You're confusing the GPL, which is an affirmative grant of use and copy rights, with patent law, which would apply to a bare algorithm. Microsoft would, of course, be perfectly free to re-implement an algorithm that was embodied in a GPL'ed application or library, assuming that it wasn't patented by a free source advocate unclear on the concept.;-)
No, I realize that the algorithm can be re-implemented, but Joe's software house may not have the money to re-implement it.
Money is saved, for Microsoft, Intuit, Adobe, Joe's software house, and fubar computing if they can use code developed from government funded code.
Corporations are not by definition evil. As another poster pointed out, corporate profit helps people too and Microsoft isn't the only company with an interest in government funded research.
Additionally, every time someone re-implements an algorithm, there is a chance that they will implement it incorrectly. This is especially true in with encryption algorithms. Everyone would be better off if base algorithms weren't re-written by every single company that needs to use them. This is where government funded BSD licensed code (sorry, I haven't researched LGPL enough to know if it solves GPL's problems with sharing code with closed source groups or not) could improve software quality for everyone.
"Two points. First, GPL'ed software runs on Windows all the time.. see Cygwin, Xemacs, etc. That doesn't lead to a requirement that Windows be licensed under the GPL, anymore than the fact that the Linux kernel is licensed under the GPL means that my copy of DB2 for Linux has to be licensed under the GPL."
Runs on...not can run *IN*. If I am doing government research and write an application for analyzing data and invent a new algorithm for doing a mathematical transform that is a few orders of magnitude more efficient, but I just GPL the app, I have just locked 3/4 of the software development world out of this code.
If they want to use the algorithm in closed source code, to protect themselves they need to cleanroom reverse engineer the code.
Assume that someone makes a standalone GPL app for transforming images. MS couldn't take that and integrate it into the display processing because it needs to be a standalone app.
So either: They can use the software in an inefficient kludgefest of a hack. or: They can cleanroom reverse engineer the code.
By selecting GPL over BSD for public research, you force the closed source companies to either reinvent the wheel, or to reverse engineer and rewrite the wheel from scratch.
If the closed source company is reverse engineering the wheel in the first place, there is a fair amount of individual programmer temptation to improve on the wheel when they re-write it.
If they weren't re-writing it, but just dropping it in place, would they have nearly as much temptation to "fix" open standards.
Besides, I believe that the kerberos standard listed the field that was change as "UNDEFINED". The standard didn't say "You can't put anything here." If you want people to stick to a standard, make it more explicit.
"I pay company for their product - thus paying twice."
Are you buying the entire windows product only for Kerberos? If not, then shouldn't MS get some compensation for the rest of the product?
My argument is that by placing code in BSD license so that it can be dropped into place and used instead of re-written from scratch using cleanroom techniques, you reduce the cost to develop the commercial package.
Since the means to reduce the cost is available to all, even closed source software houses, everyone can use it, it brings down everyone's costs, and everyone can afford to sell at a lower price. With market forces in effect on all products, even an operating system (you can always go to unix or OS/2) and linux is a huge market force, prices will be driven down.
You give no proof that the company will not decrease the price of their product as their costs decrease.
Even assuming that the price will not decrease, it results in lower overhead, and more profits for the company which are invested and result in lower bank interest rates for me.:) Enabling profit is NOT a bad thing.
"The Microsoft donations, a 20-fold increase over 1996, analysts say, is part of a strategy to curry favor with the GOP in the hope a Republican president would be more eager to settle the antitrust case without the break up ordered by a U.S. judge"
If your company was threatened with distruction, wouldn't you lobby hard?
It is my opinion that the anti-trust trial is mostly the result of lobbying by Sun, AOL, Novell, Kodak, and other competitors combined with a lack of lobbying by Microsoft in the past.
I believe that in 1995, Microsoft's lobbying budget was around $15k. That isn't even enough to hire a 1/4 time contractor to talk to people!
Why does everyone keep saying that I am paying twice? I think that everyone is ignoring the fact that more than just MS can benefit from this.
Okay, I pay for the research the first time, granted. Now Collaborate exists.
Then, say Collaborate is useful to general ledger programs so MS uses it in MS Money and Intuit uses it in Quicken and Quickbooks. This decreases the costs to both companies and market forces bring down the cost of the packages relative to the benefit gained by the tax funded library.
So now MS extends Collaborate, they have a slight advantage over Intuit and market forces let them charge a little more. Seems entirely fair to me.
Besides, you are ignoring the fact that the legal issues of GPL require the larger work to be open sourced...MS couldn't use a GPL library inside windows even if they wanted to because they place too much value on maintaining control over the codebase that they paid billions to write and maintain.
Active Directory
Multiple Domain controllers per domain
Multiple Domains per tree
Multiple Trees per forest
Single Signon to access resources anywhere inside the forest.
Data Replication between domain controllers.
The ability to manage users, resources, and services like DHCP and DNS from anywhere in the forest making the changes in one place and the configuration changes are replicated.
Microsoft's W2k Server give a *LOT* of features and functionality that I can't get in the unix world unless I want to do a lot of scripting.
1) Microsoft lets the world know about their problems. Other companies like Sun force customers to sign NDAs before agreeing to fix known critical problems. Of course, the Sun issue was with hardware, but Sun HID the problem.
2) Security *WAS* a low priority compared to features, but that is changing. MS is not a heavy handed beurocracy and many decisions are made by individuals at the bottom of the organizational chart. It takes time for 40,000 programmers to change the way they work, but it is the new high priority. BTW, most of the design for XP was made a year or more before releasing the product.
3) "Trustworthy computing" means that MS will be focusing on both improved quality and improved security, but this could be a 10 year process.
4) W2k and XP are substantially higher quality than Win95...that is a trend that was ignored.
5) Over time, with the Online Crash Analysis, the end user problems that MS hasn't seen in the past will now be seen and fixed.
6) The assumption that money fixes bugs is incorrect. If one programmer works on a piece of code for 5 years and the code is very complex, you can't just hire an additional programmer to jump in and help fix bugs in this complex codebase. There is a time factor involved. Infinite money will not fix even one complex bug instantly.
Locking government research into GPL is just as bad as locking it out of open standards.
GPL fanatics act as if Microsoft is the only closed source software company in the world.
If a government researcher creates a better code library, wouldn't you like everyone to be able to benefit? Locking this code into GPL keeps it out of most Adobe, Oracle, Intuit, Sun, Microsoft, and even Idaho Computer Services software.
If my tax dollars pay for something, I want is shared with EVERYONE, I want all tax funded software development to be under the BSD license.
I think what you find lacking are wizard-type interfaces like Windows tries to do. Since each Linux distribution is unique, the responsiblity of providing them lies mostly on the distributors. Red Hat and others have made progress in ease of use, but they do need more time to mature.
Actually, what I find lacking is that to install windows software I rarely need to spend more than 5 minutes before it is up and running, but with linux I usually need to reserve an entire weekend.
Linux usability studies?
on
GUIs for Everyone
·
· Score: 3, Informative
Usability isn't just for the framework, it is also for the individual applications. Windows has standards that are recommended for applications.
First, are there application or user experience standards for KDE, Gnome, X, or command line apps? I know that there are a few de-facto standards on the command line, but is anything codified (especially for gui)?
Second, how many open source projects have done a usability study to see if your aunt, cousin, grandmother, or neighbor can easily use your cool new application or tool without significant assistance?
Formal usability studies are expensive and time consuming, but they do work.
Then again, if you are building a car in your garage, do you just care about yourself, or do you spend the extra week to make an adjustable seat so that it is comfortable for other drivers?
If you want me to move back to using linux as my main desktop machine, you need to make it much easier to install and configure the OS, the desktop, and all of the applications. Linux may be powerful, but I don't necessarily want the power to cut my leg off if I don't spend an hour reading the docs before I attempt to compile and install a new program.
Through the use of crash reports, Microsoft's next service packs and OS should withstand much more cruft without crashing. It might run slowly, but the only excuse for un-fixed crashes is a lack of knowledge.
Microsoft has already released a number of hotfixes for issues that were discovered because of the crash reports that users have submitted.
Microsoft is also using the reports to identify 3rd party drivers and other software that causes crashes. Microsoft *does* fix anything that causes a crash in NT/2k/XP that they know about...this is their way of finding out why end users experience crashes.
That's the problem with cruft, it is hard to reproduce it in a controlled environment, but because developers usually need to understand what causes the problem, they want all testing to happen in controlled environments. Crash reports give developers enough information to know where to look for the problem and they give managers the big stick of "This problem has hit 100 users" so they can justify the programmer spending the time needed to understand and fix the problem.
I went to a talk recently where a researcher was explaining human factors applied to military jet aircraft. The explanation that he gave of reboots in these systems was a 1/10th of a second or less pause - the pilot pushes a button to say "No, the computer has it wrong, it is giving me a different display than I need, reboot and give me the default display again."
A "personal" computer reboot takes > 30 seconds nd would be unacceptable. These reboots are near instantanious.
(I could be wrong, maybe this is a different aircraft and a different type of reboot than the researcher was talking about.)
The difference is that with Remote Assistance, my mother can demonstrate her problem instead of just describing how "it breaks when she clicks on the thingamajig" and she can see how I fix the problem.
For example, if she can't figure out how to switch her printer properties from landscape to portrait, I can demonstrate it for her (remotely) so she can do it herself next time.
As far as the "connecting to the machine" aspect, to me, remote desktop is "as easy" as SSH.
The entire support experience is "much easier" but the connection aspect is "as easy".
I would agree that firewalls, cable modems with built-in NAT, and flakey connections can make it painful to connect to personal computers remotely, but that is not an OS issue.
I'd prefer to install Windows XP on my mom's computer and use Remote Assistance to show her how I fix the problem so that she can do it herself next time. Remote Desktop Connection, and Remote Assistance make the latest versions of Windows much easier to support from a distance than Linux and they seem to be as stable as linux.
My parents currently run about 50 programs that are not available on Linux. Even if they were available, they use windows at work so it dosen't make sense for them to switch operating systems and applications all of the time.
I still have linux and bsd boxen around, but I have switched *back* to windows after being a unix and OS/2 die hard since suffering from windows 95's crashes.
Can you point me to the clause that explains this?
If I write GPL code, and because I write GPL code I'm going to use a GPL library, or maybe someone else give me a bug fix that is GPLed, I can't sell the exclusive rights to this code, I can't sell the rights to redistribute without the source, there *IS* a lot of taint here - and maybe it is just because of that library or that bugfix, but what good is a library if you can't sell your original work if it used the library?
Maybe I am "the uniformed", but I've read through the GPL legalese several times and that seems to be the entire POINT of GPL. If I am wrong, please explain it to me, and please quote the license when you do.
First, I don't really care about credability, I just don't get why anyone would want to allow something they enjoy to get involved with anything as disgusting as a government contract.
The government can't buy anything without a large series of tests/verifications/support contracts/etc.
I've worked for a government research lab for a year - there is so much red-tape that you can't do anything. We order a computer, pay 50% above retail price, wait 4 months for it to be delivered, and then wait another 3 months for custom software and firmware to be installed that (1) monitors the users (2) can't be uninstalled (3) crashes 10x more than the bare OS would have (4) runs 1/100th of the speed that it should (5) can only be upgraded/maintained/repaired/touched by an authorized technician of the company who had the computer support contract.
Once you get in with the government, you now need to meet their requirements. The typical developer that writes an app as a hobby dosen't have the time or resources to test for accessibility, testing security, testing for easter eggs, writing specs, ensuring that the spec is up-to date, verifying the nationality of everyone who contributed code, etc, etc, etc...
Go for "legitimacy" as a platform if you like - I just think that the requirements of the corporate and government worlds will cause OSS to no longer be a fun project for the current set of developers. There is a reason why places like MS pay people to do these things.
Except that I *can't* use GPL for just *anything* that I write.
*IF* I want to use a GPL sort routine in my database app, I must release the source to my app. If I release the source code, there is a significant percentage of people that will not pay to use my app.
GPL is anti-business for any business that makes money from selling software.
The transition to selling support dosen't work - that either results in hard to use/expensive to maintain software or in the developer going out of business.
Take a look at PKzip - we all used it, only 1% of us licensed it, the author was miserable when he died - we should all be ashamed - my understanding of GPL is that it forces this on anyone who dips from the GPL pool.
Re: Your comment on how GPL is better than BSD: So, you give your code away, but only to people who won't profit from it? Dosen't that just make you feel warm and fuzzy without actually doing anything to help others?
IP is traditionally held by an individual or corporation. GPL forces it to be owned by everyone - it can't even be put under a more restrictive license later on. Unless it is a superset of GPL, it can't even be put under a less restrictive license later on. You can't re-negotiate the license for a private party. In a lot of ways, from the view of traditional software, GPL *is* anti-IP. GPL is IP that can no longer be privately owned.
I'm probably going to get flamed for this, but these arguments are getting silly.
It seems to me that the whole proprietary vs "open source" boils down to an argument of how we share things.
Proprietary = I have cool toys, make it worth my while to let you play. Maybe we can share toys, maybe you keep the bully away from me if you get to play with my toys, or maybe we can make another arrangement. I'm not going to tear your head off if you borrow my toys without permission, but we might end up in a little spat. ie - capitalism
BSD = we do it for the fun of it, do what you want - if you can make a profit, cool! ie - pure research (historically good for buinesses)
GPL = you can play with my toys, but if you do you have to let me play with your toys. Mom and Dad are going to enforce this. These toys are magic, they can't even be traded for non-sharing toys. ie - communism - although somehow people are arguing that this is pro-business?
At this point in the game, the commune is small and can still function, but as it grows it will self destruct.
Do you *really* want to get government burocracy involved in your OSS commune?
In fact, MS is a nice place to work. If you have your own office, you can have your ORA book set within arm's reach, your plants nearby to provide oxygen and relaxation, and a door that you can close when you need to keep out the interruptions.
This is an interesting idea when looking at commute times, but I can Terminal Service from home to my office machines too. I don't need to be "hotelled" to get out of traffic.
One of the problems that I can see with Hotelling is that if everyone is on the move, I can't walk around the corner and ask a co-worker or manager about a problem, demonstrate it on their computer, and cooperatively sketch out solutions on their whiteboard.
BTW - I know several MSofties who had previously sworn to themselves that they would never work for MS and are currently very happy working at MS.
Fixing the software to the customer's satisfaction may imply a lot more than one might initially assume.
1) Is the fix made in a timely manner? If I base my business on a webstore that you wrote in your spare time, will you be able to get me a fix within 8 hours in the middle of your final exams?
2) Does the fix solve the customer's problem? What if it is a performance / scalability problem, you designed the app to handle 100 transactions per day and the user wants to do 10,000? It dosen't need to be a technical problem either - what if the software is just too complex to use?
Then, consider all of the overhead that the distributer needs in order to (a) test the problem to make sure that they can reproduce the bug - you don't want your developers to get thousands of bogus bug reports (b) find and communicate the problem to the developer (c) and communicate the status to the user.
I think that this is a larger issue than most slashdot readers realize.
Slashdot Mirror
You're confusing the GPL, which is an affirmative grant of use and copy rights, with patent law, which would apply to a bare algorithm. Microsoft would, of course, be perfectly free to re-implement an algorithm that was embodied in a GPL'ed application or library, assuming that it wasn't patented by a free source advocate unclear on the concept. ;-)
No, I realize that the algorithm can be re-implemented, but Joe's software house may not have the money to re-implement it.
Money is saved, for Microsoft, Intuit, Adobe, Joe's software house, and fubar computing if they can use code developed from government funded code.
Corporations are not by definition evil. As another poster pointed out, corporate profit helps people too and Microsoft isn't the only company with an interest in government funded research.
Additionally, every time someone re-implements an algorithm, there is a chance that they will implement it incorrectly. This is especially true in with encryption algorithms. Everyone would be better off if base algorithms weren't re-written by every single company that needs to use them. This is where government funded BSD licensed code (sorry, I haven't researched LGPL enough to know if it solves GPL's problems with sharing code with closed source groups or not) could improve software quality for everyone.
"Two points. First, GPL'ed software runs on Windows all the time.. see Cygwin, Xemacs, etc. That doesn't lead to a requirement that Windows be licensed under the GPL, anymore than the fact that the Linux kernel is licensed under the GPL means that my copy of DB2 for Linux has to be licensed under the GPL."
Runs on...not can run *IN*. If I am doing government research and write an application for analyzing data and invent a new algorithm for doing a mathematical transform that is a few orders of magnitude more efficient, but I just GPL the app, I have just locked 3/4 of the software development world out of this code.
If they want to use the algorithm in closed source code, to protect themselves they need to cleanroom reverse engineer the code.
Assume that someone makes a standalone GPL app for transforming images. MS couldn't take that and integrate it into the display processing because it needs to be a standalone app.
So either:
They can use the software in an inefficient kludgefest of a hack.
or:
They can cleanroom reverse engineer the code.
By selecting GPL over BSD for public research, you force the closed source companies to either reinvent the wheel, or to reverse engineer and rewrite the wheel from scratch.
If the closed source company is reverse engineering the wheel in the first place, there is a fair amount of individual programmer temptation to improve on the wheel when they re-write it.
If they weren't re-writing it, but just dropping it in place, would they have nearly as much temptation to "fix" open standards.
Besides, I believe that the kerberos standard listed the field that was change as "UNDEFINED". The standard didn't say "You can't put anything here." If you want people to stick to a standard, make it more explicit.
"I pay company for their product - thus paying twice."
:) Enabling profit is NOT a bad thing.
Are you buying the entire windows product only for Kerberos? If not, then shouldn't MS get some compensation for the rest of the product?
My argument is that by placing code in BSD license so that it can be dropped into place and used instead of re-written from scratch using cleanroom techniques, you reduce the cost to develop the commercial package.
Since the means to reduce the cost is available to all, even closed source software houses, everyone can use it, it brings down everyone's costs, and everyone can afford to sell at a lower price. With market forces in effect on all products, even an operating system (you can always go to unix or OS/2) and linux is a huge market force, prices will be driven down.
You give no proof that the company will not decrease the price of their product as their costs decrease.
Even assuming that the price will not decrease, it results in lower overhead, and more profits for the company which are invested and result in lower bank interest rates for me.
USA Today article
"The Microsoft donations, a 20-fold increase over 1996, analysts say, is part of a strategy to curry favor with the GOP in the hope a Republican president would be more eager to settle the antitrust case without the break up ordered by a U.S. judge"
If your company was threatened with distruction, wouldn't you lobby hard?
It is my opinion that the anti-trust trial is mostly the result of lobbying by Sun, AOL, Novell, Kodak, and other competitors combined with a lack of lobbying by Microsoft in the past.
I believe that in 1995, Microsoft's lobbying budget was around $15k. That isn't even enough to hire a 1/4 time contractor to talk to people!
Why does everyone keep saying that I am paying twice? I think that everyone is ignoring the fact that more than just MS can benefit from this.
Okay, I pay for the research the first time, granted. Now Collaborate exists.
Then, say Collaborate is useful to general ledger programs so MS uses it in MS Money and Intuit uses it in Quicken and Quickbooks. This decreases the costs to both companies and market forces bring down the cost of the packages relative to the benefit gained by the tax funded library.
So now MS extends Collaborate, they have a slight advantage over Intuit and market forces let them charge a little more. Seems entirely fair to me.
Besides, you are ignoring the fact that the legal issues of GPL require the larger work to be open sourced...MS couldn't use a GPL library inside windows even if they wanted to because they place too much value on maintaining control over the codebase that they paid billions to write and maintain.
Windows Server has:
Active Directory
Multiple Domain controllers per domain
Multiple Domains per tree
Multiple Trees per forest
Single Signon to access resources anywhere inside the forest.
Data Replication between domain controllers.
The ability to manage users, resources, and services like DHCP and DNS from anywhere in the forest making the changes in one place and the configuration changes are replicated.
Microsoft's W2k Server give a *LOT* of features and functionality that I can't get in the unix world unless I want to do a lot of scripting.
Just a few commnets on "Why so buggy?"
1) Microsoft lets the world know about their problems. Other companies like Sun force customers to sign NDAs before agreeing to fix known critical problems. Of course, the Sun issue was with hardware, but Sun HID the problem.
2) Security *WAS* a low priority compared to features, but that is changing. MS is not a heavy handed beurocracy and many decisions are made by individuals at the bottom of the organizational chart. It takes time for 40,000 programmers to change the way they work, but it is the new high priority. BTW, most of the design for XP was made a year or more before releasing the product.
3) "Trustworthy computing" means that MS will be focusing on both improved quality and improved security, but this could be a 10 year process.
4) W2k and XP are substantially higher quality than Win95...that is a trend that was ignored.
5) Over time, with the Online Crash Analysis, the end user problems that MS hasn't seen in the past will now be seen and fixed.
6) The assumption that money fixes bugs is incorrect. If one programmer works on a piece of code for 5 years and the code is very complex, you can't just hire an additional programmer to jump in and help fix bugs in this complex codebase. There is a time factor involved. Infinite money will not fix even one complex bug instantly.
It is about choice.
Locking government research into GPL is just as bad as locking it out of open standards.
GPL fanatics act as if Microsoft is the only closed source software company in the world.
If a government researcher creates a better code library, wouldn't you like everyone to be able to benefit? Locking this code into GPL keeps it out of most Adobe, Oracle, Intuit, Sun, Microsoft, and even Idaho Computer Services software.
If my tax dollars pay for something, I want is shared with EVERYONE, I want all tax funded software development to be under the BSD license.
I think what you find lacking are wizard-type interfaces like Windows tries to do. Since each Linux distribution is unique, the responsiblity of providing them lies mostly on the distributors. Red Hat and others have made progress in ease of use, but they do need more time to mature.
Actually, what I find lacking is that to install windows software I rarely need to spend more than 5 minutes before it is up and running, but with linux I usually need to reserve an entire weekend.
Usability isn't just for the framework, it is also for the individual applications. Windows has standards that are recommended for applications.
First, are there application or user experience standards for KDE, Gnome, X, or command line apps? I know that there are a few de-facto standards on the command line, but is anything codified (especially for gui)?
Second, how many open source projects have done a usability study to see if your aunt, cousin, grandmother, or neighbor can easily use your cool new application or tool without significant assistance?
Formal usability studies are expensive and time consuming, but they do work.
Then again, if you are building a car in your garage, do you just care about yourself, or do you spend the extra week to make an adjustable seat so that it is comfortable for other drivers?
If you want me to move back to using linux as my main desktop machine, you need to make it much easier to install and configure the OS, the desktop, and all of the applications. Linux may be powerful, but I don't necessarily want the power to cut my leg off if I don't spend an hour reading the docs before I attempt to compile and install a new program.
Through the use of crash reports, Microsoft's next service packs and OS should withstand much more cruft without crashing. It might run slowly, but the only excuse for un-fixed crashes is a lack of knowledge.
Microsoft has already released a number of hotfixes for issues that were discovered because of the crash reports that users have submitted.
Microsoft is also using the reports to identify 3rd party drivers and other software that causes crashes. Microsoft *does* fix anything that causes a crash in NT/2k/XP that they know about...this is their way of finding out why end users experience crashes.
That's the problem with cruft, it is hard to reproduce it in a controlled environment, but because developers usually need to understand what causes the problem, they want all testing to happen in controlled environments. Crash reports give developers enough information to know where to look for the problem and they give managers the big stick of "This problem has hit 100 users" so they can justify the programmer spending the time needed to understand and fix the problem.
I went to a talk recently where a researcher was explaining human factors applied to military jet aircraft. The explanation that he gave of reboots in these systems was a 1/10th of a second or less pause - the pilot pushes a button to say "No, the computer has it wrong, it is giving me a different display than I need, reboot and give me the default display again."
A "personal" computer reboot takes > 30 seconds nd would be unacceptable. These reboots are near instantanious.
(I could be wrong, maybe this is a different aircraft and a different type of reboot than the researcher was talking about.)
The difference is that with Remote Assistance, my mother can demonstrate her problem instead of just describing how "it breaks when she clicks on the thingamajig" and she can see how I fix the problem.
For example, if she can't figure out how to switch her printer properties from landscape to portrait, I can demonstrate it for her (remotely) so she can do it herself next time.
As far as the "connecting to the machine" aspect, to me, remote desktop is "as easy" as SSH.
The entire support experience is "much easier" but the connection aspect is "as easy".
I would agree that firewalls, cable modems with built-in NAT, and flakey connections can make it painful to connect to personal computers remotely, but that is not an OS issue.
Neither!
I'd prefer to install Windows XP on my mom's computer and use Remote Assistance to show her how I fix the problem so that she can do it herself next time. Remote Desktop Connection, and Remote Assistance make the latest versions of Windows much easier to support from a distance than Linux and they seem to be as stable as linux.
My parents currently run about 50 programs that are not available on Linux. Even if they were available, they use windows at work so it dosen't make sense for them to switch operating systems and applications all of the time.
I still have linux and bsd boxen around, but I have switched *back* to windows after being a unix and OS/2 die hard since suffering from windows 95's crashes.
1. Regarding the start bar, you can go back to the W2k theme and start menu style to see at least a 10% increase in speed.
2. A MSCP should be able to disable the "bells and whistles" that the user didn't want.
3. Again, the MSCP should be able to set some regkeys to disable the undesired functionality.
So, the 40k MS employees don't pay federal income tax?
Besides, I suspect that MS does pay quite a few taxes even if they have wiggled out of some.
Can you point me to the clause that explains this?
If I write GPL code, and because I write GPL code I'm going to use a GPL library, or maybe someone else give me a bug fix that is GPLed, I can't sell the exclusive rights to this code, I can't sell the rights to redistribute without the source, there *IS* a lot of taint here - and maybe it is just because of that library or that bugfix, but what good is a library if you can't sell your original work if it used the library?
Maybe I am "the uniformed", but I've read through the GPL legalese several times and that seems to be the entire POINT of GPL. If I am wrong, please explain it to me, and please quote the license when you do.
First, I don't really care about credability, I just don't get why anyone would want to allow something they enjoy to get involved with anything as disgusting as a government contract.
The government can't buy anything without a large series of tests/verifications/support contracts/etc.
I've worked for a government research lab for a year - there is so much red-tape that you can't do anything. We order a computer, pay 50% above retail price, wait 4 months for it to be delivered, and then wait another 3 months for custom software and firmware to be installed that (1) monitors the users (2) can't be uninstalled (3) crashes 10x more than the bare OS would have (4) runs 1/100th of the speed that it should (5) can only be upgraded/maintained/repaired/touched by an authorized technician of the company who had the computer support contract.
Once you get in with the government, you now need to meet their requirements. The typical developer that writes an app as a hobby dosen't have the time or resources to test for accessibility, testing security, testing for easter eggs, writing specs, ensuring that the spec is up-to date, verifying the nationality of everyone who contributed code, etc, etc, etc...
Go for "legitimacy" as a platform if you like - I just think that the requirements of the corporate and government worlds will cause OSS to no longer be a fun project for the current set of developers. There is a reason why places like MS pay people to do these things.
So, I can take a piece of code, GPL it, and then sell it to SoftCo for $100k under a license that allows them to redistribute it *WITHOUT* the source?
What if I take my code + GPL code and try to license it with a redistribution without source code clause?
I have never seen any indication that either of these scenarios are ever allowed.
Can you point me to the clause that I am missing?
Except that I *can't* use GPL for just *anything* that I write.
*IF* I want to use a GPL sort routine in my database app, I must release the source to my app. If I release the source code, there is a significant percentage of people that will not pay to use my app.
GPL is anti-business for any business that makes money from selling software.
The transition to selling support dosen't work - that either results in hard to use/expensive to maintain software or in the developer going out of business.
Take a look at PKzip - we all used it, only 1% of us licensed it, the author was miserable when he died - we should all be ashamed - my understanding of GPL is that it forces this on anyone who dips from the GPL pool.
Re: Your comment on how GPL is better than BSD: So, you give your code away, but only to people who won't profit from it? Dosen't that just make you feel warm and fuzzy without actually doing anything to help others?
IP is traditionally held by an individual or corporation. GPL forces it to be owned by everyone - it can't even be put under a more restrictive license later on. Unless it is a superset of GPL, it can't even be put under a less restrictive license later on. You can't re-negotiate the license for a private party. In a lot of ways, from the view of traditional software, GPL *is* anti-IP. GPL is IP that can no longer be privately owned.
I'm probably going to get flamed for this, but these arguments are getting silly.
It seems to me that the whole proprietary vs "open source" boils down to an argument of how we share things.
Proprietary = I have cool toys, make it worth my while to let you play. Maybe we can share toys, maybe you keep the bully away from me if you get to play with my toys, or maybe we can make another arrangement. I'm not going to tear your head off if you borrow my toys without permission, but we might end up in a little spat.
ie - capitalism
BSD = we do it for the fun of it, do what you want - if you can make a profit, cool!
ie - pure research (historically good for buinesses)
GPL = you can play with my toys, but if you do you have to let me play with your toys. Mom and Dad are going to enforce this. These toys are magic, they can't even be traded for non-sharing toys.
ie - communism - although somehow people are arguing that this is pro-business?
At this point in the game, the commune is small and can still function, but as it grows it will self destruct.
Do you *really* want to get government burocracy involved in your OSS commune?
In fact, MS is a nice place to work. If you have your own office, you can have your ORA book set within arm's reach, your plants nearby to provide oxygen and relaxation, and a door that you can close when you need to keep out the interruptions.
This is an interesting idea when looking at commute times, but I can Terminal Service from home to my office machines too. I don't need to be "hotelled" to get out of traffic.
One of the problems that I can see with Hotelling is that if everyone is on the move, I can't walk around the corner and ask a co-worker or manager about a problem, demonstrate it on their computer, and cooperatively sketch out solutions on their whiteboard.
BTW - I know several MSofties who had previously sworn to themselves that they would never work for MS and are currently very happy working at MS.
1) Curiosity
:)
2) A desire to learn
3) A good work ethic
Oh wait, those aren't allowed in college!
Curiosity - Universities won't let you find out what's in those steam tunnels.
A desire to learn - Universities won't let you ask other students how they approached the problem.
A good work ethic - Too many parties, too little time.
Fixing the software to the customer's satisfaction may imply a lot more than one might initially assume.
1) Is the fix made in a timely manner? If I base my business on a webstore that you wrote in your spare time, will you be able to get me a fix within 8 hours in the middle of your final exams?
2) Does the fix solve the customer's problem? What if it is a performance / scalability problem, you designed the app to handle 100 transactions per day and the user wants to do 10,000? It dosen't need to be a technical problem either - what if the software is just too complex to use?
Then, consider all of the overhead that the distributer needs in order to (a) test the problem to make sure that they can reproduce the bug - you don't want your developers to get thousands of bogus bug reports (b) find and communicate the problem to the developer (c) and communicate the status to the user.
I think that this is a larger issue than most slashdot readers realize.