Packet switching wasn't designed for a different workload than circuit switching. It was designed to virtualize circuit switching in a more robust manner. TCP is designed to have an expensive connection setup followed by an efficient sustained session. The only catch there is that ISPs generally work with things at the IP level, but this doesn't really change things. What's easier to provide quality of service for, a consistent sustained stream, or a wildly bursty and unpredictable usage pattern? The only downside of streaming media and bursty traffic sharing the same connection is to the streaming media.
If you aren't resourced to help them, point them to someone who is. Even if it's a consultant who will charge them money, at least you're giving them an option, instead of just hanging up.
I couldn't find this software on sourceforge or freshmeat. It really troubles me that the US government is using proprietary software to violate our constitutional rights.
I'm not even an Ubuntu user, but I think the whole community would benefit if some major distro said "Okay, stop everything, we're going to spend six weeks on making the distro usable by normal people." Thanks and Kudos to Ubuntu if they lead the way on this.
High power wi-fi card: 50 mW at 2.4 GHz Low power dorm microwave: 700 W at 2.4 GHz
My parents have two identical 2.4 GHz wireless phones. One has the base station next to the wi-fi router, the other has the base station next to the microwave. When the router is going full-blast at 54 Mbps transferring files between a laptop and a desktop, you hear nothing. When someone is making hot chocolate, sometimes you have to cycle through a few channels on the phone to be able to continue your conversation.
I've sat on a committee that's implementing this same sort of initiative at a major university, and the consensus from all factions was that there's still a need (though substantially reduced) for labs. There are many reasons for this:
1) Maintenance: Users suck at keeping their systems in good working order. I personally got my start in IT at the suggestion of a fellow student who said "you should work at the helpdesk" when I got a virus off her computer after she came into the lab at 3 a.m. the night before a major assignment was due, since she couldn't complete it on her system.
2) Reference systems: Grading any kind of programming assignment is an absolute nightmare if you don't know what libraries someone installed to get something working. This isn't really a problem with low-level "hello world" stuff, but when you get into higher-level stuff (particularly graphics) it becomes a big deal. If a TA can say "I will be grading on a lab machine." and all the lab machines have a standard build, everyone wins.
3) High-end workstations: No laptop in the world will hold a candle to that dual G5 in the media lab for video editing, or to the 2 GB RAM, Quadro FX-endowed CAD workstation in the engineering lab. Students really will need these things.
4) Productivity: Why do groups of computer science majors, who each own a desktop, a server, and a laptop, go to the lab? TO GET WORK DONE.
Depending on how well wired your campus is, you can probably lose 3/4 of your labs, but you still need some.
The sun gives us about 1 kw/(m^2) of energy. Regardless of how you're capturing energy, we can't do better than that in the long-long run, and if we're using more than that, we will eventually run out.
The article has been corrected. Note that the maximum number of cached pages, regardless of the number of tabs, defaults to 8, and that's only if you have at least 1 GB of RAM. RTFSC:
I got started on an 8086, but the first one that was really mine to play with was an 8088. We even had an EGA monitor, so I played a lot of games. My dad deleted a bunch of them (after the install disks had been damaged) when I got behind on my homework, but fortunately I had already learned how to hide files and directories in DOS. I've moved on, but he's addicted to one of them to this day.
There was also a wireframe F-14 flight simulator, with practically unplayably poor graphics, that was clock-timed for the 4.77 MHz 8086 processor. It ran fine on the 10 MHz 8088 thanks to the turbo button, but it was physically impossible to send it keyboard input fast enough on a 133 MHz 486 to play it, no matter how much caffeine you were on. Maybe it's playable in qemu now...
What really got me started though was a little dabbling with GW-BASIC, and using WordPerfect 5.1 with CGA graphics. Since WYSIWYG was impossible, managing your formatting carefully required using the Reveal Codes feature. It seems silly, but viewing that showed me how computers think. That was back in second grade, when you're still capable of learning languages easily. I learned how to think recursively in second grade thanks to that lousy interface, and I've been doing it ever since. For most people, it seems to be a concept they'll never truly understand.
Some police departments are using DRM in cameras to prove that photographic evidence has not been tampered with. This is just one of many, many examples people benefit from limiting the capabilities of the user. If you've ever worked in IT, you know how dangerous users can be. Imagine never having to remove gator from someone's computer again, while still giving them privileges to manage their own system.
An anti-DRM software license is just as stupid as RMS deliberately making su insecure because he was mad that he couldn't root a box.
I'd love to tell a few stories, but I don't want to get anyone fired. The short version is that when companies create bureaucracy that reaches down to technical workers, you read about their bankruptcies, restructurings, and layoffs in the newspaper. When they create no bureaucracy at all, they don't last long enough for you to read anything about them in the newspaper. When they shield technical workers from the bureaucracy and let them do their jobs in whatever innovative ways they dream up, you read in the newspapers about how they're making money hand over fist and changing the entire market. If you're a technical worker and you have to fill out more than one form a month that's not an actual part of your primary technical work or related directly to your employment and compensation, your upper management has screwed up.
If they won't pay to send the primary administrator of a critical production system to the standard training course for the product in question, they are doomed to disaster. If your boss refuses to do something about this, polish your resume and then go over his head. Maybe you'll get your training, or maybe you'll get your boss's job, or maybe you'll get a job with another company that knows what it's doing.
Not just that, but having a router makes a lot more sense when you've got broadband. A layer of NAT between you and the outside world is a nice security boost.
A lot of jobs in IT happen because of networking. For my first IT job, there wasn't exactly an opening -- they weren't actively looking for someone -- but I personally knew the manager, and he had a bit of flexibility since it was part-time. I got the job strictly on my merits, but my social connections were what got my foot in the door.
People tend to socialize with their own ethnicity. There need not be anything inherently racist about this (though there sometimes is) but it means that white people tend to know white people. If white people are the ones already working in IT, then other white people are going to be the ones who get the leads, personal recommendations, etc. This will take a very long time to equilibrate, and is in my opinion a very good justification for affirmative action *recruiting* policies.
1) BellSouth attempts to extort Google. 2) Google tells BellSouth to go fuck itself. 3) BellSouth applies routing policies that harm performance for connections with Google. 4) For all IP addresses whose traffic with Google flows across BellSouth's network, Google inserts an extremely prominent explanation of why they're getting degraded performance, and links to local broadband ISPs that do not use BellSouth's backbone to route to Google. 5) Local ISPs that *do* use BellSouth's backbone become very upset. 6) The problem (either BellSouth's policy, or BellSouth itself) goes away.
Of course, if Google decided to do away with their "don't be evil" policy, it would look something like this:
1) BellSouth attempts to extort Google. 2) Google decides to make an example of BellSouth, and demands money only from BellSouth in exchange for the right to carry traffic from Google. 3) BellSouth tells Google to go fuck itself. 4) BellSouth applies routing policies that harm performance for connections with Google. 5) Google completely replaces all pages for users connecting through BellSouth's network with an explanation of what's going on, saying that BellSouth's discriminatory routing policies impose an additional load on their servers and networks and that they are blocking access to protect their network, unless BellSouth is willing to fund additional resources to compensate for the load. Google also includes localized links to other broadband ISPs. 6) Television news picks up the story. 7) Google says "they started it" and tearfully laments how they had to cut off some users for the good of the network. 8) Irate users demand that their ISPs do something about it. 9) Local ISPs take their business elsewhere. 10) BellSouth goes bankrupt. 11) Google buys BellSouth's backbone to prevent future problems of this nature. 12) Nobody ever fucks with Google again.
I'm really glad to see that there's x86-based consumer hardware on the market that finally ditches the BIOS. Its time has come and gone. You could certainly say the same thing about x86 itself, but the BIOS method of interacting with your hardware has made every attempt to update it buggy and idiosyncratic (ACPI on top of BIOS is the latest in a long line of dissapointing arguable enhancements), while updates to the x86 ISA have been been almost completely transparent.
While I have other reservations about EFI, at least it's carrying less legacy baggage. It seems like a reasonable design decision to me. If Microsoft wants to include EFI support in their consumer x86 version of Windows, I'm sure Apple would be happy to watch Microsoft spend the money on it.
We all know this'll be booting Linux and BSD within the week anyway.
When you use a laptop, the screen is closer to your eyes than a standalone LCD monitor, so it makes sense to have smaller pixels. As for losing screen real estate switching from CRT to LCD, I had the same feeling initially when my office got new boxes with LCDs. Now I'm coming to realize just how ridiculous it was to be running at the maximum resolution my CRT was capable of. I no longer suffer from eyestrain or headaches while working long hours. I've got a nice gigantic CRT at home and at my current job, both at a relaxing 1280x1024. I'm guessing you're using a Windows box at work (if you're using a Mac, get Desktop Manager) so you're probably SOL on virtual desktops, but I've found virtual desktops to handle 98% of the cases where I'm ever tempted to go dual-head.
Good point. I would be completely in favor of a law requiring DRM publishers to put master keys in escrow with the copyright office. Unfortunately, our legislators don't seem to honor the "limited time" clause, and the courts are allowing them to get away with it, so good luck convincing them of the need.
If the content providers choose to only distribute their copyrighted works when DRM is in the loop, that's their prerogative. It's our prerogative to ignore it and give our business to those who do not use DRM.
Voluntary DRM is not evil. What is evil is when DRM is legislated into the system, even interfering with those who choose not to have anything to do with it.
Slashdot Mirror
Packet switching wasn't designed for a different workload than circuit switching. It was designed to virtualize circuit switching in a more robust manner. TCP is designed to have an expensive connection setup followed by an efficient sustained session. The only catch there is that ISPs generally work with things at the IP level, but this doesn't really change things. What's easier to provide quality of service for, a consistent sustained stream, or a wildly bursty and unpredictable usage pattern? The only downside of streaming media and bursty traffic sharing the same connection is to the streaming media.
If you aren't resourced to help them, point them to someone who is. Even if it's a consultant who will charge them money, at least you're giving them an option, instead of just hanging up.
No, this is super-duper-scalar!
I couldn't find this software on sourceforge or freshmeat. It really troubles me that the US government is using proprietary software to violate our constitutional rights.
http://www.kinderstart.com:8080/kindertoday/addPos tingForm
I'm submitting this story:
http://news.com.com/2100-1032_3-1011740.html
Thanks AKAImBatman!
I'm not even an Ubuntu user, but I think the whole community would benefit if some major distro said "Okay, stop everything, we're going to spend six weeks on making the distro usable by normal people." Thanks and Kudos to Ubuntu if they lead the way on this.
High power wi-fi card: 50 mW at 2.4 GHz
Low power dorm microwave: 700 W at 2.4 GHz
My parents have two identical 2.4 GHz wireless phones. One has the base station next to the wi-fi router, the other has the base station next to the microwave. When the router is going full-blast at 54 Mbps transferring files between a laptop and a desktop, you hear nothing. When someone is making hot chocolate, sometimes you have to cycle through a few channels on the phone to be able to continue your conversation.
I've sat on a committee that's implementing this same sort of initiative at a major university, and the consensus from all factions was that there's still a need (though substantially reduced) for labs. There are many reasons for this:
1) Maintenance: Users suck at keeping their systems in good working order. I personally got my start in IT at the suggestion of a fellow student who said "you should work at the helpdesk" when I got a virus off her computer after she came into the lab at 3 a.m. the night before a major assignment was due, since she couldn't complete it on her system.
2) Reference systems: Grading any kind of programming assignment is an absolute nightmare if you don't know what libraries someone installed to get something working. This isn't really a problem with low-level "hello world" stuff, but when you get into higher-level stuff (particularly graphics) it becomes a big deal. If a TA can say "I will be grading on a lab machine." and all the lab machines have a standard build, everyone wins.
3) High-end workstations: No laptop in the world will hold a candle to that dual G5 in the media lab for video editing, or to the 2 GB RAM, Quadro FX-endowed CAD workstation in the engineering lab. Students really will need these things.
4) Productivity: Why do groups of computer science majors, who each own a desktop, a server, and a laptop, go to the lab? TO GET WORK DONE.
Depending on how well wired your campus is, you can probably lose 3/4 of your labs, but you still need some.
The sun gives us about 1 kw/(m^2) of energy. Regardless of how you're capturing energy, we can't do better than that in the long-long run, and if we're using more than that, we will eventually run out.
The article has been corrected. Note that the maximum number of cached pages, regardless of the number of tabs, defaults to 8, and that's only if you have at least 1 GB of RAM. RTFSC:
s history/src/nsSHistory.cpp#161
http://lxr.mozilla.org/seamonkey/source/docshell/
If you're unhappy with the memory usage with 50 tabs open, I advise the following workaround:
DON'T DO THAT.
I got started on an 8086, but the first one that was really mine to play with was an 8088. We even had an EGA monitor, so I played a lot of games. My dad deleted a bunch of them (after the install disks had been damaged) when I got behind on my homework, but fortunately I had already learned how to hide files and directories in DOS. I've moved on, but he's addicted to one of them to this day.
There was also a wireframe F-14 flight simulator, with practically unplayably poor graphics, that was clock-timed for the 4.77 MHz 8086 processor. It ran fine on the 10 MHz 8088 thanks to the turbo button, but it was physically impossible to send it keyboard input fast enough on a 133 MHz 486 to play it, no matter how much caffeine you were on. Maybe it's playable in qemu now...
What really got me started though was a little dabbling with GW-BASIC, and using WordPerfect 5.1 with CGA graphics. Since WYSIWYG was impossible, managing your formatting carefully required using the Reveal Codes feature. It seems silly, but viewing that showed me how computers think. That was back in second grade, when you're still capable of learning languages easily. I learned how to think recursively in second grade thanks to that lousy interface, and I've been doing it ever since. For most people, it seems to be a concept they'll never truly understand.
Why's the work more interesting? If it's a company that's going somewhere, you might not be making less for very long.
Some police departments are using DRM in cameras to prove that photographic evidence has not been tampered with. This is just one of many, many examples people benefit from limiting the capabilities of the user. If you've ever worked in IT, you know how dangerous users can be. Imagine never having to remove gator from someone's computer again, while still giving them privileges to manage their own system.
An anti-DRM software license is just as stupid as RMS deliberately making su insecure because he was mad that he couldn't root a box.
...I think there's a much bigger swords/plowshares issue with the John Deere engine than the OS.
I'd love to tell a few stories, but I don't want to get anyone fired. The short version is that when companies create bureaucracy that reaches down to technical workers, you read about their bankruptcies, restructurings, and layoffs in the newspaper. When they create no bureaucracy at all, they don't last long enough for you to read anything about them in the newspaper. When they shield technical workers from the bureaucracy and let them do their jobs in whatever innovative ways they dream up, you read in the newspapers about how they're making money hand over fist and changing the entire market. If you're a technical worker and you have to fill out more than one form a month that's not an actual part of your primary technical work or related directly to your employment and compensation, your upper management has screwed up.
If they won't pay to send the primary administrator of a critical production system to the standard training course for the product in question, they are doomed to disaster. If your boss refuses to do something about this, polish your resume and then go over his head. Maybe you'll get your training, or maybe you'll get your boss's job, or maybe you'll get a job with another company that knows what it's doing.
Not just that, but having a router makes a lot more sense when you've got broadband. A layer of NAT between you and the outside world is a nice security boost.
A lot of jobs in IT happen because of networking. For my first IT job, there wasn't exactly an opening -- they weren't actively looking for someone -- but I personally knew the manager, and he had a bit of flexibility since it was part-time. I got the job strictly on my merits, but my social connections were what got my foot in the door.
People tend to socialize with their own ethnicity. There need not be anything inherently racist about this (though there sometimes is) but it means that white people tend to know white people. If white people are the ones already working in IT, then other white people are going to be the ones who get the leads, personal recommendations, etc. This will take a very long time to equilibrate, and is in my opinion a very good justification for affirmative action *recruiting* policies.
1) BellSouth attempts to extort Google.
2) Google tells BellSouth to go fuck itself.
3) BellSouth applies routing policies that harm performance for connections with Google.
4) For all IP addresses whose traffic with Google flows across BellSouth's network, Google inserts an extremely prominent explanation of why they're getting degraded performance, and links to local broadband ISPs that do not use BellSouth's backbone to route to Google.
5) Local ISPs that *do* use BellSouth's backbone become very upset.
6) The problem (either BellSouth's policy, or BellSouth itself) goes away.
Of course, if Google decided to do away with their "don't be evil" policy, it would look something like this:
1) BellSouth attempts to extort Google.
2) Google decides to make an example of BellSouth, and demands money only from BellSouth in exchange for the right to carry traffic from Google.
3) BellSouth tells Google to go fuck itself.
4) BellSouth applies routing policies that harm performance for connections with Google.
5) Google completely replaces all pages for users connecting through BellSouth's network with an explanation of what's going on, saying that BellSouth's discriminatory routing policies impose an additional load on their servers and networks and that they are blocking access to protect their network, unless BellSouth is willing to fund additional resources to compensate for the load. Google also includes localized links to other broadband ISPs.
6) Television news picks up the story.
7) Google says "they started it" and tearfully laments how they had to cut off some users for the good of the network.
8) Irate users demand that their ISPs do something about it.
9) Local ISPs take their business elsewhere.
10) BellSouth goes bankrupt.
11) Google buys BellSouth's backbone to prevent future problems of this nature.
12) Nobody ever fucks with Google again.
...the entire database would fit on just one sheet of A(-24) paper. (Yes, I actually did the math.)
I'm really glad to see that there's x86-based consumer hardware on the market that finally ditches the BIOS. Its time has come and gone. You could certainly say the same thing about x86 itself, but the BIOS method of interacting with your hardware has made every attempt to update it buggy and idiosyncratic (ACPI on top of BIOS is the latest in a long line of dissapointing arguable enhancements), while updates to the x86 ISA have been been almost completely transparent.
While I have other reservations about EFI, at least it's carrying less legacy baggage. It seems like a reasonable design decision to me. If Microsoft wants to include EFI support in their consumer x86 version of Windows, I'm sure Apple would be happy to watch Microsoft spend the money on it.
We all know this'll be booting Linux and BSD within the week anyway.
When you use a laptop, the screen is closer to your eyes than a standalone LCD monitor, so it makes sense to have smaller pixels. As for losing screen real estate switching from CRT to LCD, I had the same feeling initially when my office got new boxes with LCDs. Now I'm coming to realize just how ridiculous it was to be running at the maximum resolution my CRT was capable of. I no longer suffer from eyestrain or headaches while working long hours. I've got a nice gigantic CRT at home and at my current job, both at a relaxing 1280x1024. I'm guessing you're using a Windows box at work (if you're using a Mac, get Desktop Manager) so you're probably SOL on virtual desktops, but I've found virtual desktops to handle 98% of the cases where I'm ever tempted to go dual-head.
If you're doing it right, the girlfriend *is* your laptop.
Good point. I would be completely in favor of a law requiring DRM publishers to put master keys in escrow with the copyright office. Unfortunately, our legislators don't seem to honor the "limited time" clause, and the courts are allowing them to get away with it, so good luck convincing them of the need.
If the content providers choose to only distribute their copyrighted works when DRM is in the loop, that's their prerogative. It's our prerogative to ignore it and give our business to those who do not use DRM.
Voluntary DRM is not evil. What is evil is when DRM is legislated into the system, even interfering with those who choose not to have anything to do with it.