Re:BeOS _is_ a member of the UNIX family.
on
Zeta Goes Gold
·
· Score: 1
Every argument you make for BeOS as a functional Unix applies equally to Windows NT with POSIX.
There's a big difference, and you touch on it yourself:
It could understand Bash scripts, and if you want to, you could use a remote BeOS system with a Bash interface very comfortably if you are used to Bash interfaces.
Since the POSIX subsystem is not how the Windows NT box boots and manages itself, you can't log into a Windows NT system remotely and use POSIX subsystem commands and applications to manage your Windows NT box. Even if you replace the POSIX subsystem with Interix you're continually running into the differences between the native Win32 subsystem and the POSIX-derived Interix subsystem every time you go poking around outside the Interix box.
Basically... Interix may be a member of the UNIX family, but it runs under an OS that's only indirectly related. The fact that there's no line you can draw like that in BeOS means there's no hard limit to what you can do from the shell.
If GCC and top are part of your definition of a member of the UNIX family, then there was no UNIX before the late '80s. If "a decent bash environment" or even "a decent bourne shell environment" are part if your definition, then the first UNIX systems I used definitely weren't UNIX. If X11 is part of your idea of what a "UNIX" GUI is, then that leaves out ALL the early graphics workstations running UNIX.
Even now, as nice as GNOME and KDE have become, [...]
I would never use the phrase "nice" to decribe either of these steaming piles.
Come on, in a world where George Bush can be elected President not once but twice, do you really expect anything sane to come from the courts?
In a world where people are willing to use an operating system that's been deliberately designed for insecurity, do you really expect people to care if 9/11 became anything but a pretext for the New World Order?
Noticed a story in the WSJ about how product sales were projected to drop dramatically
No kidding. It's not rational, but it's certainly understandable. People aren't rational, if they were Bill Gates would be making a living asking "Do you want fries with that?", assuming he'd stayed out of jail.
I just said I wasn't going to buy a Mac - which I had planned to do this month - until later maybe in January/Feb when the new stuff should finalize and I might know what to get that would keep working.
The new stuff will be out next June.
This is the first time that Apple has pre-announced a new product line in years, they're normally really tight on details.
1. All current models and all models coming out between now and next June should be supported with OS upgrades for the next 4-5 years. This is based on Apple's history, they haven't actually said this is the case, but if the last PPC macs go out in 4Q07 according to plan, that's only 3 years from then... and they didn't pull the plug within 3 years on the last two transitions. So this is actually a pretty conservative prediction.
2. There won't be any major new models or major upgrades in the next 12 months until the Intel-based models are available. Well, they *might* do an e600-based Powerbook, but I'd be real surprised. They're not likely to have a G5 eMac or anything like that.
3. If you're looking for gaming, you want a model with at least a 64M video card for Quartz Extreme 2d. The low end desktop that has that is the new eMac. The iBook and Mac mini have 32M Radeon 9200, and may or may not have a refresh before next June.
4. The first new models will be the Intel-based Mac mini and at least one Intel-based laptop. The current G5 models will be the last to be refreshed. So that means an Intel mini next June, followed by the laptops and maybe an Intel eMac by the end of the year.
All this is based only on what Apple has already announced and what's on their website.
So...
If you buy an eMac now, you'll get a QE2d-capable Powermac. It's unlikely it will be refreshed between now and 2H06, when they come out with an Intel version. It's unlikely you'll get a better low-end from Apple between now and next June unless they refresh the Mac mini, and holding off on a refresh of the mini actually makes less sense now... it was more likely before 6/6.
If you buy an iMac now, you'll get a G5, and it's unlikely you'll get a better model before 1Q07... and, again, the news on 6/6 makes an upcoming refresh on the iMac less likely.
If you wait for the Intel-based models, you'll be waiting until next June. If you want an Intel-based iMac, you'll likely be waiting until 2007... they've already said they'll update the low-end models and the laptops first.
So I don't see much point in doing more than waiting a few weeks to see if the current shortage of the mini (Estimated Ship just went up to 6-8 business days) indicates a refresh with a Radeon 9600 to support QE2d under Tiger... if you were looking at the mini, OR waiting until 2H06 or 1H07 and getting an IntelliMac.
Steve Jobs could have a hell of a surprising "Just one more thing" at MWSF'06 and release an Intel box before next June, but June's what they've committed to.
Because just because he's 14 doesn't mean I buy what he wants in games. If he wants to play Doom V he'll have to play it at a friend's house - it's called parenting.
So that brings us back to "How long do you keep using a computer anyway?"
He's 14 now. When will he be old enough to play Doom V? 16? 18? You're looking at replacing it in 2-4 years anyway unless it's running Game-OS. Whatever you buy now will still be well supported that long, and in the meantime you'll be saving the guy from premature exposure to Windows.:)
IBM sees it, and puts a TrackPoint in the top of their mice for two-axis scrolling.
The IBM "scrollpoint" mouse I tried looked just like a trackpoint on a mouse, but it was only an up-down controller. I'd love a full trackpoint on a mouse, do you have an actual model number for the 2-d version?
I won a Kensington mouse with a scroll panel a year or so back, and gave it away. The lack of feedback is really annoying.
There's a couple of better alternatives that I've never seen. One is the IBM trackpoint... the button between G and H on thinkpads. It's a small, reliable, and well understood 2d controller that would fit on a mouse really easily. IBM even put a trackpoint on a mouse... they called it a scrollpoint. Unfortunately they only made it into an up/down scroller. *doh*
The other alternative that's good is a software solution. Logitech came up with a MUCH better idea a few years back. They implemented it in the mouse software instead of the mouse, and that poorly, but the basic idea is good.
You could configure the middle button as a "grab" button. When you held it down, moving the mouse would scroll the document in all four directions. Unfortunately, it also put an obscure icon down where you clicked and didn't clean it up properly, but there's no reason that part couldn't be left out.
We're still on NT4 for our domain controllers, because going to Active Directory would force us to completely redesign our DNS environment, either to support AD with a UNIX-based DNS, or migrate all our DNS-related scripts over to Windows... and last I checked some of the things we're doing, like synchronising subdomains across split DNS, aren't actually possible in AD.
The one new feature in Windows XP that I've ever had occasion to miss on Windows 2000 is the integration of the Citrix multi-user and remote desktop support, specifically virtual consoles (AKA fast user switching). And I'm actually kind of ticked off that Microsoft restricted it to one console per user-id, because when Apple did their version they put the same stupid restriction in, instead of following the lead of BSD and Linux and just letting you have as many concurrent sessions as you have VM for.
In general, though, I agree with you. There's no killer new technology in Windows XP, just more integration of third party packages, and the user interface "improvements" aren't any such thing. Ironically, the main effect of the "improved security" is that if you're using Microsoft IRM (Intellectual Rights Management, their DRM solution) you can't use Windowblinds or any similar software to fix the "improved" user interface.
The question started out as: will the move to Intel threaten Linux. [...] Moving to Intel will not threaten anyone [...]
Great point. It's a shame that you're making it to someone who doesn't disagree with it. Why don't you try reading what I've actually written? I didn't write "the author's conclusions are wrong". My point was "the author's argument's are irrelevant".
Even if Apple did release OS X for generic intel boxes, it wouldn't make any difference to Linux, because Linux and Mac OS X are good for different things. Because on the desktop there are only two operating systems that have any significant commercial software base. Linux isn't one of them and it has no prospect of becoming one of them any time soon, and there's no magic vaguely-hinted-at software that will show up in the next two years that will change that. Because as far as any software is concerned Linux and Mac OS X are barely distinguishable.
So... the question "is Mac OS X on intel a threat to Linux" is a category error. It's like asking "is vanilla ice cream a threat to caterpillar tractors".
I'm a UNIX guy. I'm part of the core constituency for an open source desktop operating system. I'd love to see one show up. I was using X11 and UNIX on Wintel hardware as my main desktop environment until about three years ago, and if Linux on the desktop was an option I'd be using it, because I've been using UNIX as my desktop environment of choice for longer than Linux has existed.
So, I don't use Mac OS X because it's going to knock Windows on its ass, I use Mac OS X because it's the only way to get UNIX with a desktop environment that doesn't suck. On the server side, I've got FreeBSD, Linux, Tru64, HPUX, and Solaris boxes. But none of them are an option on the desktop.
So what does "is Mac OS X a threat to Linux" mean? Mac OS X and Linux aren't in competition with each other. "One understands this is not very nice for Linux fans, and they would much rather focus on whether Mac OS X is a threat to Windows. That isn't the issue."
I didn't raise that question. The Fine Article didn't raise that question. You seem to have decided that's what you want to talk about. Well, great, have fun. But please find someone who's using Windows XP and talk about it with them, OK?
It's pretty clear that the switch to Intel means in the next year or two there are going to be new Macs that are significantly faster than today's Macs (especially the aging G4 based notebook line), and quite likely they are going to cost less than today's Macs.
Notebooks, yeh, I suppose that's likely (though closer to two years than one... there's not going to be any x86 based Macs for a year, and it's going to start at the low end of the product line), but I don't think there's going to be desktop macs faster than the iMac G5 until well into 2007, and the Powermac G5 is going to be the last update near the end of the year, and I doubt they'll be all that much cheaper than they would be with the G4 or G5: the CPU isn't what makes Macs cost that much, most of the "Mac Tax" goes to their high margins to subsidise the OS.
And the first few x86 macs... the ones you're waiting on tenterhooks for... are the ones I'd personally stay away from. They're likely to be the x86 equivalent of the Nubus Powermacs, the Rev 1 Beige G3, the "Yikes" G4.
And while they seem to have picked a pretty conservative release schedule, it could still slip.
So, unless you need the new computer right now, you'd be crazy not to wait.
Only if you completely discount the time value of money. After all, there's always going to be faster and cheaper computers coming in the future so it's always rational to wait. It's all a matter of... how long can you wait? And base your decision on the assumption that you'll have to wait at least a refresh cycle longer than that. You might luck out, and there won't be any early release bugs, but do you want to plan on it?
So... if you were planning on buying a Mac now, it may still be rational to buy one now. Especially if you've been delaying a purchase, double if you don't have a Mac at all yet, triple if you're using Windows. This is even a good time to get an eMac, since they've just had a refresh and an updated GPU for Tiger. And don't forget, Tiger is an OS sweet-spot: the next OS release, Leopard, is likely to be at least as much about changing its spots as adding new features.
You seem to imply a mutual exclusivity between a good server OS and a good desktop OS.
Not at all, it's just that there's new ideas and new ways of doing things developed all the time. When NeXT started working on NeXTSTeP in the early '90s the world was very different... there was barely an internet and no google, no apache, no Netscape or Internet Explorer, a 68030 was a pretty fast chip, and expecting a desktop OS to run multithreaded webservers when there was no web and no webservers is a but much... and since neither NeXT nor Apple has concentrated on the operating system's performance running the kinds of server apps ArsTechnica was testing...
Well, for those kinds of apps, it's like comparing a 15 year old OS to one from five-minutes-ago. That's all.
I'm running Panther in 256M right now, but on a 1.33 GHz G4. No tweaking. It's sometimes pretty doggy: I've found that the real "knee" in the performance curve is about 360M. Once you get over that, things get a lot better.
Well, I don't agree that Linux isn't in the running for the home, but was making a different point. That is, XP and X are not in the same league.
OK, I get it. What's that got to do with the original message... which is that if XP and X aren't in the same league, Linux isn't in the same game as either of them? AND its market share problems are at least as bad as OS X.
That is, let's say you're absolutely right and the future of OS X is black as Bill Gates' heart.
It's still in a better position than Linux, because it's still a better option than Linux for someone who's using Windows, or looking for an alternative to Windows. Why? Because it runs almost everything Linux does, and it runs some version of almost everything Windows does, and it's easier to run any of that than on either Linux or Windows.
Installing a program on Windows XP, best case:
Download installer. Open the installer. Click "yes" or "no" a few times.
Installing a program on Windows XP, typical case.
Download installer. Open the installer. Click "yes" or "no" a few times. Reboot.
Installing a program on Linux, best case:
If it's on the CD that Linux came on, you select it from the CD, run it, then open up a terminal program and use "rpm -q" to figure out what the command to run it is.
Installing a program on Linux, typical case #1:
Download the RPM, run "rpm -i", download the packages it says it needs, "rpm -i" again, repeat a couple of times, then use "rpm -q" to figure out what the command to run it is.
Installing a program on Linux, typical case #2:
Download the tarball, unpack it, cd "program-1.4.5",./configure, read the error message from configure or INSTALL file, download and install the stuff it depends on,./configure, make, make install.
Installing a program on OS X, best case:
Download the file, double-click it, drag the icon to your applications directory.
Installing a program on OS X, typical case:
Download the file, double-click it, double-click the installer, answer a couple of questions.
Personally, I prefer unpacking a tarball and running configure to running some unpredictable GUI installer. But even for me it's hard to beat "drag icon to application directory".
There is by the way a big difference between a few percent and rising, and a few percent and falling.
Then you should be enthusiastic about OS X, because it's 2.3% and rising, and while it's going to lose some to people who freak out over the whole Intel thing that's a one-time hit unless Apple does something weird like switching back to PPC in 2009.
according to the grandparent sited article, linux is much faster than OSX (five times in some cases) in basic thread handling functions, which is very important for server performance.
That's why I started out by saying "OS X isn't designed as a server OS" (except by comparison to OS 9, which was horrid) and finished up by pointing out that a Mac OS X desktop is a great accompaniment to whatever UNIX variant you're using on your server.
The Mac's for my son, and we're talking game software [...]
Um, so why was a Mac even on the shortlist in the first place? I'm a UNIX geek, I hate Windows, but I'm not crazy enough to try and convince a gamer to run anything else.
But I'm thinking my trusty old iBook [with 32M of RAM] would probably explode the first time I hit F12 to bring up dashboard.
If you've got 32MB of RAM, and you're running OS X, your geek-fu is better than mine. I've never installed OS X in less than 96M and that took over 24 hours to complete the install.
Well, no, it isn't really. OS X is a "BSD Single Server" design, which avoids the overhead of a microkernel by using a conventional monolithic UNIX kernel to provide most server tasks. It's really quite efficient, and runs UNIX software very well. Where it falls down is in massively multithreaded applications because, well, Mach isn't really a microkernel either and it doesn't have the low-overhead thread creation and IPC that a microkernel really needs.
Which is why the GNU approach of using the Mach kernel as a microkernel didn't work, and the Single Server model used by Tru64, NeXTstep, and Lites works quite well. Particularly for the desktop, so, to repeat the point you apparently missed... regardless of what UNIX you use on your servers, a Mac OS X desktop is a great accompaniment to it.
The real debate is, who cares, when you have less than two percent market share?
2.3 last I checked, and rising. It'll probably go down a bit as people freak out about Intel, but unless Apple really screws up badly between now and the end of 2006 it'll be heading up again by then.
Anyway... somehow I feel compelled to ask you why you think that question applies any less to Linux than it does to Apple. Right now Linux on the desktop has a miniscule market share, and almost all of that is due to companies that have switched en-mass: Linux/UNIX strength is as a server OS, and the highly controllable environment makes it a natural for IT departments.
So why don't you think Linux tiny desktop market share makes it even LESS relevant than Apple?
I mean, I hate Windows, I'm a major UNIX booster and I've admired Apple from a distance, but until OS X came out, I was honestly unable to recommend anything but Windows to a home user who wasn't technically oriented. Mac OS 9 was horrid, and none of the free UNIXes is an option for Joe Average unless they can buy it preinstalled or get a geek friend to set it up for them. Heck, I didn't even really feel good about recommending OS X until Jaguar.
Even for business, where Linux is more practical, Windows on the desktop and UNIX in the backroom is still a rational approach. Until this year Linux had an edge on Apple because Apple simply didn't make a desktop machine with a price that made sense for a business.
But for the home? There's only two options: Mac OS X and Windows XP. Linux isn't even in the running.
No sense rushing if I'll have to buy new software anyway.
Let's see...
Apple has typically supported old hardware and operating systems for five years after a transition (68000-68020, 68k-PPC, Nubus-PCI, OS9-OSX, and hardware abandoned by OS X has generally been that far behind the curve). There's really no good reason for them to drop PPC software support this time, because they're using an OS that makes old platform support easy. There's less reason than ever for software vendors to drop the old hardware, with the single exception of game software, because by the time Leopard comes out they'll have had to become portable... and there's an awful lot of recently purchased Powermac G5s that are still going to represent potential customers 5 years from now given the way Moore's Law has hit the wall in the last two years (remember, IBM didn't hit 3 GHz but Intel's speed improvement over the same period was even less).
And most companies are likely to provide "Intel-compatibility" upgrades fairly cheaply to avoid losing their customer base.
So if you buy a Mac now, it's going to be 7 years before Apple drops PPC, and at least 5 before your apps are unavailable for PPC... so what's the new software you're worried about having to buy?
In the meantime, I'm uninstalling OS X on my older Mac gear. Tiger left two of my machines behind.
What, without even waiting for XPostFacto 4 to get out of beta?
Tiger's so far shown itself faster on unsupported Macs than Panther, just as Panther was faster than Jaguar. Or, you can just keep using Panther... I've still got Jaguar on one of my boxes because it does everything I need there. It's not going to be any harder to install Linux on your iBook in 2012 or so when Apple finally drops support for PPC, or in 2009 when Whatevercorp drops YourCriticalApp for PPC, so why jump the gun? Pique?
The truth is, I might end up buying one or two of the software packages that I "try out" on my computer if I were prevented from obtaining them in other ways.
Yeh, I know that lots of people don't buy shareware. My own shareware non-income shows that. But if you're one of the guys who extends "trial use" indefinitely, so you don't consider lost sales a problem, why did this outrage you?
Slashdot Mirror
Every argument you make for BeOS as a functional Unix applies equally to Windows NT with POSIX.
There's a big difference, and you touch on it yourself:
It could understand Bash scripts, and if you want to, you could use a remote BeOS system with a Bash interface very comfortably if you are used to Bash interfaces.
Since the POSIX subsystem is not how the Windows NT box boots and manages itself, you can't log into a Windows NT system remotely and use POSIX subsystem commands and applications to manage your Windows NT box. Even if you replace the POSIX subsystem with Interix you're continually running into the differences between the native Win32 subsystem and the POSIX-derived Interix subsystem every time you go poking around outside the Interix box.
Basically... Interix may be a member of the UNIX family, but it runs under an OS that's only indirectly related. The fact that there's no line you can draw like that in BeOS means there's no hard limit to what you can do from the shell.
If GCC and top are part of your definition of a member of the UNIX family, then there was no UNIX before the late '80s. If "a decent bash environment" or even "a decent bourne shell environment" are part if your definition, then the first UNIX systems I used definitely weren't UNIX. If X11 is part of your idea of what a "UNIX" GUI is, then that leaves out ALL the early graphics workstations running UNIX.
Even now, as nice as GNOME and KDE have become, [...]
I would never use the phrase "nice" to decribe either of these steaming piles.
Come on, in a world where George Bush can be elected President not once but twice, do you really expect anything sane to come from the courts?
In a world where people are willing to use an operating system that's been deliberately designed for insecurity, do you really expect people to care if 9/11 became anything but a pretext for the New World Order?
Why, this is Wonderland, nor am I out of it.
Noticed a story in the WSJ about how product sales were projected to drop dramatically
No kidding. It's not rational, but it's certainly understandable. People aren't rational, if they were Bill Gates would be making a living asking "Do you want fries with that?", assuming he'd stayed out of jail.
but MacTel's supposed killer feature, dual-booting Windows
That's a silly idea for a "killer feature".
There's two potential "killer feature"s for MacIntel over MacPPC:
First, a faster version of SoftWindows.
Second, it's likely to make game ports easier.
Neither of them excite me, but they're more likely to benefit the average user than "dual-booting Windows".
"I was going to buy a computer this year, but I heard there will be a better one in two years, so I'm going to wait."
Am I the only one who think that's the silliest decision ever?
Not silliest by a long way, but it's not exactly rational.
I just said I wasn't going to buy a Mac - which I had planned to do this month - until later maybe in January/Feb when the new stuff should finalize and I might know what to get that would keep working.
The new stuff will be out next June.
This is the first time that Apple has pre-announced a new product line in years, they're normally really tight on details.
1. All current models and all models coming out between now and next June should be supported with OS upgrades for the next 4-5 years. This is based on Apple's history, they haven't actually said this is the case, but if the last PPC macs go out in 4Q07 according to plan, that's only 3 years from then... and they didn't pull the plug within 3 years on the last two transitions. So this is actually a pretty conservative prediction.
2. There won't be any major new models or major upgrades in the next 12 months until the Intel-based models are available. Well, they *might* do an e600-based Powerbook, but I'd be real surprised. They're not likely to have a G5 eMac or anything like that.
3. If you're looking for gaming, you want a model with at least a 64M video card for Quartz Extreme 2d. The low end desktop that has that is the new eMac. The iBook and Mac mini have 32M Radeon 9200, and may or may not have a refresh before next June.
4. The first new models will be the Intel-based Mac mini and at least one Intel-based laptop. The current G5 models will be the last to be refreshed. So that means an Intel mini next June, followed by the laptops and maybe an Intel eMac by the end of the year.
All this is based only on what Apple has already announced and what's on their website.
So...
If you buy an eMac now, you'll get a QE2d-capable Powermac. It's unlikely it will be refreshed between now and 2H06, when they come out with an Intel version. It's unlikely you'll get a better low-end from Apple between now and next June unless they refresh the Mac mini, and holding off on a refresh of the mini actually makes less sense now... it was more likely before 6/6.
If you buy an iMac now, you'll get a G5, and it's unlikely you'll get a better model before 1Q07... and, again, the news on 6/6 makes an upcoming refresh on the iMac less likely.
If you wait for the Intel-based models, you'll be waiting until next June. If you want an Intel-based iMac, you'll likely be waiting until 2007... they've already said they'll update the low-end models and the laptops first.
So I don't see much point in doing more than waiting a few weeks to see if the current shortage of the mini (Estimated Ship just went up to 6-8 business days) indicates a refresh with a Radeon 9600 to support QE2d under Tiger... if you were looking at the mini, OR waiting until 2H06 or 1H07 and getting an IntelliMac.
Steve Jobs could have a hell of a surprising "Just one more thing" at MWSF'06 and release an Intel box before next June, but June's what they've committed to.
Because just because he's 14 doesn't mean I buy what he wants in games. If he wants to play Doom V he'll have to play it at a friend's house - it's called parenting.
:)
So that brings us back to "How long do you keep using a computer anyway?"
He's 14 now. When will he be old enough to play Doom V? 16? 18? You're looking at replacing it in 2-4 years anyway unless it's running Game-OS. Whatever you buy now will still be well supported that long, and in the meantime you'll be saving the guy from premature exposure to Windows.
IBM sees it, and puts a TrackPoint in the top of their mice for two-axis scrolling.
The IBM "scrollpoint" mouse I tried looked just like a trackpoint on a mouse, but it was only an up-down controller. I'd love a full trackpoint on a mouse, do you have an actual model number for the 2-d version?
I won a Kensington mouse with a scroll panel a year or so back, and gave it away. The lack of feedback is really annoying.
There's a couple of better alternatives that I've never seen. One is the IBM trackpoint... the button between G and H on thinkpads. It's a small, reliable, and well understood 2d controller that would fit on a mouse really easily. IBM even put a trackpoint on a mouse... they called it a scrollpoint. Unfortunately they only made it into an up/down scroller. *doh*
The other alternative that's good is a software solution. Logitech came up with a MUCH better idea a few years back. They implemented it in the mouse software instead of the mouse, and that poorly, but the basic idea is good.
You could configure the middle button as a "grab" button. When you held it down, moving the mouse would scroll the document in all four directions. Unfortunately, it also put an obscure icon down where you clicked and didn't clean it up properly, but there's no reason that part couldn't be left out.
We're still on NT4 for our domain controllers, because going to Active Directory would force us to completely redesign our DNS environment, either to support AD with a UNIX-based DNS, or migrate all our DNS-related scripts over to Windows... and last I checked some of the things we're doing, like synchronising subdomains across split DNS, aren't actually possible in AD.
The one new feature in Windows XP that I've ever had occasion to miss on Windows 2000 is the integration of the Citrix multi-user and remote desktop support, specifically virtual consoles (AKA fast user switching). And I'm actually kind of ticked off that Microsoft restricted it to one console per user-id, because when Apple did their version they put the same stupid restriction in, instead of following the lead of BSD and Linux and just letting you have as many concurrent sessions as you have VM for.
In general, though, I agree with you. There's no killer new technology in Windows XP, just more integration of third party packages, and the user interface "improvements" aren't any such thing. Ironically, the main effect of the "improved security" is that if you're using Microsoft IRM (Intellectual Rights Management, their DRM solution) you can't use Windowblinds or any similar software to fix the "improved" user interface.
The question started out as: will the move to Intel threaten Linux. [...] Moving to Intel will not threaten anyone [...]
Great point. It's a shame that you're making it to
someone who doesn't disagree with it. Why don't you try reading what I've actually written? I didn't write "the author's conclusions are wrong". My point was "the author's argument's are irrelevant".
Even if Apple did release OS X for generic intel boxes, it wouldn't make any difference to Linux, because Linux and Mac OS X are good for different things. Because on the desktop there are only two operating systems that have any significant commercial software base. Linux isn't one of them and it has no prospect of becoming one of them any time soon, and there's no magic vaguely-hinted-at software that will show up in the next two years that will change that. Because as far as any software is concerned Linux and Mac OS X are barely distinguishable.
So... the question "is Mac OS X on intel a threat to Linux" is a category error. It's like asking "is vanilla ice cream a threat to caterpillar tractors".
I'm a UNIX guy. I'm part of the core constituency for an open source desktop operating system. I'd love to see one show up. I was using X11 and UNIX on Wintel hardware as my main desktop environment until about three years ago, and if Linux on the desktop was an option I'd be using it, because I've been using UNIX as my desktop environment of choice for longer than Linux has existed.
So, I don't use Mac OS X because it's going to knock Windows on its ass, I use Mac OS X because it's the only way to get UNIX with a desktop environment that doesn't suck. On the server side, I've got FreeBSD, Linux, Tru64, HPUX, and Solaris boxes. But none of them are an option on the desktop.
So what does "is Mac OS X a threat to Linux" mean? Mac OS X and Linux aren't in competition with each other. "One understands this is not very nice for Linux fans, and they would much rather focus on whether Mac OS X is a threat to Windows. That isn't the issue."
I didn't raise that question. The Fine Article didn't raise that question. You seem to have decided that's what you want to talk about. Well, great, have fun. But please find someone who's using Windows XP and talk about it with them, OK?
It's pretty clear that the switch to Intel means in the next year or two there are going to be new Macs that are significantly faster than today's Macs (especially the aging G4 based notebook line), and quite likely they are going to cost less than today's Macs.
Notebooks, yeh, I suppose that's likely (though closer to two years than one... there's not going to be any x86 based Macs for a year, and it's going to start at the low end of the product line), but I don't think there's going to be desktop macs faster than the iMac G5 until well into 2007, and the Powermac G5 is going to be the last update near the end of the year, and I doubt they'll be all that much cheaper than they would be with the G4 or G5: the CPU isn't what makes Macs cost that much, most of the "Mac Tax" goes to their high margins to subsidise the OS.
And the first few x86 macs... the ones you're waiting on tenterhooks for... are the ones I'd personally stay away from. They're likely to be the x86 equivalent of the Nubus Powermacs, the Rev 1 Beige G3, the "Yikes" G4.
And while they seem to have picked a pretty conservative release schedule, it could still slip.
So, unless you need the new computer right now, you'd be crazy not to wait.
Only if you completely discount the time value of money. After all, there's always going to be faster and cheaper computers coming in the future so it's always rational to wait. It's all a matter of... how long can you wait? And base your decision on the assumption that you'll have to wait at least a refresh cycle longer than that. You might luck out, and there won't be any early release bugs, but do you want to plan on it?
So... if you were planning on buying a Mac now, it may still be rational to buy one now. Especially if you've been delaying a purchase, double if you don't have a Mac at all yet, triple if you're using Windows. This is even a good time to get an eMac, since they've just had a refresh and an updated GPU for Tiger. And don't forget, Tiger is an OS sweet-spot: the next OS release, Leopard, is likely to be at least as much about changing its spots as adding new features.
You seem to imply a mutual exclusivity between a good server OS and a good desktop OS.
Not at all, it's just that there's new ideas and new ways of doing things developed all the time. When NeXT started working on NeXTSTeP in the early '90s the world was very different... there was barely an internet and no google, no apache, no Netscape or Internet Explorer, a 68030 was a pretty fast chip, and expecting a desktop OS to run multithreaded webservers when there was no web and no webservers is a but much... and since neither NeXT nor Apple has concentrated on the operating system's performance running the kinds of server apps ArsTechnica was testing...
Well, for those kinds of apps, it's like comparing a 15 year old OS to one from five-minutes-ago. That's all.
they will no longer be able to hide behind the Megahertz "myth."
What myth?
Both AMD64 boxes and PPC 970s currently get at least comparable performance to Intel P4s running at a far higher clock speed.
Megahertz == performance has always been a myth.
288 MB of system RAM.
I'm running Panther in 256M right now, but on a 1.33 GHz G4. No tweaking. It's sometimes pretty doggy: I've found that the real "knee" in the performance curve is about 360M. Once you get over that, things get a lot better.
Well, I don't agree that Linux isn't in the running for the home, but was making a different point. That is, XP and X are not in the same league.
./configure, read the error message from configure or INSTALL file, download and install the stuff it depends on, ./configure, make, make install.
OK, I get it. What's that got to do with the original message... which is that if XP and X aren't in the same league, Linux isn't in the same game as either of them? AND its market share problems are at least as bad as OS X.
That is, let's say you're absolutely right and the future of OS X is black as Bill Gates' heart.
It's still in a better position than Linux, because it's still a better option than Linux for someone who's using Windows, or looking for an alternative to Windows. Why? Because it runs almost everything Linux does, and it runs some version of almost everything Windows does, and it's easier to run any of that than on either Linux or Windows.
Installing a program on Windows XP, best case:
Download installer. Open the installer. Click "yes" or "no" a few times.
Installing a program on Windows XP, typical case.
Download installer. Open the installer. Click "yes" or "no" a few times. Reboot.
Installing a program on Linux, best case:
If it's on the CD that Linux came on, you select it from the CD, run it, then open up a terminal program and use "rpm -q" to figure out what the command to run it is.
Installing a program on Linux, typical case #1:
Download the RPM, run "rpm -i", download the packages it says it needs, "rpm -i" again, repeat a couple of times, then use "rpm -q" to figure out what the command to run it is.
Installing a program on Linux, typical case #2:
Download the tarball, unpack it, cd "program-1.4.5",
Installing a program on OS X, best case:
Download the file, double-click it, drag the icon to your applications directory.
Installing a program on OS X, typical case:
Download the file, double-click it, double-click the installer, answer a couple of questions.
Personally, I prefer unpacking a tarball and running configure to running some unpredictable GUI installer. But even for me it's hard to beat "drag icon to application directory".
There is by the way a big difference between a few percent and rising, and a few percent and falling.
Then you should be enthusiastic about OS X, because it's 2.3% and rising, and while it's going to lose some to people who freak out over the whole Intel thing that's a one-time hit unless Apple does something weird like switching back to PPC in 2009.
according to the grandparent sited article, linux is much faster than OSX (five times in some cases) in basic thread handling functions, which is very important for server performance.
That's why I started out by saying "OS X isn't designed as a server OS" (except by comparison to OS 9, which was horrid) and finished up by pointing out that a Mac OS X desktop is a great accompaniment to whatever UNIX variant you're using on your server.
That's my point. What's your point?
The Mac's for my son, and we're talking game software [...]
Um, so why was a Mac even on the shortlist in the first place? I'm a UNIX geek, I hate Windows, but I'm not crazy enough to try and convince a gamer to run anything else.
But I'm thinking my trusty old iBook [with 32M of RAM] would probably explode the first time I hit F12 to bring up dashboard.
If you've got 32MB of RAM, and you're running OS X, your geek-fu is better than mine. I've never installed OS X in less than 96M and that took over 24 hours to complete the install.
Linux is macrokernel, OS X is microkernel
Well, no, it isn't really. OS X is a "BSD Single Server" design, which avoids the overhead of a microkernel by using a conventional monolithic UNIX kernel to provide most server tasks. It's really quite efficient, and runs UNIX software very well. Where it falls down is in massively multithreaded applications because, well, Mach isn't really a microkernel either and it doesn't have the low-overhead thread creation and IPC that a microkernel really needs.
Which is why the GNU approach of using the Mach kernel as a microkernel didn't work, and the Single Server model used by Tru64, NeXTstep, and Lites works quite well. Particularly for the desktop, so, to repeat the point you apparently missed... regardless of what UNIX you use on your servers, a Mac OS X desktop is a great accompaniment to it.
The real debate is, who cares, when you have less than two percent market share?
2.3 last I checked, and rising. It'll probably go down a bit as people freak out about Intel, but unless Apple really screws up badly between now and the end of 2006 it'll be heading up again by then.
Anyway... somehow I feel compelled to ask you why you think that question applies any less to Linux than it does to Apple. Right now Linux on the desktop has a miniscule market share, and almost all of that is due to companies that have switched en-mass: Linux/UNIX strength is as a server OS, and the highly controllable environment makes it a natural for IT departments.
So why don't you think Linux tiny desktop market share makes it even LESS relevant than Apple?
I mean, I hate Windows, I'm a major UNIX booster and I've admired Apple from a distance, but until OS X came out, I was honestly unable to recommend anything but Windows to a home user who wasn't technically oriented. Mac OS 9 was horrid, and none of the free UNIXes is an option for Joe Average unless they can buy it preinstalled or get a geek friend to set it up for them. Heck, I didn't even really feel good about recommending OS X until Jaguar.
Even for business, where Linux is more practical, Windows on the desktop and UNIX in the backroom is still a rational approach. Until this year Linux had an edge on Apple because Apple simply didn't make a desktop machine with a price that made sense for a business.
But for the home? There's only two options: Mac OS X and Windows XP. Linux isn't even in the running.
No sense rushing if I'll have to buy new software anyway.
... and there's an awful lot of recently purchased Powermac G5s that are still going to represent potential customers 5 years from now given the way Moore's Law has hit the wall in the last two years (remember, IBM didn't hit 3 GHz but Intel's speed improvement over the same period was even less).
Let's see...
Apple has typically supported old hardware and operating systems for five years after a transition (68000-68020, 68k-PPC, Nubus-PCI, OS9-OSX, and hardware abandoned by OS X has generally been that far behind the curve). There's really no good reason for them to drop PPC software support this time, because they're using an OS that makes old platform support easy. There's less reason than ever for software vendors to drop the old hardware, with the single exception of game software, because by the time Leopard comes out they'll have had to become portable
And most companies are likely to provide "Intel-compatibility" upgrades fairly cheaply to avoid losing their customer base.
So if you buy a Mac now, it's going to be 7 years before Apple drops PPC, and at least 5 before your apps are unavailable for PPC... so what's the new software you're worried about having to buy?
In the meantime, I'm uninstalling OS X on my older Mac gear. Tiger left two of my machines behind.
What, without even waiting for XPostFacto 4 to get out of beta?
Tiger's so far shown itself faster on unsupported Macs than Panther, just as Panther was faster than Jaguar. Or, you can just keep using Panther... I've still got Jaguar on one of my boxes because it does everything I need there. It's not going to be any harder to install Linux on your iBook in 2012 or so when Apple finally drops support for PPC, or in 2009 when Whatevercorp drops YourCriticalApp for PPC, so why jump the gun? Pique?
The truth is, I might end up buying one or two of the software packages that I "try out" on my computer if I were prevented from obtaining them in other ways.
Yeh, I know that lots of people don't buy shareware. My own shareware non-income shows that. But if you're one of the guys who extends "trial use" indefinitely, so you don't consider lost sales a problem, why did this outrage you?