Slashdot Mirror
If Microsoft Went Open Source
From an Anonymous Reader: "The BBC's Bill Thompson has written a speculative article about the possibility of Microsoft attempting to secure their place in the future of operating systems by creating an open operating system. From the article: 'They allocate a billion dollars worth of programmers to shine and polish [The new OS] for a year, improving its compatibility with Windows Server technologies, donating parts of the Windows and Office code bases under the GPL and turning it into the world's best operating system.' Could this ever happen?
Microsoft's role shouldn't be in improving the OS, it should be in creating the infrastructure necessary to allow the umpteen-zillion Windows developers out there to improve the OS instead.
I don't know how many of you have contributed to an OSS project, but, at least for those projects that are well-established the process can be a lot of work and not a little bit intimidating. Some progress has been made on the tool front to make it easier but it still takes way too much effort to get a patch mainstreamed on the really big projects.
What Microsoft should do is open up their software, and invest their money in more programmers, but not to do coding, to act as support for the rest of us who do the coding.
Make it so that if I find a bug, all I have to do is fix it and submit a patch. That's it. Nothing more. Nothing less.
This is the one opportunity they have that I don't see Linux/*BSD ever possessing. The kind of work necessary to support large projects is the very last thing most of us want to do. Sourceforge is littered with the remains of OSS projects that were fun to code and get working, but that nobody wants to maintain anymore.
They'd still make gobs of money. Ever browse their help wanted section? Sometimes it seems as if half the listings there are for build engineers. Guys whose only job it is to build Windows and all the other projects. Casual/notive users are never going to attempt this on their own (Gentoo/LFS users notwithstanding), and you'd be crazy to accept builds from third-parties given the complexity we're talking about and the potential for malware.
It's the best thing Microsoft could do right now. Which is why they won't do it. It's like what they say about generals always fighting the last war. Gates and Ballmer got where they are by hewing to a specific ideology. They're not changing their minds in this lifetime or the next, even if its clear that that ideology is antiquated and obsolete.
--
Why didn't you know?
I believe what he is suggesting is that Microsoft spend a billion bucks and a year to embrace and extend Linux, starting from some existing distribution. Then when they release their flood of changes in a year, under the GPL, no one will be able to catch up because of that billion buck one year lead.
But that one year lag works the other way too. Microsoft would then be a year behind the open source baseline with which they started.
If they kept merging mainline changes into their internal codeset during that year of secret development, it would no longer have a year's worth of changes in it, it would only have enhancements, which would be a lot easier to pick and choose from for the rest of the world to merge back into the mainline.
If Microsoft kept their baseline "pure", they would be behind the world as much as the world would be behind them. If they kept their internal codeset up to date, they would not be a year ahead.
Wham! Paradox City Arizona, baby.
Infuriate left and right
Releasing anything resembling the source code to windows would be laden with problems for Microsoft. Opening their customers to a whole range of security holes created by decades of patch-fixes and arcane support layers for retired API's would possibly leave them with a public relations disaster on their hands, not to mention the financial repercussions.
However, it is interesting to imagine a truly level playing field between Windows & Unix based operating systems, in freedom and price terms. Would end users choose unix based systems over windows based systems given the full freedom of choice and knowledge that applications could run on either? Also the possibilities for code and standards interaction between two entirely open systems and the continued improvement of both in competetive and meaningful ways is something that could potentially be extremely beneficial to the computing ecosystem at large
Business Voyeur
but hey, it'd be nice if Microsoft did it too. I like UNIX ;-)
This isn't a speculative article. It's a quiet attempt by Microsoft to gauge the community's reaction to a possible open source product.
Recently I was paid $10 to take a survey geared towards IT professionals about "current trends within the Software and PC Industry". The questions were clearly written by Microsoft, and one possible plan was obvious:
-Microsoft will compose a list of dozens of software patents allegedly violated by Linux and will offer total indemnification for Red Hat users only. If necessary, it will use its own patent portfolio as leverage.
-Microsoft will strengthen Red Hat's source offerings to emphasize "interoperability", which means that it will be possible to administer a RH install from Windows.
-Microsoft will buy Red Hat for considerably more than it seems to be worth and will immediately cripple it just as it's crippled every other worthy competitor it has bought out.
This is a clever plan to defeat Linux.
(Part of the survey really bugged me because it seemed like a push poll - see here.)
We recently had heard in the office over one of the Yellow Machine that's made by Anthology Solutions.
Yes. If Windows were open sourced, it would be incorporated into Wine. In very short order, every legacy Windows application would run seamlessly under Linux and Windows as a stand alone operating system would simply fade away. So... another good (from MSFT's perspective at least) reason for them not to do so.
Portability. If MS were to move Office to X under MS-Redhat (or whatever), that would mean it should be possible to get Office running under Gentoo (which isn't the kind of lock-in MS would like). So they'd have to do something like make a special toolkit (which they would probably do anyways). But that toolkit would have to use X, so it could still be put on Gentoo. So they'd have to change X. That means either writing their own X server or adding patches to the existing one. If the patch it, they have to release it so that won't work unless you need their special kernel stuff. But they'd have to release that too (it couldn't be a module, so it'd have to be GPL). In the end, anyone with Gentoo (or whatever) and some time should be able to run the program that would run only on MS-Redhat.
The only way to fix it is patching the kernel or X, and then they'd have to release code. The other option is to write their own kernel/X from scratch... but that's what Longhorn is supposed to be (a complete rewrite). So... why bother?
Given the way MS operates, it doesn't make sense. Now to provide a better Unix on Windows environment (like better POSIX compliance, a version of BASH, etc) in the form of a good Services for Unix so that applications that are cross-platform can be run easily on Windows, that could help them (making it easy to run Unix/Linux/BSD programs on Windows opening up tons of applications and such). Out of the two, that would be FAR more likely.
But I doubt that would happen, because to allow people to easily port Unix stuff would mean allowing people to write Unix stuff and trade in their Windows servers down the line for Linux. To make it easier to keep running the platform that way would make it easier to switch off it. So it won't happen, it will stay crippled.
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
What 'open source product' did Apple take in? They were acquired by (or they acquired, depends on how you look at it) NeXT, who had a closed-source operating system. They essentially 'open sourced' big chunks of it, enough to run a 'bare UNIX-like OS' which has been called Darwin. As part of making it a bare UNIX-like OS that would be USABLE they grafted on a FreeBSD derived userland.
In no sense of the word did they 'take an open source product' and kinda create their own fork. Unless you can tell me where to download NeXT's Source Code. I wouldn't mind having NextStep/OpenStep to run on some of the various hardware (PA-RISC, Sparc, Intel, etc.) hardware I have around here. . . It's not freely available by any means except the warez route. Certainly the source code is not available.
Microsoft still suffers from the "All Star" syndrome. Hire the best people money can buy, and the rest will take care of itself. Sorry folks, it doesn't work this way. The most productive teams I have ever worked in have consisted of the most gifted, and the most brain dead, with a generous distribution in between. You need a broad view, those who can see universe, the sky, and the ground below us. All make a substantial contribution to a truly great product.
"To those who are overly cautious, everything is impossible. "
I come here to read news, not some editorial guesses at what might be news in the future
/.?
You must not come here often!
There are plenty of tech sites out there if you just want news, and most deliver fresher right-off-the-presses news. But here you get news, editorials, book reviews, interviews, and tons of sometimes funny, sometimes insightful member commentary.
so why settle for vanilla, when you can get it and more in the neopolitan that is
If it were GPL, I wouldn't care...as long as it was an official release from the MS corp. A release from a team of their engineers would leave me coldly skeptical. I would be expecting that at some point MS, the corp, would swoop down with a bunch of concealed patents, and start suing everyone they didn't like for patent infringement.
They haven't earned much in the way of trust.
I think we've pushed this "anyone can grow up to be president" thing too far.
I can't help but wonder why they would do what the author of the article thinks they should do. There's no reason they would have to fork Linux and open their own code, if they wanted to do something like this. They would simply have to take the Linux kernel, port their own window manager and development tools and desktop environment to it (you know, the one everyone's used to and is the reason they can't switch), and get programs running. They could do that without too much trouble. Run Apple-style emulation layers if you have to. Fat binaries, perhaps, that run on Window with NT kernel and Windows with Linux kernel?
... everything.
They would be able to keep their own code closed, since they wouldn't have to alter the Linux kernel, and they would be able to update the OS with Linux kernel upgrades as they happen. Whenever they make a fix to the kernel, it goes back to the community, yes. But they do it because it benefits them to make the fix. The fact that everyone else benefits from their fix should be a good thing for them. When they make a patch that improves security or stability for everyone, well, you just can't buy that kind of good PR.
It might work better in a legal sense if they did this with FreeBSD, just as Apple did. And that's how they can beat Apple. Do the same thing, with the same kernel baseline, but rely on their massive resources and programming ability to outpace them on the UI and applications front, meanwhile benefiting from every addition Apple contributes to BSD.
Then Windows is UNIX, and there would be no reason not to use it. They would win the desktop, the server, the handheld
Damn it.
Lack of eloquence does not denote lack of intelligence, though they often coincide.
But, were I Microsoft, I could think of ways to leverage the Linux development progress cheaply and easily, and piss off all the OSS people all at once.
First, MS should buy Transgaming. They own Cedega, which is a closed fork of the Wine tree. No need to support the WINE project with actual patches, since there's no licensing requirements.
Second, knock together, say, a FreeBSD or Linux distribution. X11, standard userland, everything.
Third, use their internal OS programmers to turn Cedega into the greatest thing since sliced bread. A -perfect- implementation of the Win32 API on top of Linux.
Fourth, get all the hardware manufacturers on board for drivers. Institute a driver program. Ta-da, everyone has drivers, but only on platforms MS wants to support. IE, x86. OSS driver development continues, but at a slower pace with fewer people actively testing.
Fifth, make the install as painless as a standard Windows install. No text-mode, no kernel boot stuff, just the splash we all know and love(/hate)
Fifth, sell for the price of a Windows license, or a little less. Allow the base OS to be downloaded freely, ala Darwin, but keep the WINE/Win32 API closed and sealed off.
Since their Win32 API is perfect, Visual Studio should run flawlessly. AND, with the proper window manager on X11 (as they will likely do this), it would be visually indistinguishable from standard Windows. Power-users could install Gnome/KDE/fluxbox/windowmaker/whatever, and the Win32 API would still be perfectly available, exportable over the network as any X11 app, etc.
Leverage the community to build the kernel and userland. Use their own people to maintain just the API - keep the total lock-in.
You're assuming Red Hat is the fundamental player of the free software world. It is not. There's not such a thing, and that's why free software is here to stay: No one owns it. No one can frigging own it. If tomorrow Red Hat disappears, there's a dozen companies with the same potential that will fill the gap. They just have less US market share at the moment, but that means nothing.
To do list for Windows
We already have non-us in debian. As of today software patents are not valid in Europe, so i'd like to see MS try messing here.
As for the USA, you need to deal away with software patents.
It takes a man to suffer ignorance and smile
Be yourself no matter what they say
Microsoft bought Interix years ago. It's now available in the form of Microsoft Windows Services For Unix (SFU) currently version 3.5.
It is decent - I'd probably give it 3 stars (out of a possible 5). What makes it decent (in my opinion) is the porting of NetBSD's pkgsrc (http://pkgsrc.org/ to the Interix toolchain.
This means you can compile your own shit (OpenSSH, bash, etc.) from the pkgsrc collection. For those of you who are unfamiliar with pkgsrc, think Gentoo's portage.
This is where it gets really interesting. Daniel Robbins (of Gentoo fame) as we all know - now an employee of MS. Microsoft could do something really cool here if they "shared sourced" SFU as a standalone kernel and OS. They have everything in place to do it. Even if they don't open it up, at least their users might get a reliable, kick-ass OS for once.
Just a thought. I think it's (only slightly) more likely than what TFA's author imagines.
Is it me or are people forgetting that Microsoft is a public NASDAQ company? People will always rant at the industry leader but google manages to avoid the publicp pressure due to their enthusiasm and indirect methods of aquiring profit.
I doubt one year/billion would do it. How much have they spent so far to create the worst?
I for one, having used Office at the office, want nothing to do with its code base beyond a compatibility library for a different product whose stability, UI, performance, etc, aren't an example of how bad software can be and still sell with the right name on the box.
Because the name isn't the important part. If you want to have such a reaction, you must do something significant to provoke it, and revealing the most superficial thing about the OS, its name, half of which was obvious already, that's not going to be enough.
That's admirable! If my coding skills were better I'd take over a couple of stagnent Sourceforge projects.
I wish someone would take over cinelerra. I haven't been able to get recent builds to compile, and previous builds (when it did compile) I spent ~2 hours tracking down obscure source packages they were linking to, and several more hours compiling and fixing THOSE before I could get cinelerra compiled. Once the program is compiled, it runs great but the current maintainer broke the package.
In the meantime I'm learning Broadcast 2000 for personal projects, but when it comes to projects for client I'll have to boot Windows and go back to MediaStudio Pro (Broadcast 2000 is FAR too picky about what video formats it will import)
On that note: hey Ulead! Port MediaStudio Pro to Linux and I'll buy it!