One part of engineering is applying science. Much more of engineering is predicting problems and designing ways to avoid or mitigate them. Very often, these problems are outside the scope of the field's science.
For example, one hard part of software engineering is writing good requirements. The customer (or end user, or whoever is generating the requirements) seldom has a very good idea about what they need until they have something in hand that mostly works, and it's even rarer for them to have a good understanding of what is practical. What is practical boils often down to in-house expertise, reusable work products and development budgets rather than what the science allows. Once you know how what is needed and practical, the next challenge is to write a requirement that captures that while also being verifiable. Most systems that people care about cannot (practically) be formally proven -- they are too complicated to be affordable, techniques for doing so are unknown, and so forth; instead, the project's engineers need to have good judgment about what kind of design verification are appropriate and how much is needed on which components in order to meet the project's goals.
Congratulations on your 20-line programs written in languages that hardly anyone cares about.
(I kid a little: People have developed formal proofs of separation kernels that are hundreds of lines long -- and published because doing so requires fairly novel techniques. And people sometimes do formal proofs of Java source code, which some people care about. But most programs that get formal proofs are tiny -- or tiny parts of larger, more complicated systems -- and an awful lot of formal provers only work on niche languages.)
Your opinion is idiotic and wrong; the title "software engineer" omits the key qualifiers "consulting", "professional" and "-in-training" (or abbreviations thereof). The law is refers to abbreviations such as "P.E." for professional engineer. Someone calling himself a "software consulting engineer" would probably run afoul of the law, and someone using the title "consulting software engineer" would be on very questionable ground, but the title "software engineer" is not covered by the law.
Legal restrictions on "Engineer" as a job title are in some ways like laws requiring interior designers or casket salesmen to be accredited by industry groups -- motivated by a desire to squelch competition. On the other hand, most engineering accreditation groups also put a strong emphasis on ethics: For example, someone who is a member of one of those groups has an obligation to keep the public interest in mind as they design objects or processes, and to push back (or report to the authorities) efforts to cut corners in ways that harm safety. There is also usually an effort to provide workplace mentoring so that new graduates are paired with someone who has more experience with the more practical (and less schoolbook-related) challenges of the work.
As a software developer (and sometimes engineer) who has worked on systems that involve mechanical, electrical, software and systems engineering, I try to restrict "software engineering" to projects that involve careful and informed analysis of trade-offs to meet specified goals, planning and managing the development cycle (from requirements analysis and derivation through programming, integration, testing against the requirements and testing against the user's need, and delivery), and then carrying through with something that more or less resembles the plan. If a project is less rigorous, I prefer to call it just software development.
In my experience, the usual areas where a project fails to qualify as "software engineering" are the analysis (most common), testing (sadly common) and requirements (less common). That isn't necessarily a bad thing -- for example, time to market or budget might be more important than delivering a well-understood product with known high quality. On the other hand, a project with a foreseeable impact on people's health or safety is a project that needs analysis, planning and testing that are proportionate to the risks. Because they do not directly help build a deliverable product, analysis, requirements and testing seldom get the attention or credit they deserve.
There are a number of good reasons to believe that email voting is less secure than snail mail. Among them: it is easier to change, forge or destroy electronic records than physical ones; there fewer legal protections for email than postal mail; and there is much less experience with email voting, so mistakes are easier to make and fraud is easier to commit.
When I see a 2.6 MB kernel module (for fglrx; I recall Nvidia's being closer to 9 MB), my assumption is that it is mostly code that executes on the host. Code or data that gets loaded onto the peripheral can easily live in a file, or in a executable section that gets discarded after use. Given how specialized the devices are, how much translation and optimization is needed to convert application-level APIs to hardware operations, and the fact that the vendors have begged off on providing host-side source code because of NDAs with third parties, it will take a lot of evidence to convince me that the size of the binary blobs is due to peripheral code rather than (intentionally proprietary) host code.
The BTS's cost estimate for 2002 was still significantly lower than any of the public transportation rates cited for that year, and that's a pessimistic number because there is often more than one person in a car. I didn't cherry-pick cost per passenger-mile; it is the right number to look at to analyze your (frankly idiotic) claim that private-vehicle commuters would have gotten more value by spending money on public transportation than on their cars.
On top of the fact that the average cost to drive a car is already cheaper than every form of public transit, the cost per passenger-mile for existing public routes is lower than what you would expect to see for new routes: existing routes generally serve the highest densities of residences and workplaces, so they have more people per vehicle than new public transportation routes would. In that respect, you are right that transportation costs do not scale linearly. Reaching more people with public transportation increases per capita costs, making it even less cost-competitive with privately owned automobiles.
Public transportation works better in high-density urban areas than in most the US. If you look outside the cities of Europe and Japan, for example, you'll find that most households own and use cars -- and for the same reason that so many Americans drive themselves to work. As long as the US has suburbs, it will have a lot of people for whom driving a car is the best way to get around.
You could continue to pull out statistics, yes, but you should look and think a bit about what you already found.
For example, the cost per passenger-mile -- in 2002 -- listed in the VTPI white paper (your link [8]) ranged from $0.57 to $2.78 for public transit. Even after a decade of particularly steep fuel-cost increases, the 2012 GSA reimbursement rate for privately owned vehicles is still less than that: $0.555 per mile. (The GSA reimbursement rate is intended to cover wear and tear, gas, and maintenance costs for a typical car, so I would say it is a close equivalent to cost per passenger-mile of public transit. It was $0.365/mile in 2002.) If public transit spends much less money than private vehicle owners, it is primarily because it carries an even smaller fraction of the total traffic!
It is probably just as well that you did not delve into the costs of road maintenance: Small cars, even in large numbers, do not contribute all that much to annual maintenance needs; weather damage (e.g. potholes that develop over winter) and truck traffic are bigger drivers of those costs.
If your numbers are accurate and the factors are uncorrelated, about 2 fatal crashes per thousand would be attributable to an inebriated illegal-immigrant teen texting while speeding on a poorly maintained road. Combining that with NHTSA's count of just over 30,000 fatal crashes in 2010... would you consider 60 such crashes per year "a lot"?
[And yes, before some humor-impaired commenter gets on my case, I know the numbers don't work like that: "alcohol involved" crashes include cases where a drunk driver who was otherwise driving properly gets T-boned by someone else; the numbers do not reflect plausible causes, just presence.]
Most cars on the road today cost $20,000 new (+/- 50%). How much do you think a tenth of that would help public transit? It might be enough to employ a driver for a month -- but don't forget the mechanics and managers that you seldom see, or the depots where buses or trains are parked and serviced, or (pshaw) the capital costs of the vehicle.
Leasing a car -- which means you'd get a brand new one every three years or so -- typically runs under $300/month. Even if one adds in fuel and maintenance, and compares the full cost of that to the unsubsidized per capita cost of public transit, almost all of the US works out to be cheaper for car than bus or train (or whatever else).
If your idea of reality is that public transit in America is more cost-effective than privately owned cars (even without addressing the convenience aspects), no wonder you don't understand libertarianism... or reality.
The efficiency of mass transit goes up at least linearly with population density. In the US, only some large city routes reach break-even for mass transit versus individual transit, and in most of those one pays a cost in transit time in order to realize the relative gains for other resources. (Side note: Many of those routes depend on subsidies to gain enough riders for break-even, and those cities' transit systems tend to have a lot of other routes that don't break even.)
It is implausible because of the scales being presented as representative. Wal-Mart's aggressive sourcing and price management strategies don't make the difference between spending $100 at Wal-Mart and spending $150 anywhere else. Because the argument continues that real customers wouldn't give up such big savings in order to participate in a boycott, the argument is fundamentally unrealistic. More often, the relative price difference is on the order of five or ten percent, and many people have in fact given up that kind of savings in order to boycott a merchant. Which, funnily enough, is the exact behavior that the page claims never happens.
It is precisely because you (quite intentionally) stuck your poem in a place where automated systems are used to copying stuff that it makes it okay for automated systems to copy it. You put it there with the full expectation and intent that they would copy it, which is why I think courts would treat it as an implicit license to copy and redistribute.
A court would not find your "experiment" either very interesting or very amusing. To the extent that you told your computer to present your poem as the browser's identification and configuration, I expect it would be interpreted as a nearly unlimited license to do all the things that web servers normally do with browser agent strings.
Only federal courts have subject matter jurisdiction over copyright claims, so no small claims (or other state) court will hear such a case. Federal courts also have a $75,000 threshold for jurisdiction -- i.e. you can't bring case in federal court after only $74,999.99 in damages -- but copyright law's statutory minimum damages tend to make that fairly easy to meet (as we famously saw in the Jammie Thomas case). Unfortunately, it's relatively hard to succeed in federal courts with pro se representation.
You visiting Slashdot is not your property; it is a fact, and is not treated as anyone's property under the law. Non-trivial documents are often eligible for protection under copyright law, but web-browsing history generally is not. To my knowledge, the only theory that might apply to your web-browsing history at a particular site is some kind of contract between you and the site operator, but it's hard for me to think of why a DNT header would establish or modify such a contract, and it's equally hard for me to think of why a court would think violation of DNT requests would be worthy of much in damages (even in class-action form, assuming the relevant laws would authorize class status in that kind of case).
No. The HTTP request your computer sends them is data that your computer stored at their request (cookies), data that is necessary to identify what you are asking for, or other data that are not protected by copyright. What the server sends back -- usually, but not always -- contains enough creative expression to be protected by copyright. Except in cases of people trying to pull stunts (as in another response to one of my comments), that will essentially always be true of the request headers that are at issue with DNT.
Copyright does not protect "identifiable data". It protects creative expression, in order "To promote the Progress of Science and useful Arts".
I would be curious to know how a court would treat someone who made an argument as frivolous as "they infringed copyright in my custom User-Agent string". I suspect most courts would settle for relatively minor sanctions, like throwing the case out and making the plaintiff pay the defense's costs. If you try it, please keep us posted!
Maybe in intent, but no one has given a reason that (US) courts should treat DNT as binding. If you stick a "do not enter" sign on someone else's door, it has no legal effect, and may even land you in trouble.
As I said, the bits you send to the web server are not your property. You seem to think that sending a Do-Not-Track header line gives you some claim about what the web server (and its owner(s) and their partners) can do with their property. That is why I made the exaggerated analogy to a Give-Me-Money header.
Exactly what theory would you (or your lawyer) use in court? That by sending an HTTP request with a "Give-Me-Money: $1000" line, and then not giving you $1000, they violated a court-enforceable contract?
The data that you send to a web server is not your property; without actual laws that limit the recording and sharing of personally identifiable data -- for example, Europe's data privacy laws -- or an actual contract between you and the web site's provider, you have no legally reasonable expectation that advertisers will not get, keep, massage, and exploit anything they can determine about who is viewing the web page or what else that person has looked at or done.
While this is the first traceable ballot I've heard of being used in the modern-day US, it was not always so. In the Eighteenth Century, for example, there were a variety of extreme partisan tactics: Separate ballot boxes for each candidate, or ballots being pre-printed by the major parties on different colored papers, and little or no privacy when casting a ballot, all combined with people who would (yes, literally) beat up voters who voted for the "wrong" party.
There have also been no constitutional amendments that have mentioned ballot privacy or secrecy, so I think the judge is right to say that there is no constitutional right to a secret ballot. I am a bit surprised that there are also no state or federal laws that would forbid a bar code on the ballot (assuming it gets associated in any way with the person casting the ballot, rather than just being an identifier to detect duplicated ballots).
Boycotts do not have to drive companies out of business to make businesses change their practices. The most effective boycotts are an integral part of a broader push -- efforts to get media coverage for the cause, demonstrating in front of stores or corporate offices, and so forth -- that are generally trying to change behaviors rather than put a company out of business.
As a side note, if your best anti-boycott source is someone who uses an imaginary word as a gender-neutral pronoun and an implausible scenario to argue that boycotts don't work, you might want to find a better source. (On the other hand, credible sources are not likely to comport nicely with the kind of statist philosophy that your source apparently holds, so maybe you would not think there are any "better" sources.)
Slashdot Mirror
Don't forget to include the almost equally textbook identity theft claims when filing the defamation action...
In this case, MAC = Mandatory Access Control, and the GP was right.
One part of engineering is applying science. Much more of engineering is predicting problems and designing ways to avoid or mitigate them. Very often, these problems are outside the scope of the field's science.
For example, one hard part of software engineering is writing good requirements. The customer (or end user, or whoever is generating the requirements) seldom has a very good idea about what they need until they have something in hand that mostly works, and it's even rarer for them to have a good understanding of what is practical. What is practical boils often down to in-house expertise, reusable work products and development budgets rather than what the science allows. Once you know how what is needed and practical, the next challenge is to write a requirement that captures that while also being verifiable. Most systems that people care about cannot (practically) be formally proven -- they are too complicated to be affordable, techniques for doing so are unknown, and so forth; instead, the project's engineers need to have good judgment about what kind of design verification are appropriate and how much is needed on which components in order to meet the project's goals.
Congratulations on your 20-line programs written in languages that hardly anyone cares about.
(I kid a little: People have developed formal proofs of separation kernels that are hundreds of lines long -- and published because doing so requires fairly novel techniques. And people sometimes do formal proofs of Java source code, which some people care about. But most programs that get formal proofs are tiny -- or tiny parts of larger, more complicated systems -- and an awful lot of formal provers only work on niche languages.)
Your opinion is idiotic and wrong; the title "software engineer" omits the key qualifiers "consulting", "professional" and "-in-training" (or abbreviations thereof). The law is refers to abbreviations such as "P.E." for professional engineer. Someone calling himself a "software consulting engineer" would probably run afoul of the law, and someone using the title "consulting software engineer" would be on very questionable ground, but the title "software engineer" is not covered by the law.
Legal restrictions on "Engineer" as a job title are in some ways like laws requiring interior designers or casket salesmen to be accredited by industry groups -- motivated by a desire to squelch competition. On the other hand, most engineering accreditation groups also put a strong emphasis on ethics: For example, someone who is a member of one of those groups has an obligation to keep the public interest in mind as they design objects or processes, and to push back (or report to the authorities) efforts to cut corners in ways that harm safety. There is also usually an effort to provide workplace mentoring so that new graduates are paired with someone who has more experience with the more practical (and less schoolbook-related) challenges of the work.
As a software developer (and sometimes engineer) who has worked on systems that involve mechanical, electrical, software and systems engineering, I try to restrict "software engineering" to projects that involve careful and informed analysis of trade-offs to meet specified goals, planning and managing the development cycle (from requirements analysis and derivation through programming, integration, testing against the requirements and testing against the user's need, and delivery), and then carrying through with something that more or less resembles the plan. If a project is less rigorous, I prefer to call it just software development.
In my experience, the usual areas where a project fails to qualify as "software engineering" are the analysis (most common), testing (sadly common) and requirements (less common). That isn't necessarily a bad thing -- for example, time to market or budget might be more important than delivering a well-understood product with known high quality. On the other hand, a project with a foreseeable impact on people's health or safety is a project that needs analysis, planning and testing that are proportionate to the risks. Because they do not directly help build a deliverable product, analysis, requirements and testing seldom get the attention or credit they deserve.
There are a number of good reasons to believe that email voting is less secure than snail mail. Among them: it is easier to change, forge or destroy electronic records than physical ones; there fewer legal protections for email than postal mail; and there is much less experience with email voting, so mistakes are easier to make and fraud is easier to commit.
When I see a 2.6 MB kernel module (for fglrx; I recall Nvidia's being closer to 9 MB), my assumption is that it is mostly code that executes on the host. Code or data that gets loaded onto the peripheral can easily live in a file, or in a executable section that gets discarded after use. Given how specialized the devices are, how much translation and optimization is needed to convert application-level APIs to hardware operations, and the fact that the vendors have begged off on providing host-side source code because of NDAs with third parties, it will take a lot of evidence to convince me that the size of the binary blobs is due to peripheral code rather than (intentionally proprietary) host code.
The BTS's cost estimate for 2002 was still significantly lower than any of the public transportation rates cited for that year, and that's a pessimistic number because there is often more than one person in a car. I didn't cherry-pick cost per passenger-mile; it is the right number to look at to analyze your (frankly idiotic) claim that private-vehicle commuters would have gotten more value by spending money on public transportation than on their cars.
On top of the fact that the average cost to drive a car is already cheaper than every form of public transit, the cost per passenger-mile for existing public routes is lower than what you would expect to see for new routes: existing routes generally serve the highest densities of residences and workplaces, so they have more people per vehicle than new public transportation routes would. In that respect, you are right that transportation costs do not scale linearly. Reaching more people with public transportation increases per capita costs, making it even less cost-competitive with privately owned automobiles.
Public transportation works better in high-density urban areas than in most the US. If you look outside the cities of Europe and Japan, for example, you'll find that most households own and use cars -- and for the same reason that so many Americans drive themselves to work. As long as the US has suburbs, it will have a lot of people for whom driving a car is the best way to get around.
You could continue to pull out statistics, yes, but you should look and think a bit about what you already found.
For example, the cost per passenger-mile -- in 2002 -- listed in the VTPI white paper (your link [8]) ranged from $0.57 to $2.78 for public transit. Even after a decade of particularly steep fuel-cost increases, the 2012 GSA reimbursement rate for privately owned vehicles is still less than that: $0.555 per mile. (The GSA reimbursement rate is intended to cover wear and tear, gas, and maintenance costs for a typical car, so I would say it is a close equivalent to cost per passenger-mile of public transit. It was $0.365/mile in 2002.) If public transit spends much less money than private vehicle owners, it is primarily because it carries an even smaller fraction of the total traffic!
It is probably just as well that you did not delve into the costs of road maintenance: Small cars, even in large numbers, do not contribute all that much to annual maintenance needs; weather damage (e.g. potholes that develop over winter) and truck traffic are bigger drivers of those costs.
If your numbers are accurate and the factors are uncorrelated, about 2 fatal crashes per thousand would be attributable to an inebriated illegal-immigrant teen texting while speeding on a poorly maintained road. Combining that with NHTSA's count of just over 30,000 fatal crashes in 2010... would you consider 60 such crashes per year "a lot"?
[And yes, before some humor-impaired commenter gets on my case, I know the numbers don't work like that: "alcohol involved" crashes include cases where a drunk driver who was otherwise driving properly gets T-boned by someone else; the numbers do not reflect plausible causes, just presence.]
Most cars on the road today cost $20,000 new (+/- 50%). How much do you think a tenth of that would help public transit? It might be enough to employ a driver for a month -- but don't forget the mechanics and managers that you seldom see, or the depots where buses or trains are parked and serviced, or (pshaw) the capital costs of the vehicle.
Leasing a car -- which means you'd get a brand new one every three years or so -- typically runs under $300/month. Even if one adds in fuel and maintenance, and compares the full cost of that to the unsubsidized per capita cost of public transit, almost all of the US works out to be cheaper for car than bus or train (or whatever else).
If your idea of reality is that public transit in America is more cost-effective than privately owned cars (even without addressing the convenience aspects), no wonder you don't understand libertarianism... or reality.
The efficiency of mass transit goes up at least linearly with population density. In the US, only some large city routes reach break-even for mass transit versus individual transit, and in most of those one pays a cost in transit time in order to realize the relative gains for other resources. (Side note: Many of those routes depend on subsidies to gain enough riders for break-even, and those cities' transit systems tend to have a lot of other routes that don't break even.)
It is implausible because of the scales being presented as representative. Wal-Mart's aggressive sourcing and price management strategies don't make the difference between spending $100 at Wal-Mart and spending $150 anywhere else. Because the argument continues that real customers wouldn't give up such big savings in order to participate in a boycott, the argument is fundamentally unrealistic. More often, the relative price difference is on the order of five or ten percent, and many people have in fact given up that kind of savings in order to boycott a merchant. Which, funnily enough, is the exact behavior that the page claims never happens.
It is precisely because you (quite intentionally) stuck your poem in a place where automated systems are used to copying stuff that it makes it okay for automated systems to copy it. You put it there with the full expectation and intent that they would copy it, which is why I think courts would treat it as an implicit license to copy and redistribute.
A court would not find your "experiment" either very interesting or very amusing. To the extent that you told your computer to present your poem as the browser's identification and configuration, I expect it would be interpreted as a nearly unlimited license to do all the things that web servers normally do with browser agent strings.
Only federal courts have subject matter jurisdiction over copyright claims, so no small claims (or other state) court will hear such a case. Federal courts also have a $75,000 threshold for jurisdiction -- i.e. you can't bring case in federal court after only $74,999.99 in damages -- but copyright law's statutory minimum damages tend to make that fairly easy to meet (as we famously saw in the Jammie Thomas case). Unfortunately, it's relatively hard to succeed in federal courts with pro se representation.
You visiting Slashdot is not your property; it is a fact, and is not treated as anyone's property under the law. Non-trivial documents are often eligible for protection under copyright law, but web-browsing history generally is not. To my knowledge, the only theory that might apply to your web-browsing history at a particular site is some kind of contract between you and the site operator, but it's hard for me to think of why a DNT header would establish or modify such a contract, and it's equally hard for me to think of why a court would think violation of DNT requests would be worthy of much in damages (even in class-action form, assuming the relevant laws would authorize class status in that kind of case).
No. The HTTP request your computer sends them is data that your computer stored at their request (cookies), data that is necessary to identify what you are asking for, or other data that are not protected by copyright. What the server sends back -- usually, but not always -- contains enough creative expression to be protected by copyright. Except in cases of people trying to pull stunts (as in another response to one of my comments), that will essentially always be true of the request headers that are at issue with DNT.
Copyright does not protect "identifiable data". It protects creative expression, in order "To promote the Progress of Science and useful Arts".
I would be curious to know how a court would treat someone who made an argument as frivolous as "they infringed copyright in my custom User-Agent string". I suspect most courts would settle for relatively minor sanctions, like throwing the case out and making the plaintiff pay the defense's costs. If you try it, please keep us posted!
Maybe in intent, but no one has given a reason that (US) courts should treat DNT as binding. If you stick a "do not enter" sign on someone else's door, it has no legal effect, and may even land you in trouble.
As I said, the bits you send to the web server are not your property. You seem to think that sending a Do-Not-Track header line gives you some claim about what the web server (and its owner(s) and their partners) can do with their property. That is why I made the exaggerated analogy to a Give-Me-Money header.
Exactly what theory would you (or your lawyer) use in court? That by sending an HTTP request with a "Give-Me-Money: $1000" line, and then not giving you $1000, they violated a court-enforceable contract?
The data that you send to a web server is not your property; without actual laws that limit the recording and sharing of personally identifiable data -- for example, Europe's data privacy laws -- or an actual contract between you and the web site's provider, you have no legally reasonable expectation that advertisers will not get, keep, massage, and exploit anything they can determine about who is viewing the web page or what else that person has looked at or done.
While this is the first traceable ballot I've heard of being used in the modern-day US, it was not always so. In the Eighteenth Century, for example, there were a variety of extreme partisan tactics: Separate ballot boxes for each candidate, or ballots being pre-printed by the major parties on different colored papers, and little or no privacy when casting a ballot, all combined with people who would (yes, literally) beat up voters who voted for the "wrong" party.
There have also been no constitutional amendments that have mentioned ballot privacy or secrecy, so I think the judge is right to say that there is no constitutional right to a secret ballot. I am a bit surprised that there are also no state or federal laws that would forbid a bar code on the ballot (assuming it gets associated in any way with the person casting the ballot, rather than just being an identifier to detect duplicated ballots).
Boycotts do not have to drive companies out of business to make businesses change their practices. The most effective boycotts are an integral part of a broader push -- efforts to get media coverage for the cause, demonstrating in front of stores or corporate offices, and so forth -- that are generally trying to change behaviors rather than put a company out of business.
As a side note, if your best anti-boycott source is someone who uses an imaginary word as a gender-neutral pronoun and an implausible scenario to argue that boycotts don't work, you might want to find a better source. (On the other hand, credible sources are not likely to comport nicely with the kind of statist philosophy that your source apparently holds, so maybe you would not think there are any "better" sources.)