Time to post the obligatory Steve Albini reference. For perhaps the first honest view of the way the music industry does business... check out this essay.
1(a). You're right, of course, but there *is* a movement towards swarms of smaller bitty boxes for problems that don't require a) 64-bit address spaces and b) aren't tied to Big Iron software vendors.
1(b). Sorry, no. There's no comparison in speed between a big x86 box and even the very fastest G4. Altivec *is* marvelous, clearly the best SIMD engine available, but where're the vectorizing compilers for OS X? (*sound of crickets chirping*) This has nothing to do with the "Mhz myth", and everything to do with the much larger R&D budgets at AMD and Intel, WRT Moto semi. Just look at SPEC. x86 is gross -- true. The 7400 series from Moto is a nice design -- no argument. But in performance, general purpose real-world performance, x86 wins.
3. My point was that asking for OS X on x86 is pointless. And it's true that, while Apple has been doing yeoman's work at improving their ISV relations, they've only looked good in comparison to their historical performance. Apple is *famous* for alienating their developers, and to Jobs' credit, this is one area where he's markedly improved the company's image.
There isn't a debate if you're just starting out; but the whole reason to have Carbon is to make the job of porting pre-existing applications over. You'll recall that Carbon wasn't on the roadmaps until there was a (relatively) huge uproar on the part of Mac developers, when they heard that they were going to have to move their applications over to Obj-C and the Next frameworks.
Personally speaking, I'd rather poke my own eyes out with a fork than use Carbon (Cocoa, OTOH, is lovely), but if you're, say, Macromedia, or Adobe, and you've got bignum lines of Powerplant or whatever invested in the old-school system, I can see the rationale.
"This may sound like flamebait, but it's more of a rebuttal;)"
Actually, it sounds pretty reasonable to me.
1. The rumored G5 is a joke. No, bear with me. The numbers bandied about for the (illusory) G5 at 1.6 ghz are preposterous (1342 SpecInt2000 and 1364 SpecFP2000 -- ha!), more realistic numbers are well under the current performance of high-end x86, and Moto's semi division is in even more trouble than Carly Fiorina.
2. Yes, Linux and OS X are niche markets; different niches. Linux is squeezing Solaris, and OS X is squeezing... OS 9.
"I merely disagree with your position that OS X was designed secondarily as a server platform."
I'd say that Apple has nil interest in OS X as a server platform, actually. Mach + BSD is a holdover from NeXT that carpetbagged along with the really cool stuff that Apple bought: the OpenStep runtime, and more importantly, Steve Jobs. The fact that OS X can probably serve as a decent little server OS (albeit running on toy hardware), is tangential.
Upon rereading your post, I think that we probably aren't really disagreeing.
1. Unless Motorola (ha!) or IBM (more likely, but still... ha!) can close the performance gap with commodity x86 hardware, the scientific computing market will stick with the bang for the buck that the beige box world provides.
2. Neither Linux (currently technically incapable) or OS X (incompatible hardware) are in a position to challenge MS for the commodity desktop. This situation is not likely to change any time soon.
3. OS X will/never/ be ported to x86. Firstly, Apple has no interest in alienating MORE developers with yet another giant architectural switch-over. They're going to have enough trouble getting people to drop Carbon in favor of Cocoa without having to try and convince ISVs to start their projects over on a whole new hardware platform. And secondly, Apple makes the lion's share of their money from HARDWARE sales. Their position in the industry is unique, and they're not interested in being either Be (a dead OS provider for x86) or Compaq (a soon to be dead assembler of beige boxes).
Yeah, Smalltalk does, and Obj-C doesn't. I suppose Stepstone didn't want to have to try and reconcile the Pointer Dance with a real garbage collector. Pity.
Yeah, MySQL works great for slashdot. That's why it falls on it's ass all the time -- ever get a static page, even when you're logged in? Welcome to MySQL, here's your accordion.
Or turn them off entirely, and when you come across a site that uses Javascript/Flash/etc. responsibly, add it to your "Trusted Sites" list. IE actually handles this surprisingly well, although it could be easier to move domains from untrusted to trusted and back again.
Could you remotely log into your DOS/Windows 3.* machine while someone else uses the console? No? Cause you can do that in Windows XP. Out of the box. Can you run your system on a very fast, robust, journaling filesystem? Cause NTFS is pretty damn good. Does DOS/16-bit Windows have an SMP kernel? Does it support proper memory protection? Threads? No? Thought not.
I don't particularly/like/ Windows, less so Microsoft and their business practices, but to claim it's the same as DOS is straight up trollish pig-ignorance. It's like claiming that all the advances in Unix since AT&T are meaningless fluff.
And the only reason that graphic artists prefer the Mac is brand-loyalty? The querant is goofy-stupid. Have his friend entrust her money to someone who isn't a dipshit, that's my advice.
Did the judge order the entire DOInet down? No, not if that/direct quote/ from the decision is right. If the DOI pulled everything down, that's their own lookout. Stupid? Perhaps. But don't point the finger at the judge for that.
Well, the only place where a 500mhz G4 is within spitting distance of a 1ghz CuMine is when running tons and tons and tons of vector ops. In other words, Altivec. Which, as you astutely note, the G3 lacks.
Conclusion? A 500mhz G3 is never within spitting distance of a 1Ghz P3/K7. Sorry, them's the facts. There's plenty to like about the Motorola line (especially if you're building embedded systems), but they just can't keep up with x86, speed-wise.
The two best lightweight window managers are lwm and ion. lwm is like an even lighterweight twm; you can start xterms, move windows, close windows, and make windows disappear into a right-click menu. That's it; that's all it does. As an added bonus, the binary is something like 26k on an x86 FreeBSD box.
ion is totally different; it takes over your entire screen, which is very nice for limited resolutions where you don't want pixels wasted on borders, widgets, etc.; it's controllable completely through the keyboard; it looks nothing like any other window manager (in my eyes, a big benefit, as all other window managers are broken); however, it's also quite light on the system.
I also hear good things about wm2 and wmx, but I can't stand the little tiny title bar on the side, so I don't use 'em.
You're probably thinking of lwm. In the description of the various X resources that lwm understands, the author says, "In addition to these ample facilities, there exists a powerful configuration tool called gcc."
It's a great window manager, although I've moved to Ion, because I hate having to use the mouse.
lwm : http://www.boognish.org.uk/enh/lwm/
ion : http://www.students.tut.fi/~tuomov/ion/
Slashdot Mirror
Time to post the obligatory Steve Albini reference. For perhaps the first honest view of the way the music industry does business ... check out this essay.
Peace,
(jfb)
Well done.
(jfb)
1(a). You're right, of course, but there *is* a movement towards swarms of smaller bitty boxes for problems that don't require a) 64-bit address spaces and b) aren't tied to Big Iron software vendors.
1(b). Sorry, no. There's no comparison in speed between a big x86 box and even the very fastest G4. Altivec *is* marvelous, clearly the best SIMD engine available, but where're the vectorizing compilers for OS X? (*sound of crickets chirping*) This has nothing to do with the "Mhz myth", and everything to do with the much larger R&D budgets at AMD and Intel, WRT Moto semi. Just look at SPEC. x86 is gross -- true. The 7400 series from Moto is a nice design -- no argument. But in performance, general purpose real-world performance, x86 wins.
3. My point was that asking for OS X on x86 is pointless. And it's true that, while Apple has been doing yeoman's work at improving their ISV relations, they've only looked good in comparison to their historical performance. Apple is *famous* for alienating their developers, and to Jobs' credit, this is one area where he's markedly improved the company's image.
Peace,
(jfb)
There isn't a debate if you're just starting out; but the whole reason to have Carbon is to make the job of porting pre-existing applications over. You'll recall that Carbon wasn't on the roadmaps until there was a (relatively) huge uproar on the part of Mac developers, when they heard that they were going to have to move their applications over to Obj-C and the Next frameworks.
Personally speaking, I'd rather poke my own eyes out with a fork than use Carbon (Cocoa, OTOH, is lovely), but if you're, say, Macromedia, or Adobe, and you've got bignum lines of Powerplant or whatever invested in the old-school system, I can see the rationale.
Peace,
(jfb)
"This may sound like flamebait, but it's more of a rebuttal ;)"
... OS 9.
Actually, it sounds pretty reasonable to me.
1. The rumored G5 is a joke. No, bear with me. The numbers bandied about for the (illusory) G5 at 1.6 ghz are preposterous (1342 SpecInt2000 and 1364 SpecFP2000 -- ha!), more realistic numbers are well under the current performance of high-end x86, and Moto's semi division is in even more trouble than Carly Fiorina.
2. Yes, Linux and OS X are niche markets; different niches. Linux is squeezing Solaris, and OS X is squeezing
3. We're in total agreement.
Peace,
(jfb)
"I merely disagree with your position that OS X was designed secondarily as a server platform."
I'd say that Apple has nil interest in OS X as a server platform, actually. Mach + BSD is a holdover from NeXT that carpetbagged along with the really cool stuff that Apple bought: the OpenStep runtime, and more importantly, Steve Jobs. The fact that OS X can probably serve as a decent little server OS (albeit running on toy hardware), is tangential.
Upon rereading your post, I think that we probably aren't really disagreeing.
Peace,
(jfb)
Three points.
... ha!) can close the performance gap with commodity x86 hardware, the scientific computing market will stick with the bang for the buck that the beige box world provides.
/never/ be ported to x86. Firstly, Apple has no interest in alienating MORE developers with yet another giant architectural switch-over. They're going to have enough trouble getting people to drop Carbon in favor of Cocoa without having to try and convince ISVs to start their projects over on a whole new hardware platform. And secondly, Apple makes the lion's share of their money from HARDWARE sales. Their position in the industry is unique, and they're not interested in being either Be (a dead OS provider for x86) or Compaq (a soon to be dead assembler of beige boxes).
1. Unless Motorola (ha!) or IBM (more likely, but still
2. Neither Linux (currently technically incapable) or OS X (incompatible hardware) are in a position to challenge MS for the commodity desktop. This situation is not likely to change any time soon.
3. OS X will
Peace,
(jfb)
> What's wrong with them?
From the Python documentation:
"[lambda forms] are syntactically restricted to a single expression."
At least the language is lexically scoped now, which makes things slightly less difficult.
Peace,
(jfb)
> i already love python in functional programming mode
It'd sure be nice if Python had useful lambdas, though.
Peace,
(jfb)
No, linux is not a viable alternative, even where budgets are tight. There, that wasn't so hard.
Peace,
(jfb)
USA Today made a factual error? Never!
(jfb)
Support for mounting filesystems as directories is an NTFS feature, and was available in NT 5.0 (Windows 2000).
Peace,
(jfb)
Yes it is. It's horrible.
(jfb)
Yeah, Smalltalk does, and Obj-C doesn't. I suppose Stepstone didn't want to have to try and reconcile the Pointer Dance with a real garbage collector. Pity.
Peace,
(jfb)
Objective-C sans C would most certainly be cleaner than Java, as it would be, more or less, Smalltalk.
Peace,
(jfb)
Yeah, MySQL works great for slashdot. That's why it falls on it's ass all the time -- ever get a static page, even when you're logged in? Welcome to MySQL, here's your accordion.
Peace,
(jfb)
Saw one this morning. Supremely annoying. Off goes nytimes.com to the "Restricted Sites" zone.
Peace,
(jfb)
Or turn them off entirely, and when you come across a site that uses Javascript/Flash/etc. responsibly, add it to your "Trusted Sites" list. IE actually handles this surprisingly well, although it could be easier to move domains from untrusted to trusted and back again.
Peace,
(jfb)
Could you remotely log into your DOS/Windows 3.* machine while someone else uses the console? No? Cause you can do that in Windows XP. Out of the box. Can you run your system on a very fast, robust, journaling filesystem? Cause NTFS is pretty damn good. Does DOS/16-bit Windows have an SMP kernel? Does it support proper memory protection? Threads? No? Thought not.
/like/ Windows, less so Microsoft and their business practices, but to claim it's the same as DOS is straight up trollish pig-ignorance. It's like claiming that all the advances in Unix since AT&T are meaningless fluff.
I don't particularly
Peace,
(jfb)
And the only reason that graphic artists prefer the Mac is brand-loyalty? The querant is goofy-stupid. Have his friend entrust her money to someone who isn't a dipshit, that's my advice.
Peace,
(jfb)
Did the judge order the entire DOInet down? No, not if that /direct quote/ from the decision is right. If the DOI pulled everything down, that's their own lookout. Stupid? Perhaps. But don't point the finger at the judge for that.
Peace,
(jfb)
Well, the only place where a 500mhz G4 is within spitting distance of a 1ghz CuMine is when running tons and tons and tons of vector ops. In other words, Altivec. Which, as you astutely note, the G3 lacks.
Conclusion? A 500mhz G3 is never within spitting distance of a 1Ghz P3/K7. Sorry, them's the facts. There's plenty to like about the Motorola line (especially if you're building embedded systems), but they just can't keep up with x86, speed-wise.
Peace,
(jfb)
lwm : http://www.boognish.org.uk/enh/lwm/
ion : http://www.students.tut.fi/~tuomov/ion/
(jfb)
The two best lightweight window managers are lwm and ion. lwm is like an even lighterweight twm; you can start xterms, move windows, close windows, and make windows disappear into a right-click menu. That's it; that's all it does. As an added bonus, the binary is something like 26k on an x86 FreeBSD box.
ion is totally different; it takes over your entire screen, which is very nice for limited resolutions where you don't want pixels wasted on borders, widgets, etc.; it's controllable completely through the keyboard; it looks nothing like any other window manager (in my eyes, a big benefit, as all other window managers are broken); however, it's also quite light on the system.
I also hear good things about wm2 and wmx, but I can't stand the little tiny title bar on the side, so I don't use 'em.
Peace,
(jfb)
You're probably thinking of lwm. In the description of the various X resources that lwm understands, the author says, "In addition to these ample facilities, there exists a powerful configuration tool called gcc."
It's a great window manager, although I've moved to Ion, because I hate having to use the mouse.
lwm : http://www.boognish.org.uk/enh/lwm/
ion : http://www.students.tut.fi/~tuomov/ion/
(jfb)