You're confusing the sales pitch with the intent. It's not as if users lobbied Congress to create IP. You can bet that those who were promoting copyright were among the first to profit from it.
Well IBM, of course, doesn't really embrace the four "freedoms". When it purchased Rational, it would no longer sell a license Visual Test and it didn't make it open source either.
Why? Because Visual Test was a low-cost alternative to other Rational testing applications.
Wake up and smell the Blue. The only IP that IBM has/will made/make open source, is the commodity stuff or stuff they can't make any money on.
"It's not dependent on the engineer's opinion, it's based on business need"
I was including business need as part of "what programmers think".
"As always, being a developer with a locked in mindset is the number one risk to your usefulness, and thus job security"
I don't consider myself in that category, however, I don't agree with your theory anyway. Your relationship with your Boss is far more significant in keeping your job than any mindset is.
"could that be because computer hardware up till now has not been able to handle that level of parallelization that you would need to realistically model the real world?"
No. It's because modeling the real world wasn't the goal in many cases. Besides there's a big difference between modeling and controlling.
It's seems to me that multiple-core systems may excel in two different areas:
1) Problems for which there is already a solution, but practical applicatons require near real-time performance and we currently can't achieve that speed (assuming the problem can be solved in parallel).
2) Problems that we understand how to solve but don't yet have a solution because they would take too long to calculate (say 20 years). A large network of multicore machines might be able to significantly shorten the time (again assuming the problem can be solved more efficiently in parallel).
On the other hand, problems that we've been working on for many years without a proven approach (e.g. AI) are unlikely to be solved any quicker.
"It doesn't necessarily have to be linearly proportional. It just has to be greater than the performance we could gain by making more complex single cores."
Well, given that it requires additional time and money to take advantage of it, there needs to be a commensurate performance advantage. It's not enough to be merely "greater than" a single core.
"So whether programmers find this move acceptable or not is irrelevant because this path is probably the only way to design faster CPU:s once we've hit the nanometer wall."
I guess you should put "faster" in quotes.
In any case, it is absolutely relevant what programmers think since any performance improvements that customers actually experience is dependent on our code.
Historically a primary reason to buy a new computer is because a faster system makes legacy applications run faster. To a large extent this won't be true with a new multicore PC. So why would people buy them?
That's why Intel wants us to redesign our software - so that in the future their customers will still have a reason to buy a new PC with Intel Inside.
I understand why Intel is so interested in multiple cores - they can't make the faster single-core chips that the market wants.
The question is what's our motivation? Unless software performance is approximately linearly proportional to the number of cores (e.g. a 10 core cpu can run a software application 10x faster than it ran before it was made core-aware), it probably isn't worth converting legacy apps.
You're wrong about many of the Linux appliances. They don't change anything in the OS and thus aren't obligated to provide their own source to anyone.
What they do is take a standard PC platform, install a standard Linux distro, install their proprietary software and sell the system for hundreds or thousands of dollars more than the hardware costs. Thus they make a good profit and recoup their investment in developing their software.
I guess you missed the word "average". If the only customers who bought Red Hat were customers who asked questions of that low-level type, you'd be out of business rather quickly.
Well, given the fact that McNealy did a 360 on the value of Office suites, who knows what is true?
He made a big deal in the media about handing out white boards and markers to his employees in liu of using Powerpoint, that was widely reported at the time (probably 1997 or so).
"When they bought StarOffice, it saved them a ton of money versus buying licenses for their employees."
Actually, no. Prior to buying StarOffice, Scott McNealy forbid his employees to use a word processor with more than 4 features (i.e. a text editor instead of a word processor), so there were no licenses to save money on.
Mysteriously, after Sun bought StarOffice, McNealy stopped talking about how evil office-type applications were.
A lot of these companies "succeed" because they were purchased in an Internet-Bubble-Mindset at absurd prices.
Jboss falls in this category and is not living up to expectations at Red Hat.
Since Sun bought mysql for 1 billion dollars, it's going to be a very long time before you could prove it was a good deal.
Anyway, I'm not going to address every example. The point is that the examples are few and far between for profitable open source companies who write their own original applications and don't bundle them with hardware.
Since companies want to make a profit, they'll always go for the low-hanging fruit.
There are many companies that make specialized Linux "appliances" without modifying Linux, without paying for Linux, and without contributing to Linux.
They don't do it because there's something about Linux that they can't get elsewhere, they do it because it's cheaper than paying for a license.
What these companies are really selling is proprietary software that could run on any Linux PC, but they bundle it with a closed box so they can charge more money for it.
They could do the same with Windows (and historically people have) but then the cost of a Windows license would come out of their profits.
I didn't say anything about a lack of profit at Red Hat, I just stated that the day may come when corporations ask themselves why they should pay for something that is available for free.
As far as credit for Linux is concerned, I'd put Thompson and Ritchie at the top of the list and Linus second. I leave the rest of the ranking to the Linux community (of which I am not a part).
My broader point, however, is that Red Hat isn't a good poster boy for the business viability of open source and in particular not an example of how MS could succeed at it.
There's a big difference making a product open source that you've already spent millions of dollars on for a decade without help (e.g MS Office), and making contributions to an existing product (however limited it may have been) that was somebody else's idea.
Does Red Hat have any brand new product ideas that aren't part of the Linux OS that they'd like to spend a few years developing and release version 1.0 under the GPL license?
"Staying relevant as in staying an industry leader in computer software and remaining competitive in the industry."
Ok, but how is this goal unmet by hiring experienced developers rather then entry level graduates?
"I already mentioned that college students who contribute to open source projects are going above and beyond their school's curriculum, which gives them experience that a student who doesn't contribute to open source project won't get. "
They could be going "above and beyond" or they could be contributing to open source projects instead of fully meeting the requirements of their school's curriculum. In the real world many of your choices end once you agree to work for someone else - open source projects might not be the best place to learn that.
"You do know that there were plenty of affordable, powerful (for the time) computers before they got squeezed out by Wintel, right? Amiga ring any bells? Atari ST? MacOS?"
The Mac of course, had its own niche and was orginally marketed as the "computer for the rest of us" which didn't position it well for corporate use.
The Amiga and Atari ST had some definite potential, but suffered from the shaky condition of the companies that produced them (ironically the Amiga was really a 3rd generation Atari design and the ST a 3rd generation Commodore design).
"Computers became affordable due to Moores Law."
Moore's Law is not about cost, but performance. Although prices dropped significantly after PC clones were introduced, the average cost of a PC hasn't changed that much over the years - just their computing power.
Although Red Hat has made contributions to Linux, the OS was fundamentally done by the time Red Hat became invovled. There may be room for one company to be successful holding the corporate hand as they venture into Linux, but as corporations become more comfortable with Linux they may begin to question if Red Hat is adding any value they should be paying for.
Slashdot Mirror
You're confusing the sales pitch with the intent. It's not as if users lobbied Congress to create IP. You can bet that those who were promoting copyright were among the first to profit from it.
Well IBM, of course, doesn't really embrace the four "freedoms". When it purchased Rational, it would no longer sell a license Visual Test and it didn't make it open source either.
Why? Because Visual Test was a low-cost alternative to other Rational testing applications.
Wake up and smell the Blue. The only IP that IBM has/will made/make open source, is the commodity stuff or stuff they can't make any money on.
"It's not dependent on the engineer's opinion, it's based on business need"
I was including business need as part of "what programmers think".
"As always, being a developer with a locked in mindset is the number one risk to your usefulness, and thus job security"
I don't consider myself in that category, however, I don't agree with your theory anyway. Your relationship with your Boss is far more significant in keeping your job than any mindset is.
"could that be because computer hardware up till now has not been able to handle that level of parallelization that you would need to realistically model the real world?"
No. It's because modeling the real world wasn't the goal in many cases. Besides there's a big difference between modeling and controlling.
It's seems to me that multiple-core systems may excel in two different areas:
1) Problems for which there is already a solution, but practical applicatons require near real-time performance and we currently can't achieve that speed (assuming the problem can be solved in parallel).
2) Problems that we understand how to solve but don't yet have a solution because they would take too long to calculate (say 20 years). A large network of multicore machines might be able to significantly shorten the time (again assuming the problem can be solved more efficiently in parallel).
On the other hand, problems that we've been working on for many years without a proven approach (e.g. AI) are unlikely to be solved any quicker.
"The real world, modeled in data, is generally not sequential in nature..."
The problem is that many functions people perform on computers have nothing to do with the real world.
The problem in AI isn't speed, it's the "I".
"It doesn't necessarily have to be linearly proportional. It just has to be greater than the performance we could gain by making more complex single cores."
Well, given that it requires additional time and money to take advantage of it, there needs to be a commensurate performance advantage. It's not enough to be merely "greater than" a single core.
Sure, in the past. The point is that they're going to offer more cores instead of faster ones.
"So whether programmers find this move acceptable or not is irrelevant because this path is probably the only way to design faster CPU:s once we've hit the nanometer wall."
I guess you should put "faster" in quotes.
In any case, it is absolutely relevant what programmers think since any performance improvements that customers actually experience is dependent on our code.
Historically a primary reason to buy a new computer is because a faster system makes legacy applications run faster. To a large extent this won't be true with a new multicore PC. So why would people buy them?
That's why Intel wants us to redesign our software - so that in the future their customers will still have a reason to buy a new PC with Intel Inside.
I understand why Intel is so interested in multiple cores - they can't make the faster single-core chips that the market wants.
The question is what's our motivation? Unless software performance is approximately linearly proportional to the number of cores (e.g. a 10 core cpu can run a software application 10x faster than it ran before it was made core-aware), it probably isn't worth converting legacy apps.
I'm not against it, I'm just saying that it's not an example of the feasibility of open source.
As for an example, check out Cast Iron systems.
You don't understand that an appliance doesn't necessarily mean an embedded system. It's just a product that performs a specialized function.
To the customer it's just a box.
You're wrong about many of the Linux appliances. They don't change anything in the OS and thus aren't obligated to provide their own source to anyone.
What they do is take a standard PC platform, install a standard Linux distro, install their proprietary software and sell the system for hundreds or thousands of dollars more than the hardware costs. Thus they make a good profit and recoup their investment in developing their software.
I guess you missed the word "average". If the only customers who bought Red Hat were customers who asked questions of that low-level type, you'd be out of business rather quickly.
Well, given the fact that McNealy did a 360 on the value of Office suites, who knows what is true?
He made a big deal in the media about handing out white boards and markers to his employees in liu of using Powerpoint, that was widely reported at the time (probably 1997 or so).
"When they bought StarOffice, it saved them a ton of money versus buying licenses for their employees."
Actually, no. Prior to buying StarOffice, Scott McNealy forbid his employees to use a word processor with more than 4 features (i.e. a text editor instead of a word processor), so there were no licenses to save money on.
Mysteriously, after Sun bought StarOffice, McNealy stopped talking about how evil office-type applications were.
I can't imagine what an average user would require for support. Do they ask questions like "Where's the start menu" or "the dir command is broken".
Surely Linux isn't so crappy that people have to buy the "extended warranty" of support.
A lot of these companies "succeed" because they were purchased in an Internet-Bubble-Mindset at absurd prices.
Jboss falls in this category and is not living up to expectations at Red Hat.
Since Sun bought mysql for 1 billion dollars, it's going to be a very long time before you could prove it was a good deal.
Anyway, I'm not going to address every example. The point is that the examples are few and far between for profitable open source companies who write their own original applications and don't bundle them with hardware.
"Of course there's evidence - just look at around at all the companies that DO base their business models on FOSS and still rake in a pretty penny."
I'll let you do the looking and get back to me.
Since companies want to make a profit, they'll always go for the low-hanging fruit.
There are many companies that make specialized Linux "appliances" without modifying Linux, without paying for Linux, and without contributing to Linux.
They don't do it because there's something about Linux that they can't get elsewhere, they do it because it's cheaper than paying for a license.
What these companies are really selling is proprietary software that could run on any Linux PC, but they bundle it with a closed box so they can charge more money for it.
They could do the same with Windows (and historically people have) but then the cost of a Windows license would come out of their profits.
I didn't say anything about a lack of profit at Red Hat, I just stated that the day may come when corporations ask themselves why they should pay for something that is available for free.
As far as credit for Linux is concerned, I'd put Thompson and Ritchie at the top of the list and Linus second. I leave the rest of the ranking to the Linux community (of which I am not a part).
My broader point, however, is that Red Hat isn't a good poster boy for the business viability of open source and in particular not an example of how MS could succeed at it.
There's a big difference making a product open source that you've already spent millions of dollars on for a decade without help (e.g MS Office), and making contributions to an existing product (however limited it may have been) that was somebody else's idea.
Does Red Hat have any brand new product ideas that aren't part of the Linux OS that they'd like to spend a few years developing and release version 1.0 under the GPL license?
"Staying relevant as in staying an industry leader in computer software and remaining competitive in the industry."
Ok, but how is this goal unmet by hiring experienced developers rather then entry level graduates?
"I already mentioned that college students who contribute to open source projects are going above and beyond their school's curriculum, which gives them experience that a student who doesn't contribute to open source project won't get. "
They could be going "above and beyond" or they could be contributing to open source projects instead of fully meeting the requirements of their school's curriculum. In the real world many of your choices end once you agree to work for someone else - open source projects might not be the best place to learn that.
"You do know that there were plenty of affordable, powerful (for the time) computers before they got squeezed out by Wintel, right? Amiga ring any bells? Atari ST? MacOS?"
The Mac of course, had its own niche and was orginally marketed as the "computer for the rest of us" which didn't position it well for corporate use.
The Amiga and Atari ST had some definite potential, but suffered from the shaky condition of the companies that produced them (ironically the Amiga was really a 3rd generation Atari design and the ST a 3rd generation Commodore design).
"Computers became affordable due to Moores Law."
Moore's Law is not about cost, but performance. Although prices dropped significantly after PC clones were introduced, the average cost of a PC hasn't changed that much over the years - just their computing power.
Although Red Hat has made contributions to Linux, the OS was fundamentally done by the time Red Hat became invovled. There may be room for one company to be successful holding the corporate hand as they venture into Linux, but as corporations become more comfortable with Linux they may begin to question if Red Hat is adding any value they should be paying for.
"They spread enough FUD that it wouldn't be hard to to convince a lot of people that free=open source"
RMS already confused the issue by calling his open source projects "free".