Oh, a few other points:
- a browser is not a simple tiny thing anymore.
- Mozilla is not just a web browser.
- you didn't pay (and will likely never pay) for Mozilla.
- the Quick Launch feature works well.
Well, 0.97 takes 8 seconds on my 3 year old CAN$5000 laptop (P2/366). I could easily buy a computer today for CAN$1000 that beats this laptop senseless.
So you either bought your computer a long time ago, or some component is limiting performance in some way (i.e. you got screwed).
I had 2.4.12 running on a Via chipset motherboard (Asus K7M) with DMA and it was just fine. 2.4.7 crashed when I enabled the driver for my controller in the kernel however, so I upgraded to 2.4.12 and everything was awesome. 10MB/s on an older 5.1GB Quantum drive. 20MB/s on a year old 25G Maxtor.
You should back up your important files, and then try it. Immediately after backups, corruption can do you no harm, except for loss of time. But the possible increase in speed using UltraDMA outweighs any risk.
If you don't take into consideration sample rate when you compare the numbers, yes, a 1080p frame is bigger than a 1080i field.
Otherwise, 1080p at 30 frames/s requires exactly the same amount of bandwidth as 1080i at 60 fields/s, regardless of the way you look at it, video bandwidth or data bitrate. It's just a question of sample rate and sample size, double one and halve the other and the end result is still the same.
In fact, 24fps 1080p requires LESS bandwidth than either of those, and is more accurate for sources that are truly 24 frames per second, such as film, than 30 frames or 60 fields can EVER be.
I agree that in practice, it's not done and everything is 1080i, but that's because the tv stations are broadcasting in 1080i (or 720p), not because it can't be done.
When you say "the chips required to do it", what part do you really mean? The decoder, the display? A Hollywood Plus mpeg decoder will decode up to 10Mbits/s streams and costs about US$50 (complete), and that's three year old tech. Current chips will decode 1080i HDTV just fine. My Geforce2 MX will do 2048 x 1536 progressive, and that cost around US$100. I say current tech can do it, it's just that the display manufacturers and tv stations are entrenched in 1080i.
I seem to remember looking at the spec one time, and 1080p was in there, albeit at 24fps and 30fps. This resolution and rate would be just fine for movies.
Yes, but what filesystem format will it use? FAT, FAT32, or something else windows compatible? Or will it use HFS to be easily compatible with Mac OS machines? I think the latter.
You probably won't be able to plug it into a Windows machine and have it work right away. But Linux does have HFS support. Hmmm.
Hmmm, your two comments seem to contradict each other. If you're printing a 4M pixel image at 8x10 that would give you roughly 215dpi. I thought anything that low resolution didn't look like a photo?!
Also, supposedly prints from a photo lab give prints that are approximately 300dpi. How is a photo printer going to be worse quality (at least in terms of resolution)?
I think you misjudge the actual (print) resolution required to get photo quality results.
That's not saying that photo printers (and prints) are inexpensive and/or have high quality, since I have not seen the quality myself, but from resolution alone 300dpi sounds okay to me.
I would say FTP over the internet is a medium that is customary.
Nowhere does it say the medium has to be physical, it just says the/charge/ can be no more than your cost of performing physical distribution. If there is no charge, this is satisfied.
And, provided there is a written offer (which "go to ftp site at ftp.blah.com, the source is in the root" is), the licensor satisfies the binary distribution requirements of the GPL.
Just because RMS says so, doesn't make it that way. I would say a judge would have to decide.
Hmmm those x86 numbers doesn't quite seem right. I get 1.04 Mkeys/s on a Pentium II/366 laptop (IBM Thinkpad 600E 2645-5AU). And my Athlon 600 desktop gets 2.05 Mkeys/s.
I don't do that on my laptop because I don't think it makes sense to run dnetc, not because I couldn't.
I agree on that Altivec kicks ass on RC5-64 though. Very impressive.
That reminds me of Blade Runner, looking at Rachel's personal photos. Having the one where they are sitting on the porch suddenly move was a great shocker.
But think of digital picture frames and digital cameras today, which can express and capture a notion of time. It's not unreasonable to think that in 20 years, we'll have a printing process (digital paper) that is lifelike.
If a car was downloadable over the net, I would expect it to be given to me at a price covering development cost and reasonable profit, but not production (duplication and assembly line stuff) since there is no cost involved there. This would lower profits and thus taxes from the large car manufacturers, but the old style "hard car" profit shouldn't be afforded protection (inability for consumers to download cars) from their own inability to adapt to the technology of the day.
Laws should not protect businesses from competitors in a capitalist society, actual innovation, quality and value should. The fact that the car companies have banded together in a anti car downloading Car Industry Association proves that there is no competition between those companies in the first place. Even though it is legal to download a car from another manufacturer, the CIAA would sue to stop all car downloading, legal or not.
Actually I would hope the government would give at least one model of car away for free at that point, or I could go to the library and check out a car. I personally would buy an indie car if given the chance.
What I want to know is this: what percentage of Macintosh users actually use Photoshop, and what percentage use it the majority of the time on their computer?
I think less than five percent even use Photoshop, being generous in my guess.
The percentage will undoubtedly be higher for PowerMac purchasers rather than iMac purchasers, but I would still say it's less than 30% of those (PowerMac) users.
Only if a majority of users use Photoshop intensively would the use of Photoshop to indicate general performance be acceptable.
What Apple should do is script Office, an email application, a web browser, and any other commonly used/cross-platform/ applications (not the Finder) these days, and compare the performance between machines then. These applications/are/ used by a majority of PowerMac owners.
Never mind the fact that a GHz machine is not the fastest you can get. Think multiple processor Intel boxes. They exist, Apple.
Apple's being extremely disingenuous in regards to performance testing, and it's just one reason I laugh at anyone (excepting Photoshop users) believing anything Apple says regarding performance just because they have a flashy demo. Run the Photoshop comparisons at Seybold or similar media conferences, but not at MacWorld, please.
I have Documents and Settings at the same level as my WINNT folder, just FYI. I guess this makes it easier to completely reinstall the operating system without losing user files.
I've never had Windows Update update any drivers, and i'm not sure it's supposed to.
Nonetheless, there should be a GLSetup-like thing for *nix, which configures your system such you don't have to go to some irc channel or obscure web page for the working driver, which makes sure your configuration is how it's supposed to be. But have it be not just for OpenGL, but for all drivers required, such as joysticks, sound cards, etc.
USB support out of the box is coming in 2.4. It takes a bit to get the backport to 2.2 working but after you do it's quite nice.
Antialiased fonts are not needed for games. Full screen anti-aliasing (implemented right now by at least nVidia) solves this problem.
A little aside: has anyone got the q3a mouse acceleration working in Linux as it does in the Windows version? Like, no acceleration?
We are producing two (similar) products, and the word from the publisher is that they want both products in one box, on one cd even, with an installer that allows you to choose between them.
Supposedly the distributors are pushing for smaller DVD style packaging for software as well. I think it is that particular publisher (Microsoft) that is pushing for the larger packaging, since there definitely is not a distributor or retailer demand for such a thing.
Where has this been announced? The "from scratch" part as well as the "Cocoa" part.
By the way, Cocoa is a programing interface accessible from both Objective-C++ and Java (through a bridge), as well as other languages such as Perl or Python via third party utilities.
That's the rub: if only you didn't have to buy the hardware from Apple.
Also, by supporting them, you support closed source programming APIs. While Cocoa is by far the most ideal API to code for, it's not worth the hassle when the owner of a closed source program "deprecates" what you need the most. NeXT was notorious for dropping and changing APIs left and right, and Apple doesn't have a much better track record. Consider the very recent news about GameSprockets being only slightly supported in Mac OS X. Remember the ongoing pain unix users go through when trying to play a QuickTime movie.
The probability of Apple opening Cocoa or Carbon is relatively small, as those are intended to be their next generation APIs.
I suggest GNUstep for coders who want to use an API very similar to Cocoa. While probably a bit tougher to get into because of the lack of an integrated development environment specifically made for GNUstep, there are things in development (Project Center, Interface Builder) that will eventually make writing apps (or whatever) simpler.
Just to confirm: Carbon will be used to port from Mac OS. IE 5.1 is already ported to Mac OS X, and is included on the DP4 distribution.
> I just hope Apple keeps the networked aspect of > the rendering system or makes it easy to > extend to do networking.
They have removed the networking aspect, while I believe they've stated that the option will exist for a third party to implement this functionality, as usually is the line Apple states after removing some functionality from their new OS.
If Mac OS X ever comes to Intel, I know would be interested in working on it again, particularly in the area of networked graphics, as well as many other areas which are of no interest to me because of the limited deployment platforms of this OS.
Slashdot Mirror
Please, tell me you at least use rechargables.
Oh, a few other points:
- a browser is not a simple tiny thing anymore.
- Mozilla is not just a web browser.
- you didn't pay (and will likely never pay) for Mozilla.
- the Quick Launch feature works well.
Well, 0.97 takes 8 seconds on my 3 year old CAN$5000 laptop (P2/366). I could easily buy a computer today for CAN$1000 that beats this laptop senseless.
So you either bought your computer a long time ago, or some component is limiting performance in some way (i.e. you got screwed).
I had 2.4.12 running on a Via chipset motherboard (Asus K7M) with DMA and it was just fine. 2.4.7 crashed when I enabled the driver for my controller in the kernel however, so I upgraded to 2.4.12 and everything was awesome. 10MB/s on an older 5.1GB Quantum drive. 20MB/s on a year old 25G Maxtor.
You should back up your important files, and then try it. Immediately after backups, corruption can do you no harm, except for loss of time. But the possible increase in speed using UltraDMA outweighs any risk.
If you don't take into consideration sample rate when you compare the numbers, yes, a 1080p frame is bigger than a 1080i field.
Otherwise, 1080p at 30 frames/s requires exactly the same amount of bandwidth as 1080i at 60 fields/s, regardless of the way you look at it, video bandwidth or data bitrate. It's just a question of sample rate and sample size, double one and halve the other and the end result is still the same.
In fact, 24fps 1080p requires LESS bandwidth than either of those, and is more accurate for sources that are truly 24 frames per second, such as film, than 30 frames or 60 fields can EVER be.
I agree that in practice, it's not done and everything is 1080i, but that's because the tv stations are broadcasting in 1080i (or 720p), not because it can't be done.
When you say "the chips required to do it", what part do you really mean? The decoder, the display? A Hollywood Plus mpeg decoder will decode up to 10Mbits/s streams and costs about US$50 (complete), and that's three year old tech. Current chips will decode 1080i HDTV just fine. My Geforce2 MX will do 2048 x 1536 progressive, and that cost around US$100. I say current tech can do it, it's just that the display manufacturers and tv stations are entrenched in 1080i.
I seem to remember looking at the spec one time, and 1080p was in there, albeit at 24fps and 30fps. This resolution and rate would be just fine for movies.
Yes, but what filesystem format will it use? FAT, FAT32, or something else windows compatible? Or will it use HFS to be easily compatible with Mac OS machines? I think the latter.
You probably won't be able to plug it into a Windows machine and have it work right away. But Linux does have HFS support. Hmmm.
Hmmm, your two comments seem to contradict each other. If you're printing a 4M pixel image at 8x10 that would give you roughly 215dpi. I thought anything that low resolution didn't look like a photo?!
Also, supposedly prints from a photo lab give prints that are approximately 300dpi. How is a photo printer going to be worse quality (at least in terms of resolution)?
I think you misjudge the actual (print) resolution required to get photo quality results.
That's not saying that photo printers (and prints) are inexpensive and/or have high quality, since I have not seen the quality myself, but from resolution alone 300dpi sounds okay to me.
I would say FTP over the internet is a medium that is customary.
/charge/ can be no more than your cost of performing physical distribution. If there is no charge, this is satisfied.
Nowhere does it say the medium has to be physical, it just says the
And, provided there is a written offer (which "go to ftp site at ftp.blah.com, the source is in the root" is), the licensor satisfies the binary distribution requirements of the GPL.
Just because RMS says so, doesn't make it that way. I would say a judge would have to decide.
Hmmm those x86 numbers doesn't quite seem right. I get 1.04 Mkeys/s on a Pentium II/366 laptop (IBM Thinkpad 600E 2645-5AU). And my Athlon 600 desktop gets 2.05 Mkeys/s.
I don't do that on my laptop because I don't think it makes sense to run dnetc, not because I couldn't.
I agree on that Altivec kicks ass on RC5-64 though. Very impressive.
That doesn't make a difference to Microsoft. Washington State isn't sueing them either.
It's been less than two years since the Athlon has been released. How impatient are you?
That reminds me of Blade Runner, looking at Rachel's personal photos. Having the one where they are sitting on the porch suddenly move was a great shocker.
But think of digital picture frames and digital cameras today, which can express and capture a notion of time. It's not unreasonable to think that in 20 years, we'll have a printing process (digital paper) that is lifelike.
Sorry, you're wrong.
/could/ get Foundation for NEXTSTEP 3.2. I think it started shipping preinstalled with 3.3.
You
I know for sure because that's the only way OmniWeb 2.x would run on those OSes, and i'm absolutely sure I ran OmniWeb 2.x on 3.2.
If a car was downloadable over the net, I would expect it to be given to me at a price covering development cost and reasonable profit, but not production (duplication and assembly line stuff) since there is no cost involved there. This would lower profits and thus taxes from the large car manufacturers, but the old style "hard car" profit shouldn't be afforded protection (inability for consumers to download cars) from their own inability to adapt to the technology of the day.
Laws should not protect businesses from competitors in a capitalist society, actual innovation, quality and value should. The fact that the car companies have banded together in a anti car downloading Car Industry Association proves that there is no competition between those companies in the first place. Even though it is legal to download a car from another manufacturer, the CIAA would sue to stop all car downloading, legal or not.
Actually I would hope the government would give at least one model of car away for free at that point, or I could go to the library and check out a car. I personally would buy an indie car if given the chance.
Same thing, except: no memory card, add a 1394 port.
That's it.
The only problem is the manufacturer doesn't make loads of cash on media, so it's a bit more expensive without it.
What I want to know is this: what percentage of Macintosh users actually use Photoshop, and what percentage use it the majority of the time on their computer?
/cross-platform/ applications (not the Finder) these days, and compare the performance between machines then. These applications /are/ used by a majority of PowerMac owners.
I think less than five percent even use Photoshop, being generous in my guess.
The percentage will undoubtedly be higher for PowerMac purchasers rather than iMac purchasers, but I would still say it's less than 30% of those (PowerMac) users.
Only if a majority of users use Photoshop intensively would the use of Photoshop to indicate general performance be acceptable.
What Apple should do is script Office, an email application, a web browser, and any other commonly used
Never mind the fact that a GHz machine is not the fastest you can get. Think multiple processor Intel boxes. They exist, Apple.
Apple's being extremely disingenuous in regards to performance testing, and it's just one reason I laugh at anyone (excepting Photoshop users) believing anything Apple says regarding performance just because they have a flashy demo. Run the Photoshop comparisons at Seybold or similar media conferences, but not at MacWorld, please.
You have \WINNT\Profiles\\My Documents\
I have Documents and Settings at the same level as my WINNT folder, just FYI. I guess this makes it easier to completely reinstall the operating system without losing user files.
I've never had Windows Update update any drivers, and i'm not sure it's supposed to.
Nonetheless, there should be a GLSetup-like thing for *nix, which configures your system such you don't have to go to some irc channel or obscure web page for the working driver, which makes sure your configuration is how it's supposed to be. But have it be not just for OpenGL, but for all drivers required, such as joysticks, sound cards, etc.
USB support out of the box is coming in 2.4. It takes a bit to get the backport to 2.2 working but after you do it's quite nice.
Antialiased fonts are not needed for games. Full screen anti-aliasing (implemented right now by at least nVidia) solves this problem.
A little aside: has anyone got the q3a mouse acceleration working in Linux as it does in the Windows version? Like, no acceleration?
That's not true. I got underwear from Apple during WWDC 1997 when they were testing out Mac OS 8. I didn't get socks however.
We are producing two (similar) products, and the word from the publisher is that they want both products in one box, on one cd even, with an installer that allows you to choose between them.
Supposedly the distributors are pushing for smaller DVD style packaging for software as well. I think it is that particular publisher (Microsoft) that is pushing for the larger packaging, since there definitely is not a distributor or retailer demand for such a thing.
Where has this been announced? The "from scratch" part as well as the "Cocoa" part.
By the way, Cocoa is a programing interface accessible from both Objective-C++ and Java (through a bridge), as well as other languages such as Perl or Python via third party utilities.
That's the rub: if only you didn't have to buy the hardware from Apple.
Also, by supporting them, you support closed source programming APIs. While Cocoa is by far the most ideal API to code for, it's not worth the hassle when the owner of a closed source program "deprecates" what you need the most. NeXT was notorious for dropping and changing APIs left and right, and Apple doesn't have a much better track record. Consider the very recent news about GameSprockets being only slightly supported in Mac OS X. Remember the ongoing pain unix users go through when trying to play a QuickTime movie.
The probability of Apple opening Cocoa or Carbon is relatively small, as those are intended to be their next generation APIs.
I suggest GNUstep for coders who want to use an API very similar to Cocoa. While probably a bit tougher to get into because of the lack of an integrated development environment specifically made for GNUstep, there are things in development (Project Center, Interface Builder) that will eventually make writing apps (or whatever) simpler.
Just to confirm: Carbon will be used to port from Mac OS. IE 5.1 is already ported to Mac OS X, and is included on the DP4 distribution.
>as well as many other areas which are of no interest to me
Actually, they are of interest to me, but just not on Mac OS X.
So I should have stated "which are of interest to me but not on Mac OS X because of the limited deployment platforms of this OS."
TmT
> I just hope Apple keeps the networked aspect of
> the rendering system or makes it easy to
> extend to do networking.
They have removed the networking aspect, while I believe they've stated that the option will exist for a third party to implement this functionality, as usually is the line Apple states after removing some functionality from their new OS.
If Mac OS X ever comes to Intel, I know would be interested in working on it again, particularly in the area of networked graphics, as well as many other areas which are of no interest to me because of the limited deployment platforms of this OS.
TmT