"1. If Rosetta works as well as demonstrated (Jobs showed unmodified PPC versions of Photoshop+filters and MS Office running happily and fast on the Intel Mac box) then this will be less painful than you think."
Anything CPU intensive will be iffy on Intel processors, and emulation of Altivec code is explicitly not supported by Rosetta.
". The way the Intel and PowerPC raodmaps are going I think in three uears time there will be a HUGE difference in capability. Jobs was demoing a Pentium 3.6GHz quad for God's sake!"
That was a single processor machine.
It'll be okay for most legacy apps, but anything that actually cares how fast a machine is will have to be native. There is no way around this.
Also, while Intel processors will be faster per watt, a given thread will not be faster. All the CPU manufacturors have hit a performance wall in terms of how much work can be done in one processor, and even if you have a dual-core dual-CPU machine, one thread will have to run on one core, and under emulation that means every thread will be slower.
I agree that it's unfortunate, but if IBM isn't going to be developing PowerPC processors that Apple needs, there's no choice. Better now than after a protracted period of decay.
Apple will have influence over future revisions of the SSE extensions. It'll never equal Altivec (because they'll be extensions rather than an orthogonal set of instructions that worked from the start), but if it's going to be continually improving for the forseeable future it'll ultimately be the better choice.
We can't deffinitively deny this, and anyone that can is under an NDA.
However, I think the consensus is that we'd have to see it to believe it. And that means it'll have to come from Steve's mouth, not some IT rumor mill.
"My psychic forecast is that the PowerPC would outperform the x86 setup because Apple has tuned gcc for the PowerPC platform."
Turns out that's not the case.
Intel and AMD win in integer performance (no surprise), PowerPC wins in floating point performance (no surprise). The chips won where they are strong, and lose where they are weak.
The OS dependant things like threading and I/O performance are entirely OS dependant, they don't have much to do with the compiler.
" Tell me more! I looked into it briefly this afternoon, but most forum threads started about it were left unanswered."
I don't know how it works on Windows, but on Linux you can install as many versions as you want. "python" is a symlink to the default version, and if you want a specific one you can say "python2.4" or whatever.
"Not knowing Python (I got taught Java at university) I never really payed much attention to it."
Knowing any one language is a very dangerous specialization. Java isn't going to be in vogue forever.
The ability to learn languages trumps specialization in any one language in about 99.999% of cases.
-Get preferences from admins (priority, approximate update frequency). -Get metadata (time of last update)
They can tell most of that just by downloading the page regularly but with 8 billion pages it's probably pretty hard to do every one of them with any frequency and most of them probably change a few times a year if that.
Now they can tell if a page has changed and whether it's likely to change in the future with a few kb of gzipped xml instead of megabytes of HTML.
They've opened the format, so it seems likely that everyone else will be supporting it soon. It's too much of a competitive disadvantage to anyone that doesn't take advantage of it, and other protocols are less likely to be supported by any give site because this is first, open, and simple.
If a distro is really terrible and I want nothing to do with it, I feel perfectly free to criticize it without the slightest intention of fixing anything.
For example, I'm not going to stop criticizing Gentoo because I have specific and accurate complaints. I'm not going to help fix Gentoo because that would involve using it.
I'd rather use something else and leave Gentoo to sink or swim. Hoepfully, my criticism will let the maintainers know their focus isn't working for some things, or at least make people think twice before they take the claims of Gentoo evangelists at face value.
These lawsuits protect the consumer. They impose a financial cost on selling bad hardware, which Apple does regularly. I'm disapointed that laywers get most of the compensation, but at the end of the day if Apple has to spend millions of dollars of lawsuits because of flawed hardware they will eventually start spending more on QA to prevent future lawsuits.
"I'm not going for my compensation because I still believe it was my fault."
I expect Apple to replace the logic board for my iBook every 3-6 months and I expect them to thank me for remaining their customer after they knowingly sold me hardware with such a critical flaw.
Apple is not a religion. Apple is a company and they'll screw you if you don't stay on your toes, just like every other company out there.
Except for a few very specific niche areas, a G4 is completely outclassed by every other processor being sold in desktop machines today.
Being obsolete doesn't necessarily make a CPU useless. I have an older CPU I use daily in a firewall and server. I'm not going to throw it out, but I'm certainly not going to buy one new.
I've seen it too many times to count, and I don't get around that much. Need a server for some random task? No more money in the budget? Find an abandoned box, get someone to admit they don't need it, throw Debian-stable on it and leave it alone until the hardware dies. Repeat as necessary, because there won't be any money the next time either.
That doesn't mean much in terms of sales, but it increases shareholder value without any additional spending.
"But they are here to stay. The term "blog" may be a bit of a fad like "cyberspace". But it's basic form isn't going to go away any time soon. That's because they do one thing very well: help people communicate."
You're correct of course. I was basically making fun of the depths companies will sink to to pander to investors. Investors will flee over transitory problems, so companies need a way to say "everything is fine" more frequently and more publically. The message hasn't changed, but they think finding a new way to present it will make it more credible.
It's been possible since Internet access because widespread, but until someone automated it and called it a "blog" companies thought of it as a press release that was available online.
Now it's just another fashion to coopt for marketting reasons. These people will make lots of money for a months until everyone realizes that anyone can write blogs -- that's the entire point. Then it'll be just another job requirement for employees.
512 is enough for a few major applications and the usual background type stuff. A gig is enough for everything you want to run simultaneously (for most people's definition of everything, obviously some users need more),
For the set of applications I run, a gig is enough for me to not worry about it anymore. 512 isn't.
It's going to get worse when.NET/Java applications become common on Linux. I dunno what the situation is on Windows, they might already be common.
Slashdot Mirror
"1. If Rosetta works as well as demonstrated (Jobs showed unmodified PPC versions of Photoshop+filters and MS Office running happily and fast on the Intel Mac box) then this will be less painful than you think."
Anything CPU intensive will be iffy on Intel processors, and emulation of Altivec code is explicitly
not supported by Rosetta.
". The way the Intel and PowerPC raodmaps are going I think in three uears time there will be a HUGE difference in capability. Jobs was demoing a Pentium 3.6GHz quad for God's sake!"
That was a single processor machine.
It'll be okay for most legacy apps, but anything that actually cares how fast a machine is will have to be native. There is no way around this.
Also, while Intel processors will be faster per watt, a given thread will not be faster. All the CPU manufacturors have hit a performance wall in terms of how much work can be done in one processor, and even if you have a dual-core dual-CPU machine, one thread will have to run on one core, and under emulation that means every thread will be slower.
I agree that it's unfortunate, but if IBM isn't going to be developing PowerPC processors that Apple needs, there's no choice. Better now than after a protracted period of decay.
Apple will have influence over future revisions of the SSE extensions. It'll never equal Altivec (because they'll be extensions rather than an orthogonal set of instructions that worked from the start), but if it's going to be continually improving for the forseeable future it'll ultimately be the better choice.
No, "SSH" has been around for a long time, it predates the OpenSSH client and probably your website.
We can't deffinitively deny this, and anyone that can is under an NDA.
However, I think the consensus is that we'd have to see it to believe it. And that means it'll have to come from Steve's mouth, not some IT rumor mill.
No one, not even the article, is denying that MacOS is preferable as a desktop or workstation OS most of the time.
It's the server tasks that need work, so it may be advisible to avoid XServes.
"However, the G5s highest LW Raytracing performance is comforting, as I just bought an LW 8 license for Mac OS X the other week :-)"
The G5 wins in floating point performance, so that's easy to believe.
The server performance is dependant on the kernel, and Linux gets an easy win there.
"They say Macs are bought by creative profesisonnal so they will test open source solution such as MySQL and Apache!!!???"
Actually, they say CPU-heavy workstation tasks will be just fine on a Mac. It's server tasks that are a problem.
"My psychic forecast is that the PowerPC would outperform the x86 setup because Apple has tuned gcc for the PowerPC platform."
Turns out that's not the case.
Intel and AMD win in integer performance (no surprise), PowerPC wins in floating point performance (no surprise). The chips won where they are strong, and lose where they are weak.
The OS dependant things like threading and I/O performance are entirely OS dependant, they don't have much to do with the compiler.
In this case it's moot, because Linux wins easily in the OS dependant benchmarks.
" Tell me more! I looked into it briefly this afternoon, but most forum threads started about it were left unanswered."
I don't know how it works on Windows, but on Linux you can install as many versions as you want. "python" is a symlink to the default version, and if you want a specific one you can say "python2.4" or whatever.
"Not knowing Python (I got taught Java at university) I never really payed much attention to it."
Knowing any one language is a very dangerous specialization. Java isn't going to be in vogue forever.
The ability to learn languages trumps specialization in any one language in about 99.999% of cases.
I think that's a journalist that doesn't know what "hyper-threading" means, so they just dropped it.
Multiple versions of Python can coexist on a machine...
It seems there's two goals:
-Get preferences from admins (priority, approximate update frequency).
-Get metadata (time of last update)
They can tell most of that just by downloading the page regularly but with 8 billion pages it's probably pretty hard to do every one of them with any frequency and most of them probably change a few times a year if that.
Now they can tell if a page has changed and whether it's likely to change in the future with a few kb of gzipped xml instead of megabytes of HTML.
They've opened the format, so it seems likely that everyone else will be supporting it soon. It's too much of a competitive disadvantage to anyone that doesn't take advantage of it, and other protocols are less likely to be supported by any give site because this is first, open, and simple.
Well, I noticed two things about it...
First, the priority is a relative priority, so if you want to set every page to 1.0 (defined as the highest priority) it'll mean nothing.
Second, if you lie about update frequency or the date of the last update they'll figure it out pretty quick.
These aren't commands, they're hints.
That's absurd.
If a distro is really terrible and I want nothing to do with it, I feel perfectly free to criticize it without the slightest intention of fixing anything.
For example, I'm not going to stop criticizing Gentoo because I have specific and accurate complaints. I'm not going to help fix Gentoo because that would involve using it.
I'd rather use something else and leave Gentoo to sink or swim. Hoepfully, my criticism will let the maintainers know their focus isn't working for some things, or at least make people think twice before they take the claims of Gentoo evangelists at face value.
I am not inclined to be so forgiving.
These lawsuits protect the consumer. They impose a financial cost on selling bad hardware, which Apple does regularly. I'm disapointed that laywers get most of the compensation, but at the end of the day if Apple has to spend millions of dollars of lawsuits because of flawed hardware they will eventually start spending more on QA to prevent future lawsuits.
"I'm not going for my compensation because I still believe it was my fault."
I expect Apple to replace the logic board for my iBook every 3-6 months and I expect them to thank me for remaining their customer after they knowingly sold me hardware with such a critical flaw.
Apple is not a religion. Apple is a company and they'll screw you if you don't stay on your toes, just like every other company out there.
Except for a few very specific niche areas, a G4 is completely outclassed by every other processor being sold in desktop machines today.
Being obsolete doesn't necessarily make a CPU useless. I have an older CPU I use daily in a firewall and server. I'm not going to throw it out, but I'm certainly not going to buy one new.
There's not really any way to measure it.
I've seen it too many times to count, and I don't get around that much. Need a server for some random task? No more money in the budget? Find an abandoned box, get someone to admit they don't need it, throw Debian-stable on it and leave it alone until the hardware dies. Repeat as necessary, because there won't be any money the next time either.
That doesn't mean much in terms of sales, but it increases shareholder value without any additional spending.
"But they are here to stay. The term "blog" may be a bit of a fad like "cyberspace". But it's basic form isn't going to go away any time soon. That's because they do one thing very well: help people communicate."
You're correct of course. I was basically making fun of the depths companies will sink to to pander to investors. Investors will flee over transitory problems, so companies need a way to say "everything is fine" more frequently and more publically. The message hasn't changed, but they think finding a new way to present it will make it more credible.
It's been possible since Internet access because widespread, but until someone automated it and called it a "blog" companies thought of it as a press release that was available online.
I can remember when blogging was cool.
Now it's just another fashion to coopt for marketting reasons. These people will make lots of money for a months until everyone realizes that anyone can write blogs -- that's the entire point. Then it'll be just another job requirement for employees.
Wait, that's not true. Blogs were never cool.
512 is enough for a few major applications and the usual background type stuff. A gig is enough for everything you want to run simultaneously (for most people's definition of everything, obviously some users need more),
.NET/Java applications become common on Linux. I dunno what the situation is on Windows, they might already be common.
For the set of applications I run, a gig is enough for me to not worry about it anymore. 512 isn't.
It's going to get worse when
How much is that dollar amount offset due to the fact that you don't necessarily have to pay for Linux?
Firefox with a large number of windows and tabs (how I like it) dwarfs any WM I could use. More memory is a better solution.
A bus faster than an Athlon had in 1999 might also be nice.
Audio inputs are a "premium" feature. Apple likes to leave out "premium" features even if they're free (like dual video outputs on iBooks and iMacs).