I know it's not the best explanation but the way to think about it is this: You use linux, but you don't represent the majority.
I was merely curious, not pushing any particular agenda. In my day job now it is largely HTRI and ASPEN which are both Windows programs.
While I agree that the majority of engineers and scientists don't use supercomputers, I can say that very very few supercomputers run Windows -- the TOP500 lists only 5 such supercomputers. If you are doing any serious work in the high performance space (nuclear, molecular, weather,...) you are going to be seeing a lot of Unix(-like) systems.
I find it easier to purchase the Windows license (which is usually in built in the cost of the computer anyway) and the 5000 Euro worth of licenses I need
I did my MS in chemical engineering focused on quantum chemistry / molecular simulation / molecular modeling / "nanotechnology". In my field the mainstays all run on clustered supercomputers running some form of Unix: Gaussian (which has a Windows version too), DL_POLY, VASP, MOPAC, Cerius2,... Even the visualization tools often were Unix-only requiring an X11 server. Though some of the grad students wished for more Windows packages, it was pretty much a given that doing real work in quantum chemistry means learning to love Unix.
I'm curious: which Windows-only packages are hot in your field?
If they have to consult their notes for every question (regardless of format) then they should fail the test because they don't understand the material adequately.
You are so clearly not an engineer.
A very common "easy" question in thermodynamics is to calculate some value -- say fugacity -- using three different methods. Each method means one main complex equation plus 2-5 supporting equations, and a single character off any one of them will dramatically change the answer. Oh, and applying each equation might mean looking up 4 different constants for the substance of interest. Question 1 is only worth 20% btw.
The you get to question 2 which requires interpolating between data points from six sets of tables....repeat for the next three questions...
And when you are taking a regular course load you've got 4 such exams every month, each taking 2-3 hours. As another engineer put it to me, EVERY engineering exam is like the final exam of a "regular" (non-engineering/non-science) course. At some places we do more testing in a single semester than a business major will do across their entire undergraduate program.
I suppose though that the minute the first advertisement appeared on the web years ago the future was written in stone.
It wasn't the web, it was Usenet (Canter and Siegel).
The Internet's fatal flaw in its design was that the original users were system and network administrators. This created an immediate conflict of interest: once the Internet was capable of passing messages between hosts they thought it was capable of passing messages between users because there would always be other system and network admins around to make that work. They did not design an Internet that could pass messages when no admins were around. And now 35 years later that's what we need: an Internet that works to connect users even when every ISP and "content provider" has a compelling business interest in preventing users from directly speaking to each other.
We've returned to the world of 1981. The AT&T monopoly is still in force, only now it's the consumer ISP oligopoly. Today's ISPs oversell their capacity just as the Baby Bells did their voice lines. Hayes has just released their first Smartmodem, only we call it darknet today (I2P, TOR, etc). Today's SMTP email is 1981's mail box; today's HTML web browser is 1981's color television.
Just as 1981's BBS technology connected fringe elements of society into their own distinct sub-culture that came to consider the mainstream "just noise", so will tomorrow's darknet users be able to develop their own culture out of the noise that is the public Internet. We'll still have Google and Comcast and Verizon and everyone else, but the darknets of tomorrow will be advertisement free, censorship free, and wiretap free. Users won't mind paying a small price in network latency to gain those freedoms, just as they didn't mind spending days if necessary to download just a few treasured minutes of music and video that would never be played over the mainstream radio and television.
It's coming. They know it and we know it. We can all feel it in the air. Eben Moglin has written a call to arms. Projects are popping out of the woodwork with the stated goal of taking back our Internet. Daniel Suarez's awesome book "Daemon" which features a really cool darknet is shelved in the fiction section rather than sci-fi.
I don't believe your 73rd percentile stat at all given that the mean and median (so 50th percentile) are roughly $45k.
You can get those stats from irs.gov. Back in 2003 I checked it out: I was making $75K and that was the 80th percentile; at the time $35K was the median. At the top 5% range it gets absolutely ridiculous - see the L Curve (biased but still accurate numbers).
That's the major cause of the current recession - the fact that people didn't save and used credit cards to spend way more than they earned (at least you're not doing that).
The problem wasn't people spending - spending is great for the economy. The problem is that the economy simultaneously requires a large middle class to keep spending and a smaller workforce to keep producing. Since the workforce and consumer class are the same people, eventually layoffs must have an effect. For the last 30 years we've been aggressively pursuing greater efficiency (woo!) but have not been able to keep up with demand for new jobs.
Ultimately the problem is structural: we are moving towards an asymptote of one self-healing machine, owned by one person, that creates all of the goods of the world. Think of a single Star Trek replicator operated by one Ferrengi. The IRS has only one W-2 to audit, and the Ferrengi insists that the income tax is drastically unfair as he is the only one paying any of it.
Our purpose is to collect and report notable knowledge
Because no "true encyclopedia" would have articles on things that aren't notable, it stands to reason that anything that Wikipedia defines as notable must be so, because Wikipedia is a true encyclopedia.
If you've been here 30 years then you'll know that the various Eternal Septembers were unavoidable. GUI was going to come to wipe out the more efficient text interfaces; personal computers would have to climb the long slog to mainframe-like architecture one baby step at a time; the network effect was destined to come along and wipe out whole sectors of competition in word processors, spreadsheets, operating systems, and network protocols.
Essentially everything not made by Microsoft was better in a technical sense, but for every user willing to spend ten minutes to learn how their software worked there were a hundred users who just wanted to click on the first thing they saw and then complain to the help desk when they had no clue what was going on.
Microsoft raked in the cash, but it was the users in the end who were to blame.
"Each affected program will have to be patched separately."
And this is why Linux package managers that know how to handle shared library dependencies are better than one-click installers that bring along their own versions of the libraries.
Perhaps we could model it after the old BBS and UUCP-style networks: very very fast local meshes connected to each other over slower VPN links carried over the existing Internet. Make P2P applications prefer the local mesh and put in some large caching servers ala Akamai. Eventually people design little solar-powered buckets with 2-5W autoconfiguring repeater nodes for about $100 and you can spend an afternoon covering a 30-mile stretch off the local interstate. After enough go up people start noticing that things in the "freedom net" come down a whole lot faster than the public internet. (But then the problem become how does one anonymously *purchase* those devices. Because I could see lawmakers moving to regulate their sale ala Sudafed because of kiddie porn/terrorists/whatever.)
The harder part I think is the client-side technology: making browsers, IM clients, etc. aware that there are two Internets that must be treated distinctly and somehow expose that to the user. And making them automatically enforce privacy setting like stripping out EXIF data on every jpg picture upload and using encrypted email by default with gigantic warnings when the recipient's key isn't in the keyring or PKI.
But the possibilities...oh the possibilities. Imagine a "freedom net" without an IANA that doles out IPv6 address space only to those who can pay; dynamic DNS that costs nothing and can never be traced to a real name; the ability to have a real anonymous handle that stays that way; a new Usenet that isn't dripping in spam; end-to-end encryption both on the wire and in the application layer; no more "trusted CAs" to enable MITM for various governments; an Internet with hundreds of available routes between any two hosts, and that won't go nuts because of bad routing records. Basically, what the Internet was supposed to be before "routing around damage" was optimized away in the name of commercial efficiency.
You can SSH into it decently fast, but the lack of a hardware math unit adds around 5-10 seconds of delay when sshing from it to another computer.
I have never experienced this - my Sheevaplug (1st gen maybe?) has 1/2 second or less response time ssh'ing outbound (similar to all of my other desktops). Is it really SSH or maybe DNS that is the issue?
like one which is going to go get pelted around in the ocean
Interesting. I worked on the Texas Automated Buoy System when we moved from custom boards to PC/104-based Linux systems. Are you allowed to say more about this project?
Let's say we've all got "freedom boxes" at home. (And I've already got a Sheevaplug, it's very handy.)
So: I want to access the "freedom network" from my laptop, and I see my encrypted access point and say "use this". I use my Diaspora social networking stuff, my tweet-clone, whatever, and now I want to do a Google search: does my traffic to Google traverse the "freedom network" ala tor? If so, how does Google not remember who I am from a prior session cookie when I was on the public Net? What about performance? What about hostile trusted CAs like Etilsat and CNNIC?
But suppose all of that is solved, how am I reachable from the rest of the public Net? Do we all use one dyndns service that can be brought down ala Napster? If we don't, how do we make a distributed dyndns that is itself not subject to MITM or traffic analysis?
But suppose all of that is solved, how do we prevent one leak on the public Net from linking our "freedom network" identity(ies) to our "real life" Facebook/LinkedIn/Twitter/etc. identity?
I don't think the governments will be caught flat-footed this time around. I can't name a single government or corporate entity that would want widespread adoption of these freedom boxes. It is one of the few things that would unite such enemies as the USA and China, Israel and Iran, Apple and Microsoft.
Slashdot Mirror
Also a "person" like BP, Exxon, Halliburton, IBM, Nike, etc.
If a computer can be infected with a virus simply by plugging in a USB storage device, you're doing it wrong.
That's only true if the USB device is a generic mass storage device. If instead it pretends to be a keyboard it can do a lot of damage within seconds.
How is a virus on someone else's machine going to disable the firewall on my machine?
"It's not a virus, it's a worm!"
I know it's not the best explanation but the way to think about it is this: You use linux, but you don't represent the majority.
I was merely curious, not pushing any particular agenda. In my day job now it is largely HTRI and ASPEN which are both Windows programs.
While I agree that the majority of engineers and scientists don't use supercomputers, I can say that very very few supercomputers run Windows -- the TOP500 lists only 5 such supercomputers. If you are doing any serious work in the high performance space (nuclear, molecular, weather, ...) you are going to be seeing a lot of Unix(-like) systems.
The government inserting money anywhere is a bad idea. It mostly gets wasted with no real benefit to show for it.
Agreed. Let's stop inserting money into the large defense contractors first.
I find it easier to purchase the Windows license (which is usually in built in the cost of the computer anyway) and the 5000 Euro worth of licenses I need
I did my MS in chemical engineering focused on quantum chemistry / molecular simulation / molecular modeling / "nanotechnology". In my field the mainstays all run on clustered supercomputers running some form of Unix: Gaussian (which has a Windows version too), DL_POLY, VASP, MOPAC, Cerius2, ... Even the visualization tools often were Unix-only requiring an X11 server. Though some of the grad students wished for more Windows packages, it was pretty much a given that doing real work in quantum chemistry means learning to love Unix.
I'm curious: which Windows-only packages are hot in your field?
To a liberal arts major you sound profound. To an engineer you sound absurd.
What is the liberal arts / engineering version of Poe's Law?
If they have to consult their notes for every question (regardless of format) then they should fail the test because they don't understand the material adequately.
You are so clearly not an engineer.
A very common "easy" question in thermodynamics is to calculate some value -- say fugacity -- using three different methods. Each method means one main complex equation plus 2-5 supporting equations, and a single character off any one of them will dramatically change the answer. Oh, and applying each equation might mean looking up 4 different constants for the substance of interest. Question 1 is only worth 20% btw.
The you get to question 2 which requires interpolating between data points from six sets of tables. ...repeat for the next three questions...
And when you are taking a regular course load you've got 4 such exams every month, each taking 2-3 hours. As another engineer put it to me, EVERY engineering exam is like the final exam of a "regular" (non-engineering/non-science) course. At some places we do more testing in a single semester than a business major will do across their entire undergraduate program.
See the b43 driver and b43-fwcutter utility.
2.11 Apps that duplicate apps already in the App Store may be rejected, particularly if there are many of them
Does this mean that a few shitty apps in the App Store that do function X can prevent approval of a new app that does function X much better?
I suppose though that the minute the first advertisement appeared on the web years ago the future was written in stone.
It wasn't the web, it was Usenet (Canter and Siegel).
The Internet's fatal flaw in its design was that the original users were system and network administrators. This created an immediate conflict of interest: once the Internet was capable of passing messages between hosts they thought it was capable of passing messages between users because there would always be other system and network admins around to make that work. They did not design an Internet that could pass messages when no admins were around. And now 35 years later that's what we need: an Internet that works to connect users even when every ISP and "content provider" has a compelling business interest in preventing users from directly speaking to each other.
We've returned to the world of 1981. The AT&T monopoly is still in force, only now it's the consumer ISP oligopoly. Today's ISPs oversell their capacity just as the Baby Bells did their voice lines. Hayes has just released their first Smartmodem, only we call it darknet today (I2P, TOR, etc). Today's SMTP email is 1981's mail box; today's HTML web browser is 1981's color television.
Just as 1981's BBS technology connected fringe elements of society into their own distinct sub-culture that came to consider the mainstream "just noise", so will tomorrow's darknet users be able to develop their own culture out of the noise that is the public Internet. We'll still have Google and Comcast and Verizon and everyone else, but the darknets of tomorrow will be advertisement free, censorship free, and wiretap free. Users won't mind paying a small price in network latency to gain those freedoms, just as they didn't mind spending days if necessary to download just a few treasured minutes of music and video that would never be played over the mainstream radio and television.
It's coming. They know it and we know it. We can all feel it in the air. Eben Moglin has written a call to arms. Projects are popping out of the woodwork with the stated goal of taking back our Internet. Daniel Suarez's awesome book "Daemon" which features a really cool darknet is shelved in the fiction section rather than sci-fi.
ps. I don't code, I'm a systems analyst.
Surprise surprise.
So if you remove the WWW, the internet basically drops dead.
Email, IM, BitTorrent, VOIP, networked games, and mobile phone apps are already using more bandwidth than the WWW.
What are you gonna do ? Use Gopher to find files to FTP ?
I'm personally looking forward to a Daemon darknet, probably built out of I2P or TOR.
So important that these days "Internet" and "WWW" are used interchangably.
Only by the people who don't know what either of them really are.
I don't believe your 73rd percentile stat at all given that the mean and median (so 50th percentile) are roughly $45k.
You can get those stats from irs.gov. Back in 2003 I checked it out: I was making $75K and that was the 80th percentile; at the time $35K was the median. At the top 5% range it gets absolutely ridiculous - see the L Curve (biased but still accurate numbers).
That's the major cause of the current recession - the fact that people didn't save and used credit cards to spend way more than they earned (at least you're not doing that).
The problem wasn't people spending - spending is great for the economy. The problem is that the economy simultaneously requires a large middle class to keep spending and a smaller workforce to keep producing. Since the workforce and consumer class are the same people, eventually layoffs must have an effect. For the last 30 years we've been aggressively pursuing greater efficiency (woo!) but have not been able to keep up with demand for new jobs.
Ultimately the problem is structural: we are moving towards an asymptote of one self-healing machine, owned by one person, that creates all of the goods of the world. Think of a single Star Trek replicator operated by one Ferrengi. The IRS has only one W-2 to audit, and the Ferrengi insists that the income tax is drastically unfair as he is the only one paying any of it.
and all catholics are child molesters.
You meant to say "all catholic priests are child molesters."
Our purpose is to collect and report notable knowledge
Because no "true encyclopedia" would have articles on things that aren't notable, it stands to reason that anything that Wikipedia defines as notable must be so, because Wikipedia is a true encyclopedia.
Some people might say that their workplace requires them to use Microsoft products, but a career is a choice.
"Some people might say that their grocery stores force them to eat FDA-approved items, but buying food is a choice."
"Some people might say that their wiring forces them to use 110V AC appliances, but having electricity is a choice."
"Some people might say that their medical condition forces them to go to the pharmacy, but buying drugs is a choice."
If you've been here 30 years then you'll know that the various Eternal Septembers were unavoidable. GUI was going to come to wipe out the more efficient text interfaces; personal computers would have to climb the long slog to mainframe-like architecture one baby step at a time; the network effect was destined to come along and wipe out whole sectors of competition in word processors, spreadsheets, operating systems, and network protocols.
Essentially everything not made by Microsoft was better in a technical sense, but for every user willing to spend ten minutes to learn how their software worked there were a hundred users who just wanted to click on the first thing they saw and then complain to the help desk when they had no clue what was going on.
Microsoft raked in the cash, but it was the users in the end who were to blame.
"Each affected program will have to be patched separately."
And this is why Linux package managers that know how to handle shared library dependencies are better than one-click installers that bring along their own versions of the libraries.
Perhaps we could model it after the old BBS and UUCP-style networks: very very fast local meshes connected to each other over slower VPN links carried over the existing Internet. Make P2P applications prefer the local mesh and put in some large caching servers ala Akamai. Eventually people design little solar-powered buckets with 2-5W autoconfiguring repeater nodes for about $100 and you can spend an afternoon covering a 30-mile stretch off the local interstate. After enough go up people start noticing that things in the "freedom net" come down a whole lot faster than the public internet. (But then the problem become how does one anonymously *purchase* those devices. Because I could see lawmakers moving to regulate their sale ala Sudafed because of kiddie porn/terrorists/whatever.)
The harder part I think is the client-side technology: making browsers, IM clients, etc. aware that there are two Internets that must be treated distinctly and somehow expose that to the user. And making them automatically enforce privacy setting like stripping out EXIF data on every jpg picture upload and using encrypted email by default with gigantic warnings when the recipient's key isn't in the keyring or PKI.
But the possibilities...oh the possibilities. Imagine a "freedom net" without an IANA that doles out IPv6 address space only to those who can pay; dynamic DNS that costs nothing and can never be traced to a real name; the ability to have a real anonymous handle that stays that way; a new Usenet that isn't dripping in spam; end-to-end encryption both on the wire and in the application layer; no more "trusted CAs" to enable MITM for various governments; an Internet with hundreds of available routes between any two hosts, and that won't go nuts because of bad routing records. Basically, what the Internet was supposed to be before "routing around damage" was optimized away in the name of commercial efficiency.
You can SSH into it decently fast, but the lack of a hardware math unit adds around 5-10 seconds of delay when sshing from it to another computer.
I have never experienced this - my Sheevaplug (1st gen maybe?) has 1/2 second or less response time ssh'ing outbound (similar to all of my other desktops). Is it really SSH or maybe DNS that is the issue?
like one which is going to go get pelted around in the ocean
Interesting. I worked on the Texas Automated Buoy System when we moved from custom boards to PC/104-based Linux systems. Are you allowed to say more about this project?
Let's say we've all got "freedom boxes" at home. (And I've already got a Sheevaplug, it's very handy.)
So: I want to access the "freedom network" from my laptop, and I see my encrypted access point and say "use this". I use my Diaspora social networking stuff, my tweet-clone, whatever, and now I want to do a Google search: does my traffic to Google traverse the "freedom network" ala tor? If so, how does Google not remember who I am from a prior session cookie when I was on the public Net? What about performance? What about hostile trusted CAs like Etilsat and CNNIC?
But suppose all of that is solved, how am I reachable from the rest of the public Net? Do we all use one dyndns service that can be brought down ala Napster? If we don't, how do we make a distributed dyndns that is itself not subject to MITM or traffic analysis?
But suppose all of that is solved, how do we prevent one leak on the public Net from linking our "freedom network" identity(ies) to our "real life" Facebook/LinkedIn/Twitter/etc. identity?
I don't think the governments will be caught flat-footed this time around. I can't name a single government or corporate entity that would want widespread adoption of these freedom boxes. It is one of the few things that would unite such enemies as the USA and China, Israel and Iran, Apple and Microsoft.