Microsoft in 2008 365
r.jimenezz writes "Over at Wired there's an entertaining article written by Gary Wolf. It purports to be a memo written by a 2008-Microsoft-employed Linus Torvalds to Bill, arguing against Steve Ballmer's desire to go back to the untenable OS monopoly proposition instead of the 'new order': Windows is now some sort of desktop environment on top of an open OS!"
Re:far away ... (Score:4, Insightful)
This IS entertaining (Score:5, Insightful)
You never made me alter my goal, which was world domination for Linux. I'll never forget your line: "Come on, Linus, infect the mothership." I still believe that was the best recruiting pitch ever uttered. We both took a lot of criticism from our partisans, but look what we've accomplished.
Inflect the mothership? Just writing this makes me chuckle. Seems kind of creepy, and dare I say, 'borgish'. Oh well, I suppose getting co-opted by Mothership Microsoft had somehow warped the psuedo-Torvald's mind.
Re:BULL!!! (Score:3, Insightful)
Make a point (Score:3, Insightful)
[...]But Firefox taught people that you could replace pieces of the Windows desktop with open source software. That was a crack in the seamless facade.
Re:Go back to coding! (Score:4, Insightful)
Creative writing is interesting and the thoughts behind the piece were definitly thought provoking.
Microsoft 2000-whatever (Score:5, Insightful)
Sounds like a pretty good idea to me... (Score:3, Insightful)
There are presently efforts to dump X11 in favor of a more hardware direct interface for graphics and such in order to provide more speed and flexibility. I don't know where those projects are now, but without a big backer of the idea, getting rid of X will never happen. As far as I can see, asside from some Microsoft-blessed system services, that's what I imagine WinX would be anyway. And to run proprietary code on top of a Linux kernel? I don't see any violations, legal or moral.
with as much work and progress that has been made over the years with KDE and GNOME projects, it would be far kinder to the users if there were a strong and unified user interface from which to run their applications. It freaks people out to change and learn new things. KDE and GNOME folks have done a lot of work to get their projects into the lime light but frankly, a large player like Microsoft could easily swoop in and make it all irrelevant. This may not be the case in a year or two but it feels like it is the case right now.
For the record, I'm very anti-microsoft. But it would be a mistake to fail to embrace them if they were to attempt something like WinX. (If they did, it'd probably be a BSD kernel though... worked for Apple didn't it?)
Happy Halloween! (Score:5, Insightful)
Re:Interesting thing is... (Score:5, Insightful)
As an old OS2 user, I find Windows in every incarnation to just be a stunted imitation of the WPS, which was a truly awesome GUI still light years ahead of XP. I don't see anything particularly innovative or easy about the Windows XP interface. I deal with a lot of plain users, and half of them panic a little when you ask them "Now please press the Start button".
Most of the time their more-knowledgeable son/daughter/brother/friend/neighbor puts icons to Internet Exploder and Outlook Distress, along with the photo software and Solitaire on the desktop and people never look one bit further.
The only place where Windows has something of an edge is in software installation, and the problem there is that there's a number of different systems depending on distro. I mainly use Slackware, but when I played with Mandrake, it's package GUI installation system was every bit as easy as Windows XP.
So I can't really understand what people mean when they say "Windows is easier to use". I mean, what parts of Windows does the average user plug into on anything approaching a regular basis that's more involved than double-clicking on the program.
I'll place a bet that if you put a Linux box running KDE with Firefox and Thunderbird on the desktop as Internet Explorer and Outlook Express, many users will probably be just fine. It's the product recognition thing that gets most users.
Re:Wha...? (Score:3, Insightful)
Closed desktop on top of open OS may not be as far fetched as you'd think. Apple has done exactly that with OSX and not like MS never stole ideas from Apple (which is not specific to MS either)....
-Em
Mediocrity (Score:5, Insightful)
What makes you think that Linus would solve this problem? In all seriousness, look at the "stable" 2.6 kernel branch, and the attitude demonstrated by comments like "some kernels will be good, others will be bad... we'll find out which kernels are broken soon enough".
I'm not saying that Linus himself believes in such mediocrity; but it's a bit unreasonable to expect that he would improve things at Microsoft when Linux, under his "benevolent dictatoriship" is plagued by exactly the same problems.
Re:Sounds like a pretty good idea to me... (Score:2, Insightful)
http://www.xandros.com/images/screenshots/v3/desk
Re:I read this and found it to be terribly funny (Score:1, Insightful)
Believable (Score:5, Insightful)
Alternatively, now that Oracle has bought Peoplesoft, Oracle is vulnerable. It hasn't the money left to resist an attack from Microsoft. With Microsoft wanting more of the server market, taking over companies dealing in high-end server software would be not only logical but consistant with Microsoft's tactics in the past.
A third possibility would be for Microsoft to buy part of the Internet backbone, or one of the suppliers of it. Juniper is growing in popularity but isn't so big as to be able to resist a buyout. Cisco's not been doing too great, recently, and may be vulnerable. Lucent would be easy pickings and may even welcome such a move.
Finally, Microsoft may opt for a "strategic partnership" with Boeing. Boeing is in the middle of a massive struggle with Airbus, and it's unlikely both can survive. If Boeing wants to win, it needs more money. Microsoft doubled its profits last quarter, even after allowing for the shareholder payout AND the record EU fine. Aircraft may soon have WIFI. If Microsoft can become the only vendor who can work with such WIFI points, they'd have absolute control of the business market.
Finally, Microsoft could buy a hard drive vendor. If the OS came pre-installed on the hard drive to OEMs, then fewer OEMs would be willing to install rival Operating Systems....
Linus never would have written this memo. (Score:3, Insightful)
Second, it isn't nearly as witty as Linus is; it doesn't have any of the insults-that-make-you-feel-like-thanking-Linus-fo
Third and most vital, Linus doesn't give a damn about any of the crap the author's writing about. He doesn't care about taking over the world or marketing. He is only interested in technology.
Re:Wha...? (Score:3, Insightful)
The problem with Windows isn't the OS (i.e. the kernel-mode modules), it's the Win32 API, which was designed in a rush in the early 1990s, when Microsoft fell out with IBM and dumped the OS/2 API (which was pretty bad too). In addition to having been designed in a rush, Win32 was made similar to Win16, so that 16-bit Windows developers could easily port their software to NT. That meant its designers had to loosely follow an API designed in the 1980s, which was never a particularly good design, but had become very popular (for various reasons).
The strength of Win32 was in moving existing Windows developers from DOS/Win3.x/Win9x to NT, and in that respect it was a very good move by Microsoft. However, it's become something of a liability over time, since it limits Microsoft's flexibility to take advantange of all the advances since the late 1980s and early 1990s.
Microsoft could replace Win32 with a new subsystem/API, which would sort of be following Apple's lead in replacing the classic Mac OS API with Cocoa (and Carbon for directly porting old applications). However, Microsoft would be crazy to replace the NT kernel with something else, since NT was designed from the beginning to support multiple OS personalities, is amongst the best production kernels around and has unparalleled device support.
Unfortunately, Microsoft's current efforts seem to be based on building new layers (e.g.
Re:I read this and found it to be terribly funny (Score:3, Insightful)
With me, it's not so much what I wear or where I am but rather that I sort of mentally designate the place as somewhere to do work instead of play games or read slashdot.
Re:Mediocrity (Score:3, Insightful)
I'm not sure I see that as accepting mediocrity. I see it as more a "relaxed" approach.
Microsoft has programmers to pay, shareholders to satisfy, and all sorts of other expenses that come with being a business and, despite being a near monopoly, SHOULD be concerned with the established mediocrity and its effect on the end product. You can't do the job right if you don't have the right tools, and it sounds like Microsoft doesn't have very good tools to start.
Linux, OTOH, isn't under external pressure from a marketing department that keeps promoting superwhamodyne products by such-and-such a date. Linux is ready when it is ready, and not before.
My non-programming two cents...
Re:Interesting thing is... (Score:1, Insightful)
So I can't really understand what people mean when they say "Windows is easier to use".
They mean that they can go to a website of some little tool maker (let's say a shareware tool that helps them oraganize a shopping list), they can click once to download the tool and another click to install it and they're done. It isn't like that with Linux (no matter what distro) and it never will be until it achieves a critical mass on the desktop which won't be anytime soon.
And no, fiddling with a zillion different "package managers" is not a viable alternative.
Re:Interesting thing is... (Score:1, Insightful)
Great post!!!! Now if only someone can create a real Workplace clone for Linux (or Windows for that matter). After all, OS/2 WPS is more than 10yrs old and there not a single desktop that can equal it. WHAT ARE WE DOING WRONG???? And I don't mean the cosmetic improvements, because todays desktops are years ahead in look, but they all still years behind technically the WPS. Yeah, I know there is WPS-look alikes for Linux, but they not the same. We still do not have a true object oriented desktop like OS/2 was.
Re:I read this and found it to be terribly funny (Score:3, Insightful)
Windows has done drag-n-drop to writeable cds for years.
The $100 add-on might give you a few more options on formats etc., but to just stick some files on a cd, all you need is windows.
Re:Interesting thing is... (Score:3, Insightful)
> little tool maker (let's say a shareware tool that
> helps them oraganize a shopping list), they can
> click once to download the tool and another click
> to install it and they're done. It isn't like that
> with Linux (no matter what distro) and it never
> will be until it achieves a critical mass on the
> desktop which won't be anytime soon.
Except that this has been part of the problem with Windows. It's just simply too easy to install this sort of stuff, and suddenly find yourself plagued with spyware. I'm all for easy-of-install, but it shouldn't be too easy. It's the same thing with attachments. It's one thing to have relatively easy access to attachments, but allowing executables to be clicked on is just asking for disaster.
I think a reasonable compromise is simply allow the software to install under the user's account, and thus have the vulnerability at least not be inherited by anyone else logging on to the computer. I think the reason we haven't seen this is because the designers of *nix GUIs just don't want to introduce the worst parts of Windows.
> And no, fiddling with a zillion different
> "package managers" is not a viable alternative.
And I've admitted this is a problem, though like I said, some of the distros have pretty damn easy installation procedures, and I think if it's thought out carefully, even installing from the browser might be okay, but if what you're asking for is simply the same brain-dead methodology that has made Windows extremely vulnerable, then absolutely not.
At some point, whether its Microsoft or KDE or who or whatever, users are going to have to face the facts that they are a critical quotient in the security issue, and that means that users are either going to have to be educated or not be given the ability to install just any-ol-thing. Ease of installation isn't necessarily a good thing.
Re:ps. (Score:3, Insightful)
Linux aggregated around the kernel, it wasn't integrated into an existing system. The FSF tools were not the only options, nor in some cases were they even the best... though in the process of building Linux led to big improvements in many.
But... people had been writing the userland components for over a decade at that time, starting in the late '70s with the Software Tools Virtual OS. There were two or three alternate C compilers, a huge variety of other components, and long before Linux was solid the BSD userland became available.
Linus has stated that if BSD had been ready a year earlier, he would just have used that... Linux wouldn't have existed at all. And the only major component in the open-source BSDs that came from the FSF is the C compiler... and that's not the compiler it started with.
So, if Linus hadn't existed, we'd be debating BSD and Hurd now, not Linux and Hurd. If RMS hadn't existed we'd be using some descendant of lcc or tcc... on Linux.
Re:Another one from the "Duh!" file (Score:3, Insightful)
In contrast, the Windows NT kernel is of a high standard, although some limitations prevent it from scaling beyond 16 (I think. Maybe 32. I'm too lazy to check.) CPUs - not really a limitation anywhere other than high-throughput server space. Read Tanenbaum's Modern Operating Systems for more info. Almost nothing that's wrong with Windows is the result of the kernel (well, maybe the decision to move more drivers into ring 0 in recent versions) and so moving to a Linux kernel would be silly. The main flaws with Windows come from some design decisions at a higher level (win32 api) made before the Internet was in widespread use, and before the need to maintain backwards compatibility with some these design decisions (and even most 16-bit DOS/Windows code). Microsoft really need to do what Apple did with OS X, and run legacy code in an emulation layer, although in their case with a nice big sandbox around it. Coincidentally, I believe this is exactly what they are planning for Longhorn (which seems to be going the way of Copland at the moment...).