I think you should know, AShly simpson is an 'entertainer' not a musician. No, I don't personaly find her entertaining. PIF, most people don't. I used to think I was just out of touch with the music scene. I've been talking to be in the industry recently and it turns out a lot of these names everyone heres about don't sell many tickets. Most people in one of there concerts are people in the industry that are there to be seen.
My wife and I have been talking about this quite a bit recently. We've been watching "American Idol". Anwar Robinson is clearly the most musically talented person to ever be on that show, but the stuff he does is not what the record companies want to market. This past Monday he got up and sang Louis Armstrong's "Wonderful World". He started out singing like Louis Armstrong, then series of runs as he moved the style into something more reminicent of Sammy Davis Jr., and finally ended the song in a soul style. His talent should win, but he won't fit into the marketing machine of the record companies.
However, in Econ 301 they learned that running a cartel to fix prices is the best system of all, so that's what they did. Supply and demand have nothing to do with the record industry's prices.
Except that cartels only work 1) when comsumers can't "do without" and 2) when there is no alternative good that can satisfy their needs.
If music sales are truly suffering, something of which I'm not convinced, it could be due to either of those factors or a combination. Piracy provides an alternate source. But consumers are also capable of just not buying.
By reducing prices, they can reduce the effect of piracy, and reduce the numbers of potential consumers who choose not to consume. This can have the net effect of increasing their revenue and profit.
Sooner or later, if you like it or not, the US will be metric country as well.
No we won't. It's not that there's anything wrong with metric, it's just that Carter tried to do it. If a useful president had tried, we'd be done by now.
Actually, we're already a metric country in any way that matters. Does miles per hour on a highway sign or measuring a recipe in cups and tablespoons really matter all that much. HELL NO!. What matters is machine tools, and they've been converting over for quite a while now. What matters is that a transmission manufactured in Detroit can be mated to an engine block manufactured in Japan.
That still doesn't explain why fahrenheit can't be used for research. What is there to prevent me from doing a water heating experiment, and making all measurements in fahrenheit, or for that matter, in any scale I can imagine.
Actually, a scientist who prefered to use Fahrenheit measurements would probably use the Rankine scale. Zero Rankine is, like 0 Kelvin, absolute zero. However, the size of a degree is the same as the Fahrenhiet scale. The freezing point of water (32F) and the boiling point of water (212F) correspond to 491.67R and 671.67R, respectively.
It is all in the intent. In blocking port 25, Verizon is not attempting to use their strength to block competing services. They are trying to block spam.
It is pretty clear in this case that the telco was attempting to hamstring a competing product.
Now, if Verizon was charging a per-email fee to use their own mail servers and blocking access to external mail servers, the situation may be more comparable.
Only governments censor. This would be anti-competitive. Semantics, yes, but an important distiction nonetheless.
The situation is not so cut and dried. As a practical matter, the ISPs in the United States that can effectively carry VoIP traffic also hold their competitive positions as the result of government granted monopolies and franchises: the telephone system and cable tv system. A business that has received such a level of competitive protective from the government generally can't simultaneously be a completely private business.
If, by slap on the wrist, you mean life in prison. It's not like we'll knock it down to probation os anything.
Actually you raise an interesting issue. The SCotUS decided that, for the purposes of applying the death penalty, a person who committed their crime at the age of 17 years and 364 days must be in law treated differently than a person who was 18 years of age at the time of the crime. The trend of the law before this has been to acknowledge a grey area between youth and adulthood, and that defendents under the age of 18 would be evaluated by the court on several factors before determining if the defendent would be treated as an adult or a juvenile.
If we accept the SC's decision that a minor one day short of his eighteenth birthday should not be treated as an adult for the purpose of administering a capital sentence, why should that minor still be eligible for another severe sentence. Is Life without Possibility of Parole any less cruel and unusual?
If I was a defense attorney for such a client, I'd be running as fast as I can to file an appeal for my life sentenced minor.
Re:Neither necessary nor sufficient
on
QA != Testing
·
· Score: 1
1) deciding what you want to build, and deciding exactly (i.e., good specs);
Too often, I've stumbled across over-specified systems that, as a result, are delivered incredibly late. And then, because of time constraints, the whole project is de-scoped and bodged work-arounds are built so that functionality can be 'added later'.
At the design stage, politics often slows things down. I prefer the continuous approach: When you have enough design, start coding.
An effective team is going to wind up doing the same tasks regardless of the methodology chosen. For instance:
A team practicing XP is going to jump into coding very soon. Then they'll refactor several times and toss away code that just didn't fit the problem. All the time they are showing the customer and getting additional feedback.
Meanwhile, a team practicing a waterfall methodology is going to build several prototypes during the requirements gathering and architecture stages. They'll do several experiments to insure that they understand key parts of the design. Then they'll agree on a spec and an architecture and finally start coding.
The difference is semantic. I can say I'm doing a prototype or experimenting with a key feature, then later use key parts or my results to build the actual product. Or I can say I'm working on the actual product and throw away substantial parts of my first iteration.
But what about a web app? If a popular website is using Oracle on the backend, is that several million users? Or is it one user, because only one webserver is accessing it? Or maybe it is the size of the connection pool?
You're right, but there are additional considerations...
First, the major software vendors would love to move the market to a subscription service model, but it may be a better model for the consumer as well.
Take Microsoft, for instance. When MS sold a copy of Office 95, they made their money once. If there was no compelling reason to upgrade, then they could never make another dime selling Office to that customer. So MS's self interest is to get the Office 95 customer to buy Office 98, 2000, XP, 2003,... To do that they need to 1) add features, 2) change file formats to artificially reduce interoperability, 3) make sure that the new version didn't have some bugs that the older version had. Note that MS has little motivation to fix bugs in versions it has already sold.
In a subscription world, MS gets a revenue continuously from their customers. There is no longer a motive for MS to resell to the same customer, so their is no motive to add features of dubious value, alter file formats, or leave bugs in older versions. MS doesn't care if the customer upgrades to the latest version, so MS is more than happy to fix bugs in older versions. In the end, customers get software that is more stable and secure.
My other point is that thin clients can drasticly reduce the operating costs of an enterprise full of desktops. Back in the Windows 95 days, I actually was part of an eval to consider replacing all the desktops in a 400 person enterprise with Citrix systems. Any user could work on "their" pc from any desk in the company. The old Win3.1 boxes floating around could run the client just fine, and wouldn't need to be replaced. Server administration, backups, everything could be centralized. The technology was quite ready for prime time, but I do some consulting for a major wall street company that is today using Citrix over the web instead of a normal VPN. I log into the web site and I'm connected via Citrix to a basic machine with my Email, Instant Messaging, Word, Excel, and other basic apps right there. If needed I can do a WinXP remote desktop session to my own desktop.
For real fun, imagine a company that decides to drink the Linux-on-Desktop kool-aid. Start with an IBM mainframe. Partition it into several thousand individual Linux instances, one per employee. Give each employee a system that runs X - or maybe even a modified VNC....
Now if a system needs to be deployed for a new employee: run one Perl script and it is done is 5 seconds. If someone needs more ram or disk quota: 2 more scripts. A patch needs to be pushed to every system: no problem. Backups: easy. Disaster recovery: trivial.
If buffer overflows are getting past your unit tests, it's because you're not writing proper unit tests.
Unit tests? I've never heard of that before. Can you tell me more?
Seriously, though, that is an excellent point. In the real world, large C codebases are going to have security holes because some part of the system, be it the C code itself or the libraries, was not sufficiently unit tested. In a Java environment, insufficiently tested code is likely to abend but is less likely to open a security hole, where in a C environment insufficiently tested code is a security risk.
But these things all require a robot to consider in advance the likely outcome of several courses of action. It means that the robot must not only model its own behaviour but also model the behaviour of other objects and people in the area.
For instance, a robot might be intelligent enough to not injure a person directly, but if you take that robot to the top of a tall building and tell it to throw bricks, it is another step for that robot to determine that 1) a brick falling from a high place might injure a person, and 2) there may be people on the street below.
To make it even more complex, the robot is still on a roof top. A guest of wind knocks a person off balance, and they are in danger of falling. The robot has to consider 1) the likelihood of the person regaining their balance without intervention, 2) the likelihood that grabbing the person with a robot manipulator will injure the person, and 3) the likelihood that the person will fall from the roof to their death. Now, the robot needs to make a judgement call, to grab and potentially break an arm in the process, or to not grab and possibly see someone plunge to their death. Still a tougher judgement to make.
It is possible to build these models, but to use AI in order to build these models in an effective way is a stunning leap forward from our current state.
I don't understand why the record companies dropped the suit. Did they suddenly grow a brain that allowed them to correlate between the defendant's being dead and therefore their inability to sue her?
Actually, your the one that needs to grow a brain. They can certainly sue a dead defendant. It happens all the time. The estate of the defendant is liable for any judgement.
It works in the opposite way too. A dead person can be a plaintiff. If you were to die due to, perhaps, medical malpractice, you (through your own estate) would sue for wrongful death. Your spouse or other next of kin don't have standing to sue for wrongful death, only you do, and so your estate must make the claim.
In this case they chose to drop the suit because of the bad PR. They are probably investigating right now to determine if a family member was using grandma's name to buy internet access. They'll sue that family member once they get the evidence.
To be honest, most business programming is glue code and does not require any "top 5% developers" for success. What it does require is flexible people who can digest busines requirements and do estimates and planning correctly, and usually that's an entirely different skillset than the guys who can churn out 3000 lines of bug-free C++ every day. (Worked with a lot of really smart people that operated on a "it's done when it's done" mode which generally translates into project failure.)
I didn't say "top 5% programmer". Developer is much more encompassing. One of the skills of a top developer is requirements discovery, even requirements anticipation. Not to mention planning and estimating. Even more important, a top developer understands the architecture underlying the entire project, and therefore can write code that not only adheres to the specs, but also integrates well with other components.
Large innovative projects are frequently doomed to fail. One major problem is that it is very difficult to staff a project properly. It takes very bright developers to build innovative things. It takes a lot of very bright developers to build very large innovative things. The odds of being able to collect the kind of talent needed are often very slim.
The top 5% developers are almost never found on the unemployment line, so you can't hire them from there. They are almost never working for body shops, so you can't find them there. Generally they are 1) comfortably employed somewhere else, and not interested in jumping, 2) working for a commercial software house and not interested in working for a company where they are "overhead" and not the center of the business, or 3) they are self employed and making much bigger bucks freelancing than they could be paid within a corporate job.
Most companies have a few of these highly qualified people. The best way to insure successful projects is to make sure that there are only as many innovative projects as can be staffed by your innovative people. A large business changing project is out of reach. But a several small to midsize projects can also transform the business, and can be executed by the staff of innovators already on hand.
It detects a ground fault, e.g. a short between the power and the ground in the plug. However if you were to touch a live wire and the ground is through you and down into the real ground it would NOT trip the GFCI. (Disclaimer: IANAE)
Actually, that is quite wrong. Nearly all electrical power systems we encounter in our daily lives carry current in a two wire circuit. If everything is operating properly, the current in one wire will exactly equal the current flowing in the opposite direction in the other wire. The GFCI device constantly measures these two currents and compares them.
If, however, any of that current finds an alternate path - perhaps through a hair dryer that has been dropped into a bath tub - the current in each wire will be different. This imbalance causes the GFCI device to trip. Note that if the current flows from one power conductor, through you, and back into the other power conductor, you will still be electrocuted.
[W]hen I contribute open source code, I want it to remain open source -- not benefiting some company who isn't even going to give anything back.
Even if a company makes a proprietary derivative work, they are very likely to contribute back to the Open Source code base. Maintaining a fork indefinately is expensive. It is much better to maintain the derivative work as a limited set of additions/revisions to the publicly available base work.
However the users of the proprietary work expect support. So the proprietary maintainer is constantly doing bug fixes in areas of the code outside the proprietary modules. It is better that these fixes be contributed back to the base project, rather than maintain an ever growing set of patches for the proprietary product. There own selfish motives can therefore improve the base work.
I'm glad we have both the GPL and BSD licenses, but in my opinion GPL is "more free" because it ensures that code contributions remain open source. With either license, a company/individual may use open source code for their own gains. (and there's nothing wrong with that)
However, the GPL simply ensures that any modifications to open source code are themselves available as open source. The BSD license allows propriatary forks of open source code, which may be good from a corporate standpoint, but may be considered detrimental to the open source community at large.
It is just as easy to argue the opposite. The flip side is that BSD is more free because you can do whatever you want with derivative works, whereas the GPL is restrictive because because you can only release a derivative work under the GPL.
As a practical matter, a software vendor may have some piece of work that they want or need to keep proprietary. They cannot distribute that code integrated into a GPL package. However they can distribute that code as part of a BSD licensed package.
While in theory this doesn't seem to offer an advantage, in practise the developer community for the base package is larger. These developers inevitably contribute large amounts of work back to the base project.
I could be wrong, but IMO, they aren't the same licenses. The GPL and BSD licenses differ quite a bit.
You are correct that the GPL and BSD licenses are very different. How they differ is not a factor, however, for the 99.99% of the users who will use this software without ever modifying or redistributing the source code.
The GPL is in some ways a vestige of a time when computer users were largely computer programmers. Now, most people will never even look at a line a code. For them both Open Source and Free Software mean high quality software without BSA audits, burdensome license agreements, and closed standards. There is no lock in, so software vendors must compete on value.
This is painful just thinking about. We think of Microsoft products as being pretty lousy, but the truth is that they have some very good programmers. Why there stuff sucks, I'm really not sure. Maybe a cultural thing. Anyhow, any company with the culture that even CONSIDERS bidding for programmers has to have a pretty stupid culture -- just imagine what kinds of horrible stuff would come out of such a company. I shudder at the thought.
I believe that Microsoft's coders are toward the elite of the industry. Microsoft's quality problem is primarily this...
Given the choice between coding a new feature and improving the reliability of the existing feature set, Microsft marketing and management have always chosen the new feature.
Code can be delivered quickly, can be designed innovatively, and can be bug free. Code can never have all three at once.
For those in the Technology sector, lets all collectively agree right now to not allow this to become a common practice for IT jobs. If quality people do not participate, then the companies that host this kind of charade will get the unskilled, questionable people that they deserve.
Wake up! What do you think has been going on in the IT sector?
During the boom, how many people jumped from job to job every 6-9 months because someone else waved ten or twenty or fifty thousand dollars more at them? Was it ok then? Suddenly it isn't ok?
Our work has no intrinsic value. Our work is valued at exactly how much we can convince someone to pay us. No more and no less.
Someone that has a proven track record as a successful architect is going to get $150,000 to $300,000 a year, at least. Someone that has read HTML For Idiots is going to work at Denny's bussing tables. Anyone that expects to get an architect is going to have to pay the going rates for architects. It doesn't matter how many HTML jockeys are working at Denny's, it doesn't affect the market for architects.
I know. I'm trying to hire and I can't find good people at any price. Playing with MS Access does not qualify you to design TerraByte size DB2 or Oracle systems. Designing a web site for you cousin's deli does not prepare you to design a major J2EE infrastructure for a Fortune 500 company.
Re:Maybe they could advertise this at the hospital
on
Would You Bid for a Job?
·
· Score: 2, Insightful
If you were going to stay a hospital for a few days for a surgery or illness, Would you rather have a nurse that values her skills at $10/hour or one that thinks she is worth $50/hour. Also, a nurse that works for less will put in longer hours to maintain the same standard of living. She is more likely to be tired and overworked.
Would you rather be cared for by a nurse who was considered good enough to have a steady job at the hospital and just wanted to pick up a extra shift here or there? Or perhaps you'd rather be cared for by someone from a temp agency, who has never before been vetted by the hospital, and doesn't care all that much because they may never work there again.
Slashdot Mirror
Yes, but it's already a waste of bandwidth to let spammers spew trillions of emails at our /dev/nulls.
Not to mention what are we supposed to do when our /dev/nulls fill up.
My wife and I have been talking about this quite a bit recently. We've been watching "American Idol". Anwar Robinson is clearly the most musically talented person to ever be on that show, but the stuff he does is not what the record companies want to market. This past Monday he got up and sang Louis Armstrong's "Wonderful World". He started out singing like Louis Armstrong, then series of runs as he moved the style into something more reminicent of Sammy Davis Jr., and finally ended the song in a soul style. His talent should win, but he won't fit into the marketing machine of the record companies.
Except that cartels only work 1) when comsumers can't "do without" and 2) when there is no alternative good that can satisfy their needs.
If music sales are truly suffering, something of which I'm not convinced, it could be due to either of those factors or a combination. Piracy provides an alternate source. But consumers are also capable of just not buying.
By reducing prices, they can reduce the effect of piracy, and reduce the numbers of potential consumers who choose not to consume. This can have the net effect of increasing their revenue and profit.
Actually, we're already a metric country in any way that matters. Does miles per hour on a highway sign or measuring a recipe in cups and tablespoons really matter all that much. HELL NO!. What matters is machine tools, and they've been converting over for quite a while now. What matters is that a transmission manufactured in Detroit can be mated to an engine block manufactured in Japan.
Actually, a scientist who prefered to use Fahrenheit measurements would probably use the Rankine scale. Zero Rankine is, like 0 Kelvin, absolute zero. However, the size of a degree is the same as the Fahrenhiet scale. The freezing point of water (32F) and the boiling point of water (212F) correspond to 491.67R and 671.67R, respectively.
It is all in the intent. In blocking port 25, Verizon is not attempting to use their strength to block competing services. They are trying to block spam.
It is pretty clear in this case that the telco was attempting to hamstring a competing product.
Now, if Verizon was charging a per-email fee to use their own mail servers and blocking access to external mail servers, the situation may be more comparable.
The situation is not so cut and dried. As a practical matter, the ISPs in the United States that can effectively carry VoIP traffic also hold their competitive positions as the result of government granted monopolies and franchises: the telephone system and cable tv system. A business that has received such a level of competitive protective from the government generally can't simultaneously be a completely private business.
Actually you raise an interesting issue. The SCotUS decided that, for the purposes of applying the death penalty, a person who committed their crime at the age of 17 years and 364 days must be in law treated differently than a person who was 18 years of age at the time of the crime. The trend of the law before this has been to acknowledge a grey area between youth and adulthood, and that defendents under the age of 18 would be evaluated by the court on several factors before determining if the defendent would be treated as an adult or a juvenile.
If we accept the SC's decision that a minor one day short of his eighteenth birthday should not be treated as an adult for the purpose of administering a capital sentence, why should that minor still be eligible for another severe sentence. Is Life without Possibility of Parole any less cruel and unusual?
If I was a defense attorney for such a client, I'd be running as fast as I can to file an appeal for my life sentenced minor.
An effective team is going to wind up doing the same tasks regardless of the methodology chosen. For instance:
A team practicing XP is going to jump into coding very soon. Then they'll refactor several times and toss away code that just didn't fit the problem. All the time they are showing the customer and getting additional feedback.
Meanwhile, a team practicing a waterfall methodology is going to build several prototypes during the requirements gathering and architecture stages. They'll do several experiments to insure that they understand key parts of the design. Then they'll agree on a spec and an architecture and finally start coding.
The difference is semantic. I can say I'm doing a prototype or experimenting with a key feature, then later use key parts or my results to build the actual product. Or I can say I'm working on the actual product and throw away substantial parts of my first iteration.
But what about a web app? If a popular website is using Oracle on the backend, is that several million users? Or is it one user, because only one webserver is accessing it? Or maybe it is the size of the connection pool?
You're right, but there are additional considerations...
First, the major software vendors would love to move the market to a subscription service model, but it may be a better model for the consumer as well.
Take Microsoft, for instance. When MS sold a copy of Office 95, they made their money once. If there was no compelling reason to upgrade, then they could never make another dime selling Office to that customer. So MS's self interest is to get the Office 95 customer to buy Office 98, 2000, XP, 2003, ... To do that they need to 1) add features, 2) change file formats to artificially reduce interoperability, 3) make sure that the new version didn't have some bugs that the older version had. Note that MS has little motivation to fix bugs in versions it has already sold.
In a subscription world, MS gets a revenue continuously from their customers. There is no longer a motive for MS to resell to the same customer, so their is no motive to add features of dubious value, alter file formats, or leave bugs in older versions. MS doesn't care if the customer upgrades to the latest version, so MS is more than happy to fix bugs in older versions. In the end, customers get software that is more stable and secure.
My other point is that thin clients can drasticly reduce the operating costs of an enterprise full of desktops. Back in the Windows 95 days, I actually was part of an eval to consider replacing all the desktops in a 400 person enterprise with Citrix systems. Any user could work on "their" pc from any desk in the company. The old Win3.1 boxes floating around could run the client just fine, and wouldn't need to be replaced. Server administration, backups, everything could be centralized. The technology was quite ready for prime time, but I do some consulting for a major wall street company that is today using Citrix over the web instead of a normal VPN. I log into the web site and I'm connected via Citrix to a basic machine with my Email, Instant Messaging, Word, Excel, and other basic apps right there. If needed I can do a WinXP remote desktop session to my own desktop.
For real fun, imagine a company that decides to drink the Linux-on-Desktop kool-aid. Start with an IBM mainframe. Partition it into several thousand individual Linux instances, one per employee. Give each employee a system that runs X - or maybe even a modified VNC....
Now if a system needs to be deployed for a new employee: run one Perl script and it is done is 5 seconds. If someone needs more ram or disk quota: 2 more scripts. A patch needs to be pushed to every system: no problem. Backups: easy. Disaster recovery: trivial.
If buffer overflows are getting past your unit tests, it's because you're not writing proper unit tests.
Unit tests? I've never heard of that before. Can you tell me more?
Seriously, though, that is an excellent point. In the real world, large C codebases are going to have security holes because some part of the system, be it the C code itself or the libraries, was not sufficiently unit tested. In a Java environment, insufficiently tested code is likely to abend but is less likely to open a security hole, where in a C environment insufficiently tested code is a security risk.
But these things all require a robot to consider in advance the likely outcome of several courses of action. It means that the robot must not only model its own behaviour but also model the behaviour of other objects and people in the area.
For instance, a robot might be intelligent enough to not injure a person directly, but if you take that robot to the top of a tall building and tell it to throw bricks, it is another step for that robot to determine that 1) a brick falling from a high place might injure a person, and 2) there may be people on the street below.
To make it even more complex, the robot is still on a roof top. A guest of wind knocks a person off balance, and they are in danger of falling. The robot has to consider 1) the likelihood of the person regaining their balance without intervention, 2) the likelihood that grabbing the person with a robot manipulator will injure the person, and 3) the likelihood that the person will fall from the roof to their death. Now, the robot needs to make a judgement call, to grab and potentially break an arm in the process, or to not grab and possibly see someone plunge to their death. Still a tougher judgement to make.
It is possible to build these models, but to use AI in order to build these models in an effective way is a stunning leap forward from our current state.
Actually, your the one that needs to grow a brain. They can certainly sue a dead defendant. It happens all the time. The estate of the defendant is liable for any judgement.
It works in the opposite way too. A dead person can be a plaintiff. If you were to die due to, perhaps, medical malpractice, you (through your own estate) would sue for wrongful death. Your spouse or other next of kin don't have standing to sue for wrongful death, only you do, and so your estate must make the claim.
In this case they chose to drop the suit because of the bad PR. They are probably investigating right now to determine if a family member was using grandma's name to buy internet access. They'll sue that family member once they get the evidence.
Bruce Willis is dead throughout the whole movie.
I didn't say "top 5% programmer". Developer is much more encompassing. One of the skills of a top developer is requirements discovery, even requirements anticipation. Not to mention planning and estimating. Even more important, a top developer understands the architecture underlying the entire project, and therefore can write code that not only adheres to the specs, but also integrates well with other components.
Large innovative projects are frequently doomed to fail. One major problem is that it is very difficult to staff a project properly. It takes very bright developers to build innovative things. It takes a lot of very bright developers to build very large innovative things. The odds of being able to collect the kind of talent needed are often very slim.
The top 5% developers are almost never found on the unemployment line, so you can't hire them from there. They are almost never working for body shops, so you can't find them there. Generally they are 1) comfortably employed somewhere else, and not interested in jumping, 2) working for a commercial software house and not interested in working for a company where they are "overhead" and not the center of the business, or 3) they are self employed and making much bigger bucks freelancing than they could be paid within a corporate job.
Most companies have a few of these highly qualified people. The best way to insure successful projects is to make sure that there are only as many innovative projects as can be staffed by your innovative people. A large business changing project is out of reach. But a several small to midsize projects can also transform the business, and can be executed by the staff of innovators already on hand.
What could be more intuitive?
Actually, that is quite wrong. Nearly all electrical power systems we encounter in our daily lives carry current in a two wire circuit. If everything is operating properly, the current in one wire will exactly equal the current flowing in the opposite direction in the other wire. The GFCI device constantly measures these two currents and compares them.
If, however, any of that current finds an alternate path - perhaps through a hair dryer that has been dropped into a bath tub - the current in each wire will be different. This imbalance causes the GFCI device to trip. Note that if the current flows from one power conductor, through you, and back into the other power conductor, you will still be electrocuted.
Even if a company makes a proprietary derivative work, they are very likely to contribute back to the Open Source code base. Maintaining a fork indefinately is expensive. It is much better to maintain the derivative work as a limited set of additions/revisions to the publicly available base work.
However the users of the proprietary work expect support. So the proprietary maintainer is constantly doing bug fixes in areas of the code outside the proprietary modules. It is better that these fixes be contributed back to the base project, rather than maintain an ever growing set of patches for the proprietary product. There own selfish motives can therefore improve the base work.
It is just as easy to argue the opposite. The flip side is that BSD is more free because you can do whatever you want with derivative works, whereas the GPL is restrictive because because you can only release a derivative work under the GPL.
As a practical matter, a software vendor may have some piece of work that they want or need to keep proprietary. They cannot distribute that code integrated into a GPL package. However they can distribute that code as part of a BSD licensed package.
While in theory this doesn't seem to offer an advantage, in practise the developer community for the base package is larger. These developers inevitably contribute large amounts of work back to the base project.
You are correct that the GPL and BSD licenses are very different. How they differ is not a factor, however, for the 99.99% of the users who will use this software without ever modifying or redistributing the source code.
The GPL is in some ways a vestige of a time when computer users were largely computer programmers. Now, most people will never even look at a line a code. For them both Open Source and Free Software mean high quality software without BSA audits, burdensome license agreements, and closed standards. There is no lock in, so software vendors must compete on value.
I believe that Microsoft's coders are toward the elite of the industry. Microsoft's quality problem is primarily this...
Given the choice between coding a new feature and improving the reliability of the existing feature set, Microsft marketing and management have always chosen the new feature.
Code can be delivered quickly, can be designed innovatively, and can be bug free. Code can never have all three at once.
Wake up! What do you think has been going on in the IT sector?
During the boom, how many people jumped from job to job every 6-9 months because someone else waved ten or twenty or fifty thousand dollars more at them? Was it ok then? Suddenly it isn't ok?
Our work has no intrinsic value. Our work is valued at exactly how much we can convince someone to pay us. No more and no less.
Someone that has a proven track record as a successful architect is going to get $150,000 to $300,000 a year, at least. Someone that has read HTML For Idiots is going to work at Denny's bussing tables. Anyone that expects to get an architect is going to have to pay the going rates for architects. It doesn't matter how many HTML jockeys are working at Denny's, it doesn't affect the market for architects.
I know. I'm trying to hire and I can't find good people at any price. Playing with MS Access does not qualify you to design TerraByte size DB2 or Oracle systems. Designing a web site for you cousin's deli does not prepare you to design a major J2EE infrastructure for a Fortune 500 company.
Would you rather be cared for by a nurse who was considered good enough to have a steady job at the hospital and just wanted to pick up a extra shift here or there? Or perhaps you'd rather be cared for by someone from a temp agency, who has never before been vetted by the hospital, and doesn't care all that much because they may never work there again.