"guaranteed to be available on 100% of Windows systems (sic)"
1) Win 3.1 is not supported by WinXP. 2) Just because it is available, doesn't mean it will work/correctly/. Win 9x and ME are emulated within XP, so parts of their functionality don't work. Also, there must be a reason why the "application compatibility wizard" exists.
Note: Linux does have problems with things working properly, too.
"A set of APIs and an ABI for writing graphical programs"
You got me here, I must say. Linux doesn't have any stable graphical APIs and ABIs.
Oh, wait. I forgot. There are several that run on it though. GNUStep, Qt & KDE Libs, Gtk/2, and (if you are insane) you can even use raw X11. (And the spec for that hasn't changed, and is still supported, for over 15 years).
Now, it is true that there have been several changes within many of the APIs listed above. However, whenever they break binary compatibility, the major release number is changed. And -- get this -- you can run several versions of the same libraries on the same system. And you don't have to go through DLL HELL to locate the right one. Imagine that. So, yeah, the support is there...
I'm not saying that Linux is perfect; I'm not even saying it's better. I am saying that you really should know what you are talking about before speaking.
And, as long as we are in the neighborhood...
Microsoft as well.
Anything which could spark a war between zealots.
Ok. Now that I said it... have at me.
Not to mention the "articles early for subscribers". Now, I understand they need to support their bandwidth and server costs, but they should not do so at the expense of their viewers.
If they produced news of their own, it may be a different case. Like LWN... News is free to read and post on. Their *own* articles/editorials/bulletins are available a week earlier to subscribers.
Don't forget nVidia. Your drivers are sweeter than honey... but making them OSS means that when there is a driver API change. I can run the new kernels (have my cake) and use your drivers (and eat it too).
"Tux and Clippy playing ring around the posie (sic)"
http://en.wikipedia.org/wiki/Ring_around_the_rosie
hehe. If the "Black Death" Interpretation is correct... that would be very fitting as an "end/death of the universe" analogy.
Quoth brontus3927: "One/.er made a snide remark about a future with a MS Linux distro. The chances of that aren't great, but I would love to see it happen."
Well, you should investigate how KDE's versioning system works. Different numbers updating mean completely different things.
(For the sake of this post, the versioning is interpreted as Major.Minor.Release)
When the major number changes, it means that there has been an interface change and breaks interoperability between applications. Applications written for a 3.x release will not work with a 4.x release. New features are added in a major release, and bugs are fixed.
When the minor number changes, it means that new features added. These features will not break the API. Bugs are also fixed.
When the release number changes, it means that bugs have been fixed. (E.g. The release numbers are maintenance releases.)
Quoth FatherOfONe: "The other things they would loose would be: 1. The ability to leverage their OS to push more Office sales (~30-40% of their income). 2. The ability to leverage their office app to push more desktop sales. 3. The ability to hinder their competiton by not disclosing their API's."
1 & 2) They, instead, could leverage their distro to push their office sales. And vice-versa. Nothing is preventing them from creating their own DE or WM. (Albeit most of us probably think that it would be detrimental for them to do so, they do already.)
However, one of the reasons that everyone is ticked with MS is because they do this. Instead, MS could drop that act, and offer products that sell because they are good products. It would save them from alot of investigations from the SEC, and would make their customers less disgruntled.
3) Most of the MS APIs that are useful are exposed to developers. The implementing code isn't, but that is another matter entirely. All you need to know is where to look. (Help files and MSDN). Even the APIs that they choose not to expose are usually released in later versions of their code. (For instance, most of the usability improvements in Office 2003 are standard in Avalon, the windowing API for Longhorn).
Furthermore, some interfaces in the open source world are available to developers, but have been marked as "no touchie". They are usually:
A) very low level (e.g. kernel function calls),
B) Interfaces that are likely to change (again, mostly kernel function calls), or
C) Internal interfaces (particularly internal functions in.so's, but also a large portion of the STL and other libraries).
Now, people can still look at them because of the nature of open source. However, this can also motivate them to use the libraries, just because they exist. This eventually causes problems down the line when the interface which was marked "do not use" has to be changed, or is removed, etc. *Sometimes* closed source can be a good thing.
Now, don't get me wrong. I am an OSS nut. I am practically as large as they come. However, I do really love some parts of MS, and I think that if MS were to provide such services on linux, that it would be better for the market overall. It'd definitely be better for me. And I wouldn't mind paying for the closed-source solution, if it did things that benefited me in the long run.
"What I would love to see is the Playstation3 and or the next Nintendo system to use Linux as the OS."
Heh. This is about as likely to happen (in the manner I read) as MS going to linux. If it did happen, the effect would be very similar to how Palm is managing their next version of PalmOS. (And, come to think of it, the way I am suggesting MS switch markets).
E.G. The device would use linux, and expose the OSS libaries, but everything that matters would be built on top as a proprietary, closed-source solution. For consoles that use specialty hardware (e.g the PS2 and bretheren), special compilers could be required. Sony would be under no pressure to release the specs for their archetecture, let alone the source code to their compiler.
For the sake of argument, say the compiler was based off of a GPL'd application (such as GCC). They would only need to distribute the source code for the compiler when it is distributed. And if the price to obtain a copy of the SDK is significant (which it usually is), game developers may not want to give it out after they have a copy. It is actually in their best interests to not distribute it, as distribution equates to more market competition. As far as the SDKs go, there is no requirement that they be OSS at all.
Most of the income generated in the console manufacturing market is made by selling the SDKs to developers for games. The console market is as proprietary and lock-in as MS Windows.
"Lastly, I need to say that perhaps OpenGL and Alsa have no "gaps", I honestly don't know, but it appears that most deve
After reviewing the SEC filings for Microsoft (Year 2004, link: http://sec.gov/Archives/edgar/data/789019/00011931 2504150689/d10k.htm )
It is true that OS Sales make up a larger portion of their income than I thought they did. However, if MS were to play their cards right, they could just as easially market a Linux desktop distribution that was sold for the same amount as their OEM sales. Since most of their OS income is based on OEM sales, they shouldn't be loosing too much money, even if they made the boxed version the same price.
Hell, they could get a large audience just becase it's their next product (e.g. be less than honest, and don't even mention that it is Linux). I think it would be in their best interests to let it be known. Maybe a year after the product comes out. (As far as the GPL inclusion goes, most people don't read the license provided by the software. I'm willing to bet someone would pick up on it, though.)
They could even, if they really wanted to, make it so that in order to run windows programs (or, at least the old ones) you need to pick up their product. (Basically, do the same thing that Mac did with OSX).
However, you are wrong when you state that the OS sales make up over half their income. According to their SEC 10-K (linked to above) their revenue from OS sales is roughly $11.5 Billion in 2004. Their revenue from Office/Connectivity sales is (again, roughly) $10.8 billion in the same year. I would hardly call $700 million a dwarfing figure in the context of MS's income. This is the main area that I was talking about in my post.
The 3rd largest segment is their server market, which includes Exchange, MS-SQL, Windows Server 2003, etc. The total contribution to revenue from server sales was about $8.4 billion. They would take a hit in this area as well. However, I think that the loss would largely be mitigated (and IANA MBA/Economist/etc) because they could have products based on top of that, just like IBM, Novell, Oracle, etc do.
These products could provide the familiar "Windows way" of doing things, but they would be built over a more secure system. In addition, they should also still offer their service products (Exchange, MS SQL, IIS, etc). These products are actually worthwhile to use. However, the underlying "legacy software" (e.g. the kernel and windowing system) should be changed.
I would bring up other areas that were outlined within the 10-K, however their contribution is negligible in comparison with what has already been outlined. Furthermore, most (if not all) of the other revenue generation (Mostly XBox sales and MSN) would successfully bridge the gap, as long as the systems noted above cross over successfully.
DISCLAIMER: I know that Microsoft would rather pay for a freezer large enough to freeze hell over than let this happen, but hear me out.
I consider the obvious tissue for Microsoft's tears to be the embrace rather than destruction of Linux. They should develop their technologies (such as groupware and PIM suites, along with DirectX) to cooperate with and run on it. They would line thier pockets with what they do best. At the same time, they discard the flak caused from buggy implementations of these same systems.
Think of the possibilites:
Microsoft wouldn't have to worry about providing massive security features (and updates) for an OS. It would be handled predominantly by kernel maintainers and such. When Microsoft does implement a fix or a feature upgrade, their products can reach a wider audience as they forward it upstream.
Linux users would finally be able to use a fully implemented DirectX, and all of the goodies that come with it. On the flip side, Microsoft would benefit because DirectX would be built on a better network, threading, and paging archetecture.
This doesn't include the possibility of opening the DirectX code base to an open-source license (probably similarly to the CDDL or MPL, as MS would still want to keep an eye on their code). That would allow open-source coders (or possibly even game designers) to fix bugs and add features of their own. It would also facilitate bindings to other languages, meaning the libraries are used even more. And as Microsoft has known for a long time, ubiquity is A Good Thing (tm).
Microsoft could expand their ease-of-use/GUI mentality to other tasks, thereby making it that much easier to use Linux. They would recieve the configurability and power of the underlying systems for free. For instance, they could create *good* system management tools to configure firewalls, QoS, Quota Management, Network FSs, package/software management, etc. In effect, they may create their own distro that would come with these apps, and market that just like they do Windows. It would be similar to what the "big 3" (RH, SuSE, Mandrake) do. They could gain some control over the systems they like best by:
including them in the distro, and
providing patches that do what they want.
Windows does have many of these things now, but it is MS's responsibility to ensure that everything is handled: from security to feature additions. This includes "legacy software" (such as kernels and web browsers), which are not particularly profitable in and of themselves. (In my experience, the only way MS has been able to make legacy systems profitable is through their vendor lock-in model). With Linux, they could all but remove this aspect of their business. Granted, they would take a hit in OS sales, but that isn't their cash cow anyway.
The distilled message is that MS could add all of thier eye-candy,.net integration, etc. to Linux and have a better product overall, and they don't have to lose complete control of their market to do so. And, for the first time in years, people could be pleasantly surprised with a MS product.
Of course, Microsoft definitely has something standing in the way of doing this. They got locked into their own proprietary licensing and code. It would cost a significant amount of money to refactor their entire code-base to run on Linux.
Slashdot Mirror
Ok. It will run on x86... but that hasn't been released yet.
So, I ammend my previous statement:
"This be nitpicking, but the only *publicly released* version of OSX runs on PPC"
"guaranteed to be available on 100% of Windows systems (sic)"
/correctly/. Win 9x and ME are emulated within XP, so parts of their functionality don't work. Also, there must be a reason why the "application compatibility wizard" exists.
1) Win 3.1 is not supported by WinXP.
2) Just because it is available, doesn't mean it will work
Note: Linux does have problems with things working properly, too.
"A set of APIs and an ABI for writing graphical programs"
You got me here, I must say. Linux doesn't have any stable graphical APIs and ABIs.
Oh, wait. I forgot. There are several that run on it though. GNUStep, Qt & KDE Libs, Gtk/2, and (if you are insane) you can even use raw X11. (And the spec for that hasn't changed, and is still supported, for over 15 years).
Now, it is true that there have been several changes within many of the APIs listed above. However, whenever they break binary compatibility, the major release number is changed. And -- get this -- you can run several versions of the same libraries on the same system. And you don't have to go through DLL HELL to locate the right one. Imagine that. So, yeah, the support is there...
I'm not saying that Linux is perfect; I'm not even saying it's better. I am saying that you really should know what you are talking about before speaking.
This be nitpicking, but OSX only runs on PPC... not x86.
Hmm.. I see a joke RFC (or the likes) in the making.
And, as long as we are in the neighborhood... Microsoft as well. Anything which could spark a war between zealots. Ok. Now that I said it... have at me.
Not to mention the "articles early for subscribers". Now, I understand they need to support their bandwidth and server costs, but they should not do so at the expense of their viewers. If they produced news of their own, it may be a different case. Like LWN... News is free to read and post on. Their *own* articles/editorials/bulletins are available a week earlier to subscribers.
Don't forget nVidia. Your drivers are sweeter than honey... but making them OSS means that when there is a driver API change. I can run the new kernels (have my cake) and use your drivers (and eat it too).
Ditto here, except that I am /in/ college.
;P
I pulled out a highlighter just to pluck out the tidbits I thought I'd need. Now I have all kinds of hot pink and yellow on my screen
"Tux and Clippy playing ring around the posie (sic)" http://en.wikipedia.org/wiki/Ring_around_the_rosie
hehe. If the "Black Death" Interpretation is correct... that would be very fitting as an "end/death of the universe" analogy.
Quoth brontus3927: /.er made a snide remark about a future with a MS Linux distro. The chances of that aren't great, but I would love to see it happen."
2 11079
"One
I was going more for endearing when I said it...
http://slashdot.org/comments.pl?sid=134244&cid=11
Well, you should investigate how KDE's versioning system works. Different numbers updating mean completely different things.
(For the sake of this post, the versioning is interpreted as Major.Minor.Release)
When the major number changes, it means that there has been an interface change and breaks interoperability between applications. Applications written for a 3.x release will not work with a 4.x release. New features are added in a major release, and bugs are fixed.
When the minor number changes, it means that new features added. These features will not break the API. Bugs are also fixed.
When the release number changes, it means that bugs have been fixed. (E.g. The release numbers are maintenance releases.)
Quoth FatherOfONe:
.so's, but also a large portion of the STL and other libraries).
"The other things they would loose would be:
1. The ability to leverage their OS to push more Office sales (~30-40% of their income).
2. The ability to leverage their office app to push more desktop sales.
3. The ability to hinder their competiton by not disclosing their API's."
1 & 2) They, instead, could leverage their distro to push their office sales. And vice-versa. Nothing is preventing them from creating their own DE or WM. (Albeit most of us probably think that it would be detrimental for them to do so, they do already.)
However, one of the reasons that everyone is ticked with MS is because they do this. Instead, MS could drop that act, and offer products that sell because they are good products. It would save them from alot of investigations from the SEC, and would make their customers less disgruntled.
3) Most of the MS APIs that are useful are exposed to developers. The implementing code isn't, but that is another matter entirely. All you need to know is where to look. (Help files and MSDN). Even the APIs that they choose not to expose are usually released in later versions of their code. (For instance, most of the usability improvements in Office 2003 are standard in Avalon, the windowing API for Longhorn).
Furthermore, some interfaces in the open source world are available to developers, but have been marked as "no touchie". They are usually:
A) very low level (e.g. kernel function calls),
B) Interfaces that are likely to change (again, mostly kernel function calls), or
C) Internal interfaces (particularly internal functions in
Now, people can still look at them because of the nature of open source. However, this can also motivate them to use the libraries, just because they exist. This eventually causes problems down the line when the interface which was marked "do not use" has to be changed, or is removed, etc. *Sometimes* closed source can be a good thing.
Now, don't get me wrong. I am an OSS nut. I am practically as large as they come. However, I do really love some parts of MS, and I think that if MS were to provide such services on linux, that it would be better for the market overall. It'd definitely be better for me. And I wouldn't mind paying for the closed-source solution, if it did things that benefited me in the long run.
"What I would love to see is the Playstation3 and or the next Nintendo system to use Linux as the OS."
Heh. This is about as likely to happen (in the manner I read) as MS going to linux. If it did happen, the effect would be very similar to how Palm is managing their next version of PalmOS. (And, come to think of it, the way I am suggesting MS switch markets).
E.G. The device would use linux, and expose the OSS libaries, but everything that matters would be built on top as a proprietary, closed-source solution. For consoles that use specialty hardware (e.g the PS2 and bretheren), special compilers could be required. Sony would be under no pressure to release the specs for their archetecture, let alone the source code to their compiler.
For the sake of argument, say the compiler was based off of a GPL'd application (such as GCC). They would only need to distribute the source code for the compiler when it is distributed. And if the price to obtain a copy of the SDK is significant (which it usually is), game developers may not want to give it out after they have a copy. It is actually in their best interests to not distribute it, as distribution equates to more market competition. As far as the SDKs go, there is no requirement that they be OSS at all.
Most of the income generated in the console manufacturing market is made by selling the SDKs to developers for games. The console market is as proprietary and lock-in as MS Windows.
"Lastly, I need to say that perhaps OpenGL and Alsa have no "gaps", I honestly don't know, but it appears that most deve
After reviewing the SEC filings for Microsoft (Year 2004, link: http://sec.gov/Archives/edgar/data/789019/00011931 2504150689/d10k.htm )
It is true that OS Sales make up a larger portion of their income than I thought they did. However, if MS were to play their cards right, they could just as easially market a Linux desktop distribution that was sold for the same amount as their OEM sales. Since most of their OS income is based on OEM sales, they shouldn't be loosing too much money, even if they made the boxed version the same price.
Hell, they could get a large audience just becase it's their next product (e.g. be less than honest, and don't even mention that it is Linux). I think it would be in their best interests to let it be known. Maybe a year after the product comes out. (As far as the GPL inclusion goes, most people don't read the license provided by the software. I'm willing to bet someone would pick up on it, though.)
They could even, if they really wanted to, make it so that in order to run windows programs (or, at least the old ones) you need to pick up their product. (Basically, do the same thing that Mac did with OSX).
However, you are wrong when you state that the OS sales make up over half their income. According to their SEC 10-K (linked to above) their revenue from OS sales is roughly $11.5 Billion in 2004. Their revenue from Office/Connectivity sales is (again, roughly) $10.8 billion in the same year. I would hardly call $700 million a dwarfing figure in the context of MS's income. This is the main area that I was talking about in my post.
The 3rd largest segment is their server market, which includes Exchange, MS-SQL, Windows Server 2003, etc. The total contribution to revenue from server sales was about $8.4 billion. They would take a hit in this area as well. However, I think that the loss would largely be mitigated (and IANA MBA/Economist/etc) because they could have products based on top of that, just like IBM, Novell, Oracle, etc do.
These products could provide the familiar "Windows way" of doing things, but they would be built over a more secure system. In addition, they should also still offer their service products (Exchange, MS SQL, IIS, etc). These products are actually worthwhile to use. However, the underlying "legacy software" (e.g. the kernel and windowing system) should be changed.
I would bring up other areas that were outlined within the 10-K, however their contribution is negligible in comparison with what has already been outlined. Furthermore, most (if not all) of the other revenue generation (Mostly XBox sales and MSN) would successfully bridge the gap, as long as the systems noted above cross over successfully.
DISCLAIMER: I know that Microsoft would rather pay for a freezer large enough to freeze hell over than let this happen, but hear me out.
I consider the obvious tissue for Microsoft's tears to be the embrace rather than destruction of Linux. They should develop their technologies (such as groupware and PIM suites, along with DirectX) to cooperate with and run on it. They would line thier pockets with what they do best. At the same time, they discard the flak caused from buggy implementations of these same systems.
Think of the possibilites:
The distilled message is that MS could add all of thier eye-candy, .net integration, etc. to Linux and have a better product overall, and they don't have to lose complete control of their market to do so. And, for the first time in years, people could be pleasantly surprised with a MS product.
Of course, Microsoft definitely has something standing in the way of doing this. They got locked into their own proprietary licensing and code. It would cost a significant amount of money to refactor their entire code-base to run on Linux.
Ah well. Man can dream.
~ coolGuyZak