually needs it. You could drop a plugged-in hair drying into the tub and if it was off at the time there would be no danger.
- If the aforementioned hair dryer was on at the time, the power usage would change from 80% resistive and 20% motor to 100% resistive and the amount of current drawn would increase, and there would be a ground fault. Halt power delivery on any unexpected change in usage parameters.
- A toddler sticking a fork into an outlet would be safe. Unless the fork specifically knows how to ask for power, none would be sent.
- Devices that only need 5V only get 5V. You can charge your electric razor next to the sink and not worry about it.
IIRC, the main issue was more to do with a crappy sh or csh truncating a long path. The underlying find, xargs, and grep programs have probably been quite capable for years now.
The issue I remember most clearly is doing "find/" and having it descend into/net or/mnt and spending waaaay too long searching all the remote mounts that I'd visited recently because I forgot to say -xdev.
>> Keys you expect to repeat don't. That triple-tap thing holds firm for everything. Even backspace. Even the arrow keys.
Here's my little wish. I'd like a keyboard with pressure sensitive keys. The idea is:
1. movement keys (arrows, page up/down, backspace) repeat like normal, but repeat faster as you press harder. there is tactile feedback at the points where repeat start accelerating and at the limit point where it won't go any faster.
2. normal keys that correspond to printable characters do not repeat, unless you press somewhat harder than usual. they repeat at a constant speed.
The only real threat to your bottom line is that he'll come out with a competing product based on your IP. If he does that, you can send the lawyers after him with a clear conscience.
If he doesn't do that, then don't worry about it. He's not making any money at your expense, and you're saving money by not paying his salary anymore. Getting all paranoid and angry about this is just wasting YOUR mental effort.
In linux you can run a program as a different user without logging out. If it's a setuid program, you don't even have to type in a password. Depending on how the idea is implemented, real normal people might not even notice anything was different until they tried to download a dodgy zip file or something.
What I'd like to see is even more user granularity. One account for browsing the web, another for reading email, another for ftp'ing. Even if you download or click on some malware, not only is your OS protected, but now your user id's files are also safe.
When it comes time to actually use the files you downloaded, there should be a malware-scanning chown that checks the file is safe before assigning it over to you, perhaps on top of a check that firefox's chroot jail is not disturbed.
I'm actually more concerned that China's industrialization will ratchet them up to per capita oil consumption rates that match the US, Europe, or Japan. Their cost structure will shift from labor-cheap to petro-expensive. While they have made an incredible investment in hydro, the trucks, buses, cars, and planes still need oil-derived fuel, and the more their economy grows, the more oil they'll need.
Their industrial growth, and the resultant need for oil, is occuring just as the price and supply of oil have re-asserted themselves as global issues. It would be foolish to guess at the eventual impact, but the bottom line is that China will not have the same 100-years-of-cheap-energy growth period that the US and Europe had.
The US is actually working their consumption down (per capita, and per GDP, if not overall) and if I may say so has quite a lot of wiggle room in the form of a fleet of millions of SUVs and hundreds of oil-fired power plants that can gradually be de-commissioned as oil prices rise and make them uneconomical.
China has the advantage that they can use the latest, most efficient technology. However, the up-front costs of using for example hybrid vehicles or installing systems to recapture waste heat from industrual processes may simply be too high for now.
Not to be a downer, but replay and tivo services are actually not that expensive. I have the grandfathered-in $9.95/mo service from replay and that is my lowest monthly expense. Honestly, I spend more on cola.
What I want is a $300 PC that will replace my mortgage, property taxes, and/or car insurance.
I don't work there no more (disclaimer: my leaving had nothing to do with their network architecture).
But I know a network consult who has described to me far worse scenarios: - Unstable networks that don't behave the same if the various elements are powered up in a different order - People bringing in their own wifi equipment and creating and routing new subnets on their own initiative, and the network admin people don't see it for a couple of MONTHS - Cheapo hubs that assume all equipment is at least FDX 10baseT. Works fine until you move that vintage but still-useful sparcstation onto one of them. (The microsofties blame Sun for being broken--a similar vintage windows box wouldn't even have ethernet.) - Run out of ports in an incovenient part of the building? Don't bother with a home run, just put a linksys 8 porter resting on a ceiling tile and start plugging. Run power to it using the cheapest extension cord you can find. - Subset of things plugged into UPS's according to no particular plan except maybe some manager demanding to be able to print TPS reports during a power outage or some equally insane requirement. Other stuff on surge suppressors with no visible joules rating (i.e. useless). - Configuring application firewalls or even DMZ's takes extra effort and thus just doesn't get done.
Honestly, this kind of stuff isn't the extreme, it's the norm.
A top-quality network architecture created and installed by pros using top-notch equipment is really cool, but companies are realizing that a little network downtime doesn't mean their company is going to grind to a halt. Their web site is colo'd and independent. People can still open and edit documents, make phone calls (unless they have VOIP, suckas), have meetings, etc. Revenue generating people like sales and marketing are out at customer sites using their networks anyway.
And when the crappy consumer grade stuff fails they can swap in a new one for cheaper than two weeks maintenance on a cisco box. Because of the core strength of ethernet and TCP/IP, most failures are localized to a small group so it's really not a bad strategy if cost-effectiveness is any concern.
Count your blessings. You'd be amazed at how many small to medium sized companies (2000+ employees) have one Cisco router in a rack somewhere and use consumer grade linksys or d-link 10/100 switches everywhere else.
For sending email and word docs around, you really don't need the whole Cisco hierarchy. On the other hand, If you're sending uncompressed production video around, it's not enough.
They might as well be marketing a new kind of shampoo. The ingredients that do the cleaning are the same as before, it's just a different color and smells like lilacs. Whoop dee frackin doo.
The only benefit your get from monster cables is a perhaps slightly lower resistance. That is all. The higher resistance of standard wires can easily be overcome by "turning up the volume".
Resistance is slightly more important than you think. The additional factors are that (a) speakers don't have resistance, they have impedance and (b) the speaker's stated impedance (e.g. 8 ohms) is only nominal, or an average. The actual impedance varies with frequency and determined by the design of the crossover and the responses of the different drivers.
So long as the total resistance of the wires is minimal (fraction of an ohm) the effect on frequency response is minimal. But if it gets above an ohm, the resistance of the wires and the varying impedance of the speakers combine to form a filter with significant non-linearities. Usual symptom is weak bass, since the woofers load the most heavily.
The jitter stuff is bogus. All components capable of receiving digital signals buffer them and remove them from the buffer according to their own clock. Only the clock on the final DAC could actually effect sound quality, and then probably only by a microsecond or so. The lag time on moving a speaker cone is gigantic in comparison.
If you're going to use twistlocks at least put some fuses in there. If enough people start using mains type connectors for speakers, eventually somebody, somewhere, will be drunk enough at a party to connect a speaker up to the dryer outlet and possibly start a fire.
A more audiophile angle is that connectors are undesirable in the first place. They typically have more resistance than a little extra cable would plus they tend to corrode or dislodge faster. Just run the cable (of whatever type) straight from the amp to the speakers.
Slashdot Mirror
carbon nanotubes...that's awfully similar to the Inanimate Carbon Rod.
There is a safety aspect to this also. By distributing power only in the amount, manner and moment that a device requests it, we wouldn't need to have 120V outlets active throughout our buildings at all times. Electrocution could almost be eliminated if power was intelligently distributed and monitored for noise, over-current, and ground faults.
- The full 120V (or whatever the device needs) is only sent when the device actually needs it. You could drop a plugged-in hair drying into the tub and if it was off at the time there would be no danger.
- If the aforementioned hair dryer was on at the time, the power usage would change from 80% resistive and 20% motor to 100% resistive and the amount of current drawn would increase, and there would be a ground fault. Halt power delivery on any unexpected change in usage parameters.
- A toddler sticking a fork into an outlet would be safe. Unless the fork specifically knows how to ask for power, none would be sent.
- Devices that only need 5V only get 5V. You can charge your electric razor next to the sink and not worry about it.
IIRC, the main issue was more to do with a crappy sh or csh truncating a long path. The underlying find, xargs, and grep programs have probably been quite capable for years now.
/" and having it descend into /net or /mnt and spending waaaay too long searching all the remote mounts that I'd visited recently because I forgot to say -xdev.
The issue I remember most clearly is doing "find
Anyone remember xargs?
/dev/null
find . | xargs grep foo
ah, the good old days.
>> No constitutional right for government to give you adult content. They are paying for it.
And who is paying for the government?
>> Keys you expect to repeat don't. That triple-tap thing holds firm for everything. Even backspace. Even the arrow keys.
Here's my little wish. I'd like a keyboard with pressure sensitive keys. The idea is:
1. movement keys (arrows, page up/down, backspace) repeat like normal, but repeat faster as you press harder. there is tactile feedback at the points where repeat start accelerating and at the limit point where it won't go any faster.
2. normal keys that correspond to printable characters do not repeat, unless you press somewhat harder than usual. they repeat at a constant speed.
Is it just me or does centos remind you of breath mints or something?
The only real threat to your bottom line is that he'll come out with a competing product based on your IP. If he does that, you can send the lawyers after him with a clear conscience.
If he doesn't do that, then don't worry about it. He's not making any money at your expense, and you're saving money by not paying his salary anymore. Getting all paranoid and angry about this is just wasting YOUR mental effort.
A sensible approach is to have one secure "launcher"-like program that is setuid so that you don't have to trust random applications to be honest.
In linux you can run a program as a different user without logging out. If it's a setuid program, you don't even have to type in a password. Depending on how the idea is implemented, real normal people might not even notice anything was different until they tried to download a dodgy zip file or something.
What I'd like to see is even more user granularity. One account for browsing the web, another for reading email, another for ftp'ing. Even if you download or click on some malware, not only is your OS protected, but now your user id's files are also safe.
When it comes time to actually use the files you downloaded, there should be a malware-scanning chown that checks the file is safe before assigning it over to you, perhaps on top of a check that firefox's chroot jail is not disturbed.
I'm actually more concerned that China's industrialization will ratchet them up to per capita oil consumption rates that match the US, Europe, or Japan. Their cost structure will shift from labor-cheap to petro-expensive. While they have made an incredible investment in hydro, the trucks, buses, cars, and planes still need oil-derived fuel, and the more their economy grows, the more oil they'll need.
Their industrial growth, and the resultant need for oil, is occuring just as the price and supply of oil have re-asserted themselves as global issues. It would be foolish to guess at the eventual impact, but the bottom line is that China will not have the same 100-years-of-cheap-energy growth period that the US and Europe had.
The US is actually working their consumption down (per capita, and per GDP, if not overall) and if I may say so has quite a lot of wiggle room in the form of a fleet of millions of SUVs and hundreds of oil-fired power plants that can gradually be de-commissioned as oil prices rise and make them uneconomical.
China has the advantage that they can use the latest, most efficient technology. However, the up-front costs of using for example hybrid vehicles or installing systems to recapture waste heat from industrual processes may simply be too high for now.
Not to be a downer, but replay and tivo services are actually not that expensive. I have the grandfathered-in $9.95/mo service from replay and that is my lowest monthly expense. Honestly, I spend more on cola.
What I want is a $300 PC that will replace my mortgage, property taxes, and/or car insurance.
i wuv my ibook!
If sending email on weekends is so damned important to your business why do you only have one ISP?
Is there a way to pop a (paid for) DVD, DVD-A, SACD, or DTS-CD into my computer and get MP3's out of it?
I don't work there no more (disclaimer: my leaving had nothing to do with their network architecture).
:-(
But I know a network consult who has described to me far worse scenarios:
- Unstable networks that don't behave the same if the various elements are powered up in a different order
- People bringing in their own wifi equipment and creating and routing new subnets on their own initiative, and the network admin people don't see it for a couple of MONTHS
- Cheapo hubs that assume all equipment is at least FDX 10baseT. Works fine until you move that vintage but still-useful sparcstation onto one of them. (The microsofties blame Sun for being broken--a similar vintage windows box wouldn't even have ethernet.)
- Run out of ports in an incovenient part of the building? Don't bother with a home run, just put a linksys 8 porter resting on a ceiling tile and start plugging. Run power to it using the cheapest extension cord you can find.
- Subset of things plugged into UPS's according to no particular plan except maybe some manager demanding to be able to print TPS reports during a power outage or some equally insane requirement. Other stuff on surge suppressors with no visible joules rating (i.e. useless).
- Configuring application firewalls or even DMZ's takes extra effort and thus just doesn't get done.
Honestly, this kind of stuff isn't the extreme, it's the norm.
A top-quality network architecture created and installed by pros using top-notch equipment is really cool, but companies are realizing that a little network downtime doesn't mean their company is going to grind to a halt. Their web site is colo'd and independent. People can still open and edit documents, make phone calls (unless they have VOIP, suckas), have meetings, etc. Revenue generating people like sales and marketing are out at customer sites using their networks anyway.
And when the crappy consumer grade stuff fails they can swap in a new one for cheaper than two weeks maintenance on a cisco box. Because of the core strength of ethernet and TCP/IP, most failures are localized to a small group so it's really not a bad strategy if cost-effectiveness is any concern.
Sorry, no pictures
Count your blessings. You'd be amazed at how many small to medium sized companies (2000+ employees) have one Cisco router in a rack somewhere and use consumer grade linksys or d-link 10/100 switches everywhere else.
For sending email and word docs around, you really don't need the whole Cisco hierarchy. On the other hand, If you're sending uncompressed production video around, it's not enough.
Cool, so now all you need is a 1000 year power source for the trigger to rupture the membrane...
If someone steals my cash it isn't going to screw up my credit rating for the next 3 years.
On my planet, NASA was created in 1958 and is only 47 years old.
They might as well be marketing a new kind of shampoo. The ingredients that do the cleaning are the same as before, it's just a different color and smells like lilacs. Whoop dee frackin doo.
The only benefit your get from monster cables is a perhaps slightly lower resistance. That is all. The higher resistance of standard wires can easily be overcome by "turning up the volume".
Resistance is slightly more important than you think. The additional factors are that (a) speakers don't have resistance, they have impedance and (b) the speaker's stated impedance (e.g. 8 ohms) is only nominal, or an average. The actual impedance varies with frequency and determined by the design of the crossover and the responses of the different drivers.
So long as the total resistance of the wires is minimal (fraction of an ohm) the effect on frequency response is minimal. But if it gets above an ohm, the resistance of the wires and the varying impedance of the speakers combine to form a filter with significant non-linearities. Usual symptom is weak bass, since the woofers load the most heavily.
The jitter stuff is bogus. All components capable of receiving digital signals buffer them and remove them from the buffer according to their own clock. Only the clock on the final DAC could actually effect sound quality, and then probably only by a microsecond or so. The lag time on moving a speaker cone is gigantic in comparison.
If you're going to use twistlocks at least put some fuses in there. If enough people start using mains type connectors for speakers, eventually somebody, somewhere, will be drunk enough at a party to connect a speaker up to the dryer outlet and possibly start a fire.
A more audiophile angle is that connectors are undesirable in the first place. They typically have more resistance than a little extra cable would plus they tend to corrode or dislodge faster. Just run the cable (of whatever type) straight from the amp to the speakers.