Desktop Linux on x86 - Adapt or Die
An anonymous reader writes "The recent announcement of Apple's upcoming x86 systems has gotten a lot of people thinking. Among the conjecture, there has been much thought given to how Linux will be affected by this move. The author of this article does not believe that Linux as a whole is threatened harmed by the 'Mactel' alliance, but does point out that his could mean major trouble for distros like Xandros and Linspire which are reliant on the desktop audience. These distros are clearly not ready to take on OS X, which will soon be the primary x86 alternative to Windows XP not only because of OS X's dedicated and outspoken user base but because of its slick looks and ease of use."
See, you can buy cheap hardware and run linux. OSX wont replace linux for those who are conscious about what money they have and what the hardware will cost.
MS should be worried shitless that, one day, Apple will release OSX for all x86 desktops and put a big dent in MS's marketshare. Unless Apple signed some no-OS-compete agreement forever with MS, they have a lot more to worry about in the long run (think 10+ years).
The Doormat
If you're not outraged, then you're not paying attention.
I write this as a former Windows user, occasional Linux desktop user and new Mac user:
The reason I switched from Windows is that the features I wanted (better shell, nicer GUI, easier to use programs, better workspace, more scriptability and easier to organize folders) was already on the Mac.
Sure, Linux has some of these features. The problem, I've found is also an 'apparent' strength of other 'Nix systems: X, KDE, Gnome and a whole slew of Window Managers and DEs. I say apparent, because, frankly, with all the work that has gone into each DE and WM, Linux could have one (maybe) two really kick-ass desktop environments. Insead everything would work well together. And something has to be done with the library compatibility problems.
I only want some OSS programs. I don't really care about having an OSS (GLD' whatever) Operating System. I'll pay for the OS. Heck, I just bought a Mac and am really happy. I just like to have 'options'! Doesn't everyone?
As a long-time Macintosh user, Apple's move to Intel chips has actually sparked my interest in Linux.
It's not yet entirely clear why Apple chose Intel. There is some reason to suspect Intel hardware will ease implementation of system-wide DRM capabilities. Time will tell.
The microcomputer revolution of the 1970s and '80s was about individuals controlling machines that had once been the exclusive domain of governments and big corporations. Now DRM, product activation, live updates and other technologies are being used to take back that control. Well, I'm not going back.
I don't doubt that the Linux desktop might seem crude in comparison to Mac OS X. But if Apple chose Intel to help put DRM everywhere, then I, for one, will be more than willing to go "rough it" with the free souls of the Linux world.
When all you have is an axe, everything looks like a grindstone.
As long as Mac OS X is locked to Apple hardware, it is not a true alternative to Windows and Linux (or *BSD for that matter) that happen to run on commodity hardware. Unless Apple will sell their X86 hardware at Dell prices, there will not be competition. Also, the crowd using free (as in speech) operating systems on their computers are not likely to use closed operating system, let alone closed hardware... Just a thought.
OS X will not be greater a threat to Linspire or Xandros or the like than it already is, because OS X will never be able to (officially) run on the same machines that Linspire and Xandros run on: Generic x86 machines. And what does that mean for the market?
Well, if you wanna run OS X and you don't already have a Mac, you'll have to buy a Mac. Just like today. If people wanted to run OS X instead of their favorite Linux desktop distro, they would have already bought a Mac. Since they haven't, that's your proof right there that nothing major is going to change.
I'm sure a few (including me) will try to reach triple-boot nirvana after the Intel-based Macs come out, but not most users.
This doesn't fundamentally change anything for the average Linux desktop user.
Steve
These people need to realise that one day they will be forced to pay seperate licences. You cann't get away with it forever. And Microsoft has already started pressuring users and businesses to buy licences with the "genuine (dis?)advantage" program.
With DRM and palladium coming, I think Microsoft will be able to enforce license purchases within the next 2 years. Notice I said 'be able'-they won't do it yet, not as long as there are viable free alternatives like Linux.
So their strategy is to get as many users as possible on windows, pirated or not, and when there are only a few left on Linux, force everyone to pay. Then you'll look ofcourse for a free alternative, but it'll be too little too late.
The funny thing is, most windows users with pirated versions, think they are 'cheating' Microsoft, while infact they are playing Microsoft's game. And Microsoft treats them like criminals, like they've done something bad. The same tactic banks use. They'll give you more credit than you can afford, and when you cann't pay it back on time, they'll blame you and treat you like you've done something bad. So people usually fall into the trap, borrow more than they can afford and end up paying extraordinary fees without complaining. After all, it's their fault...isn't it?
These people need to stop thinking about short term convenience and think the long term implications of their actions.
Nowdays, Linux is very easy to use and very powerful. There really is, no excuse not to use it.
VStrider.
Actually, try to find me a 12" PC laptop, with a competitive price compared to the iBook. And a separate graphics chip with own memory is a demand.
I have a really elegant proof for Fermat's last theorem. If this sig was only a bit longer...
"Look at OS X. Take the Dock for example."
So right click on the KDE panel and install a couple docks. No big deal. I put one up on top that I keep visable with a digital clock, and a dictionary lookup entry widget, and links to documents that are currently in play alot. My math and physics software all goes on a hidden dock to my left. Office software on a hidden dock to the right. These are all drag n' drop. No big deal to configure. And yes, the icons zoom when I mouse over them. I can set transparency, etc...
"Desktop users don't want power, they want simplicity."
When you are confused, you want it simpler. Once you know your way, you want it powerful. The trick is to design things to be simple enough to start with (ease of use, entry), without forcing them to never grow into efficency (ease of power, later on).
I got a mac recently. I like it quite a bit but it does have some flaws. In fact, there is one flaw that is amazingly annoying -- no middle-click paste. The only saving grace is that quanta runs under X11 and I can middle-click there. But it drives me crazy when I'm using aqua apps and simply highlight, switch apps, try to paste, then have to redo the process with ctrl-c type strokes. It's really a downer.
What changed under Obama? Nothing Good
????
I've presented this story before. My business partner is not computer savy at all -- she still has an AOL dial-up account. She calls the computer a "cpu" -- you know the type, points at the case and says "cpu". She's the type of person who bought a computer years ago and since it still allows her to check email, she has no interest in getting something else.
When I set up our office, I used linux set up with the KDE desktop. I was surprised one day when I came in and saw she had independently changed the wallpaper to her kid's picture. She's never had a problem with openoffice (used MS Office before) or Mozilla or Kmail.
I'm thinking that if she can use it, almost anyone can. She may not be able to do a deep configuration, but she wouldn't be able to do that with her windows box -- I know this -- recently she called me up and asked me how she could burn a CD from her computer at home (she has no problem with k3b). Fact is, it's been so long since I've used any windows software it was hard to guide her (you know, I'm saying stuff like "ok, look at the menu for something suggests the concept of burning or writing
What changed under Obama? Nothing Good
Interestingly, people who don't want to use OS X or Windows could probably come up with something based on Darwin, using X for the GUI.
Device drivers developed for OS X on Intel will probably work with Darwin on Intel, which could help such an X+Darwin system gain better device support than Linux.
So this could be the real 'threat' to Linux on Intel - the ability to take advantage of OS X's drivers while not having to pay for OS X or a Macintosh.
The people who buy dirt cheap hardware will continue to do so, and that hardware will continue to run Windows or Linux. Anyone who runs important desktop applications or server stuff on a dirt cheap machine , or without enough technical expertise to make Linux/Mac/Windows 'easier' to use is an idiot.
So, this is the situation. Windows and the Mac has good usability and good interoperability. If the Mac is on Intel, then the interoperability will be better. Mac has some benifits in that it uses existing standards, so one often does not need device drives to use the basic functionility of hardware that meets those standards. Windows is slightly worse in that department, one needs to find the proper driver to match the hardware/os combination, but has the monopoly advantage.
linux is coming in with neither of these advantages. It does not enforce hardware standards, and so people feel they have a worse user experience becuase the hardware is messed up. It does not run the industry standard software. It is trying to crawl under the door on the basis of price, but it can't becuase MS forces vendors to pay them a tribute no matter what OS is shipped.
Far from damaging the Linux people, Apple is showing them the way. Not the way to 30% market share, but perhaps to 5%. Build a quality system. Be proud of it. Don't deal with the customers that want the cheapest thing. There are a good many people who would buy a linux system if they could get the work done. As long as Linux tries to copy MS and only focuses on the Geek or Cooporate deskttop, nothing will happen. People want a machine that works.
"She's a scientist and a lesbian. She's not going to let it slide." Orphan Black
The game is Unix vs. Windows.
... I think I see it starting to break now... Bill's got all ten fingers plugging leaks ... whoops Bill, there's one more leak right at about waste level...
Micosoft is breaking its noodle to crack the Unix server market. Now here come two Unix-based OSes entering the its desktop market. The more Unix that enters the desktop market, the less chance Microsoft has of taking over the Server market. Why? Because Microsoft is trying to define the server market by creating "features" and proprietary formats in the desktop market. The best example is Outlook-Exchange which is dominating the Intranet mail. That could change. What if Novell open sources Groupwise (yes, you heard it here first!)? What would the repercussions be? What if OSX becomes a clone OS running on any hardware?
I know this post has been somewhat of a rambling wreck, but really you see my point. Unix on the desktop is what MS fears. So I say OSX is good for Linux.
Now, why isn't it happening NOW? Because MS is scaring the shit out of hardware vendors with reprisal should they abet this effort. But eventually the dike will break and
That's very true.
And I love my Mac.
But truth be told, the Mac can be annoyingly toyish too. Not nearly as often as Windows, but it's there for sure.
The interface guidelines have slid downhill since OS9, the glowing gumdrops widgets annoy the hell out of anyone trying to get real work done, and of course Macs have always, and continue, to insist on the particular way they want you to do it, which is not always optimal for anyone but a rank beginner. I know, for instance, I agree with the poster that misses middle-button paste, and what he didn't mention - focus follows mouse with no autoraise. X11, whether running on top of linux or bsd or whatever, will give me that. Mac absolutely will not.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
There seems to be a lot of arguing back and forth about the user interface, but quite frankly I think that the user interfaces of gnome and kde and the distros that use them are just fine. I think, as you do, that their interfaces are just as serviceable as Windows or Mac.
There are some things that need to change, though. One thing, which I read about from time to time is the lack of a consistent clipboard in Linux. That's a big one to me.
Another factor, though, which I don't see much mention of, is the lack of good error messages and other systme-level feedback in Linux.
My typical example of this is getting printers to work. I'm no Linux expert, but I do know more about it than Mom and Pop at least. But getting a printer to work that's attached to a Windows machine is always a headache. I pray that it works the first time, and sometimes it does, but more likely than not it doesn't. So where's the problem? Is it CUPS? Is it Samba? Or somewhere else? There are no error messages whatsoever to give me clue.. Instead, the system either reports nothing, or reports that printing was successfully spooled. However, nothing emerges from the printer, so it's always some manner of guess and check. Once it finally works, it works forever, but I'm afraid to even touch the printer configuration for fear of screwing it up again.
Another example is a machine I dual boot into Linux or Windows. In Linux, all of a sudden the network connection was incredibly sporadic. I finally gave up using the computer and went to another, as I didn't have time to tweak this and discover that to figure out the problem.
Later I booted the same machine into Windows, and as Windows started up it informed me that there was an IP address conflict and that it would therefore deactivate the network card. Well, once I knew that, it was easy to fix the problem for both windows and linux.
Anyway, that's what I mean by lack of error messages and system feedback in Linux. And that's the level where I think Linux needs the most improvement.
A number of people in separate threads have made the point that, due to apple's statements that osxi will only run on apple's machines (without some hacking...) apple's x86 switch will mean little outside of apple.
Let me ask you this. What reason do you have to believe apple? Let me remind you, that not very long ago Jobs went on record saying that he had no plans to switch to x86. It is now obvious that this plan has been in the works for a long time, and jobs has only been waiting for the right time. At the time he made those statements because he knew that if the switch happened, it would be a long way off and he didn't want customers to hold off on buying hardware during that time.
In this industry it is sometimes very important to not show your hand too early. Why would jobs be bluffing about not allowing machines on generic hardware? The obvious answer is because such a switch would not happen immediately. It would be a huge transition for apple to suddenly switch to a hardware vendor with enormous revenue, to a software vendor with a much smaller revenue stream, but probably much higher profits. This switch would probably scare the hell out of apple's investors, and it would not be a good idea to do while everybody is still worried about the x86 transition.
There is another reason why apple would not show their hand at this point. Microsoft went on stage at the WWDC, and they commited to porting office to x86 osx. Would they have done this if apple was making moves to put itself into more direct competition with microsoft? I think they would be foolish to. Even if they felt compelled to do so for fear of more anti-trust aligations, microsoft could certainly take all the good developers off the office mac team, resulting in a late and buggy versions of office for mac.
Anyway, I'm not saying there is a good reason to believe that apple will make osx avaiable for generic x86. I'm just saying that if they *were* going to do such a thing, they certainly wouldn't tell *you* about it. So apple's statements about lack of support for generic x86 should be taken with a grain of salt.
Oh get a life.
I'm sick to death of people out there who reckon that Mac OSX is actually *easy* to use.
Any of you guys actually GOT a Mac?
I have 2 laptops, one a Pentium M 1.8GHz and the other a 17" Powerbook, 1.5GHz. I am *constantly* pissed off at my Mac purchase.
Both have 1GB RAM, the intel version cost $2600, my Mac $4700. Totally NOT value for money.
Software is hideously expensive for Macs, and doesn't work the way you want it. I bought the office suite as (of course) you need to have it, and Entourage isn't even compatible with Outlook, how *dumb* is that.
The adobe photoshop CS suite (or more likely, that fricking 'preview' program) managed to randomly kill a whole memory cards worth of images from my camera, say 300 pictures.
The DVD player program isn't region free. This pisses me off as Australia & USA don't share regions. If you need to play DVDs, something like VLC is needed but that's freeware ANYWAY and available for Windows too.
Windows in Aqua can only be resized down the bottom right.
Rendering takes bloody ages, at least on my 1.5GHz model.
I use FreeBSD *all* the time, and have it running permanently on my other laptop through VMWare, and the BSD subsystem that's under Mac OSX (which is *why* I bought the mac laptop in the first place) is *nothing* like it. I mean to get stuff done, how do you guys actually see this happening?
You can't run serious UNIX, you can't script the interface using button hooks, it's all shit.
The only thing that the Mac gets used for now is watching divx movies, as it *does* have a nice screen, but nothing else.
A modern desktop Linux distribution will be perfectly usable for anyone who's willing to learn *any* new desktop OS.
No, believe me, it's not. For many reasons. I will list a few of them here:
- lack of standardization
There is little standardization of menu structure, general look and feel, and most importantly, how to configure the system. How do you configure a network connection on linux. Explain in one paragraph. For the mac I can do this, for linux, maybe for specific distro's, but that knowledge is mostly useless because you constantly have to relearn how to do what you already know how to do as new releases are made.
- lack of desktop infrastructure
For example, there is no object framework for interaction between applications. Yes, there's gnome's corba, and kde's dcop, but they are incompatible, so in practice you have nothing. Just try copy/pasting an image around between apps. Often it just doesn't work.
- lack of simplicity
This is what bothers me most. Linux apps and configuration tools tend to be horribly overdesigned. Just look at your run of the mill KDE desktop. It has intolerable GUI bloat. Reams of buttons and menu's everywhere. Just try figuring out how to configure your konqueror when you're a novice user. GNOME does moderately better, but then GNOME lacks integration even more than KDE does. And also, a lot of stuff you do not find in predictable places. Why is the KDE configuration independent from the hardware configuration? There should be one big configuration panel that integrates everything, like the control panel on windows and the preferences pane on OS X.
There are many, many reasons why the linux desktop is unsuitable for regular people. I'm not going to list them all. The windows desktop admittedly is not that much simpler, but the big difference is that everyone knows at least one person who knows how to do system administration in windows. With linux this is not the case. I used to run debian on all my desktops, and when I got questions about redhat I was lost. There is little generic knowledge, other than what happens on the command line (and even there the pointless differences are plentiful).
Anyway, I run OS X on my desktops now. It's true what they say about it "just working". You end up paying more money up front, but you win it back in time. Then again, if your time is free...
HP produced CDE, under than name HP-VUE, which became the standard Unix desktop.
.xsession or .xinitrc.
I know about CDE. CDE is the biggest disaster to hit the UNIX world since SCO Open Server. It's an appallingly bad design, and completely failed to provide a usable desktop environment.
You're defining desktop as "something you put on a desk". I'm talking about a computer that you run desktop software on, that provides a Xerox-star-like working environment. So...
Now lets discuss our definition of success.
My definition of success includes actually fulfilling the requirements that providing a desktop environment created. CDE was a complete failure there: not only was the file browser completely inadequate, but the window manager and that horrid tacky dock thing was a worse tool than the vanilla menu-oriented Motif window manager it replaced. The first thing I did on our CDE-infected servers was to come up with a hook that let our users keep on using their preferred window manager, because with the introduction of CDE they lost the option of just starting their own WM from
Don't talk to me about CDE. I know CDE. CDE makes Windows 2.0 look good.
The majority of the UNIX-based "desktops" are similarly screwed up. NeXT was one of the few exceptions.
SUN were the largest player in the high-end market through out the 1990s, despite having distinctly inferior processor technology towards the end of the period.
Sparc was always a trailing edge processor, right from the start. The "register window" design made context switches appallingly expensive, but didn't provide the compiler with enough scratch registers so it suffered from almost as much register pressure as the 386. To combat the context switch problem they put multiple register contexts on chip, which helped a lot under low loads but you could always see the knee in the performance curve as soon as you got more ready processes than yuo had register contexts.