Yes it is, but it's not apple's idea. ZeroConf is an open standard invented by the IETF.
Actually, it was "invented" by Stuart Cheshire, when he was working at Apple (I put invented in quotes since this wasn't just pulled out of the air in an afternoon - but he was instrumental in setting up the IETF group that hammered out the standard, and did a lot of the groundwork to get to that point).
Of course, I don't remember Microsoft throwing money into Apple in the late 90's
Yeah, you probably don't remember it - because it didn't happen. Microsoft bought $150 million of non-voting stock, which for a company with over $4 billion in the bank was meaningless. Apple gained mindshare out of the deal, not cash.
It is those executives at Apple who get to decide what the native API is, not you or I.
If "native" is just a synonym for "Cocoa", sure, call it what you like. If "native" is meant to imply there's some kind of performance or memory penalty for other APIs, I'm afraid not.
My Mac programming experience is becoming somewhat dated since I still use a Mac OS 9 system but I believe there are some features that are only available when using Cocoa. If not now this will surely become the case since Apple bills Carbon merely as something to make the transistion from 9 to X easier.
Your Mac programming experience is pretty dated - the line you're taking here is what Apple first pushed for Rhapsody, which was basically NeXT-on-PowerPC. This was a monumental flop in trying to win Mac developer mindshare, which was the reason for Carbon. The fact is that Carbon is around for the duration, and isn't a "transitional API". If you've been to any of the WWDCs in the last couple of years, you have seen there's a phenomenal amount of stuff being added to Carbon (Carbon Events, HIObject, HIView, drawers, sheets, services, etc) and to APIs that can be used from any app framework (Quartz, CFNetwork, Keychain, etc).
The message for the last couple of years has been pretty consistent: use whatever app framework you like, and if you find something that's possible in one that's not possible in the other - log a bug.
You're quite correct in that stdout/stdin-style Unix apps can be ported to Mac OS X, but Mac OS X GUI-style apps can't be ported to other platforms (since most of the APIs they use - Carbon, Cocoa, Core Audio, HID, etc, aren't available).
But I have to take issue with your "one that is considered the native API". This is complete tripe, and something that gets regurgitated by ex-NeXT executives at Apple that really should know better.
Cocoa and Carbon are both equally "native" - they both sit on top of the lower-level APIs like Core Graphics/Core Foundation, and quite happily talk to each other (e.g., the Cocoa menu system used to call over to the Carbon Menu Manager - may still be the case). With 10.2 onwards you can place Carbon windows into a Cocoa app, and vice versa.
The term "native" on the Mac means something quite specific - it was introduced for the 68K/PowerPC transition, as a way to distinguish between apps that used the native PPC ISA vs apps that were being run in the 68K emulator. The term was misappropriated when NeXT were purchased, as an attempt to paint older Mac APIs as somehow less worthy (thankfully most people at Apple have moved past this now, but it still gets trotted out once in a while).
FWIW, 99% of the games ported to the Mac (i.e., ported by anyone other than the Omni folks, who are ex-NeXT developers) are written to Carbon. Cocoa is useful for putting UIs together quickly, but really doesn't buy you anything for a game (given that almost all of the Mac-specific code will be talking to C based APIs like OpenGL, HID, or the Sound Manager).
Although these figures are just for North America, American/Canadian sports games (hockey, basketball, American football, nascar, etc) always sell badly outside of North America.
Everywhere else the "global" sports like F1 or soccer tend to do a lot better (e.g., soccer games like Championship Manager sell by the truckload everywhere but the US).
I think this goes some way to explaining the Xbox's lacklustre sales in the other two big video game territories (Europe and Asia: just behind the GC in Europe, behind just about everything in Asia). Different cultures prefer different things, and the Xbox is very much a product of the US (the PS2 did well everywhere because it was out there first, and had an installed base of existing Playstation owners to sell into).
If you have evidence that these instructions are not supported on the PPC970, please post it.
I believe they're supported, so you won't crash, but they invoke an exception handler on the 970 (as per misaligned loads/stores on the 603 or later).
This probably makes them too slow for something like VPC, and if you use them in performance critical code you would probably be better off using a vector permute on larger blocks of data. I can't find any docs at IBM or Apple to back this up unfortunately, this was based on conversations at WWDC (so may be hearsay, but would seem to explain the issue with VPC).
It is, because the G5 DOES support pseudo little-endian mode. It must be a stupid fuck-up on MS's side (as if that'd suprise anyone).
VirtualPC does not use the PowerPC's ability to boot in big or little endian - it uses the lwbrx/stwbrx instructions, which will automatically endian-swap during a load or store. This allows them to keep data in memory in little endian form, have it swapped automatically when it's brought into a register for processing, and have it swapped back when it's written out to memory.
This is the feature which isn't present on the G5, and was responsible for the big speedup in the latest rev of VPC - and the reason it now requires a G3 or G4 (since the previous PPC chips didn't support these instructions).
Since the G5 doesn't support this feature either, they'll need to go back and resurrect some of their previous code - they will doubtless take a performance hit for having to do the swapping themselves, but the massive bandwidth in the new systems will probably help cancel some of that out.
How do you know that Apple isn't putting ECC memory into these boxes? Is the motherboard chipset incapable of handling ECC memory?
The current G5 motherboards do not support EEC memory. I asked about support for EEC at WWDC, and the reply was that they were definitely thinking about that and similar high-reliability features.
I'm quite sure that things like EEC and G5 Xserves are on the drawing boards, but realistically they can't do everything at once - and for Apple, the first rev of these machines is a replacement for their existing desktop models (big sales like this are impressive for PR, but don't sustain a company the size of Apple).
Some tests have already proven that the G5 is not overwhelmingly superior to the G4 when using Altivec code (just a linear increase with the clock rates)
I would find that quite surprising - each G5 CPU has two Altivec units within it (i.e., four total in a 2xG5 system).
Obviously you're unlikely to get perfect parallelism out of them, but I doubt a second unit would have been added if it was of no benefit in typical code - it sounds more likely that the test you quote was a particularly poorly performing example.
Didn't Voyager and Galileo take advantage of the solar wind to get way out there in a short time?
No (human) spacecraft to date has used the solar wind for propulsion - the solar sail is the only realistic mechanism for doing so, and that's never actually been tried (there was to have been a test of the Cosmos 1 couple of years ago but it suffered a launch failure).
F1 is like soccer - it's absolutely massive everywhere but the US, and the US equivalents are pretty much ignored everywhere else.
The US may/may not be the single biggest market in the world, but realistically it's just one of the top three territories for videogames (North America, Europe, and Asia). Games that sell well in the US (e.g., Madden) can easily bomb everywhere else (e.g., Madden) - games that sell well everywhere need to hit at least 2 of the 3 big markets.
Actually they changed the name of it as part of a product relaunch - SoftWindows hadn't been updated for while, and was not really regarded as offering great performance.
Connectix, who by then had built up a name for themselves as very proficient low-level developers with hacks like RamDoubler, brought out VirtualPC with a lot of PR about how much faster it was than SoftWindows.
RealPC was a direct response to that, if a little lame in the naming department...
You mean "lock" - a loch is a lake or a body of water like a fjord. I believe the plan is that ships can go directly up to the various towns en route, not up to the dam.
Of course, he did have something of a point - the three machines were code-named Piltdown Man, Cold Fusion, and Carl Sagan.
He first wrote an annoyed sounding letter to MacWEEK (who had revealed the code names) complaining about being associated with two well-known scientific frauds, which caused the whole BHA renaming thing in retaliation, which sparked off the lawsuit.
You can't fault the judge for dismissing his case though - "One does not seriously attack the expertise of a scientist using the undefined phrase 'butt-head.' "...:-)
AltiVec is important for Apple marketing because it lets them claim impressive performance figures without actually needing to push the state of the art in terms of processor design further than Intel.
No, AltiVec is important for Apple full stop - in the short term to make up for the anemic bus speeds allowed by the G4, and in the longer term because a SIMD unit is now as expected a component of modern desktop CPUs as an FPU is.
And even something like a hand-coded vectorized BLAS library doesn't help because most scientific software still doesn't use such libraries
The only thing you can really sure about "most" scientific software is that it needs an FPU. Scientists and engineers do a huge variety of simulations, some of which are vectorizable and some of which aren't.
If AltiVec has a weakness in the scientific field, it's the lack of support for double precision. And there's nothing in the instruction set which precludes this, so I wouldn't be surprised to see it appear in some future CPU.
Imagine how much better it would be if Apple could ship systems based on the 970 today, rather than after a few months additional delay due to AltiVec.
If it didn't have AltiVec, it wouldn't be what Apple needs in a desktop CPU - not much point in getting what you don't need a few months early (not like that would happen anyway: this isn't lego: you can't unplug "the AltiVec bits" without any impact on the rest of the design).
And every dollar and watt that is shaved off the AltiVec price makes it a much more viable processor for servers and blades, which would get volume up and prices down.
Except that Apple aren't currently in the blade market at all, and have a fairly small presence in the more general server market. If they can sell a few boxes there, fine, but getting the volume up means targetting consumers - not server farms.
The shared file feature only works for INTRANET LAN's using Rendezvous.
Not quite - the discovery of servers on the local net is done with Rendezvous, but you can "share" (i.e., stream) music between any two IP addresses (if you're behind a firewall, you need to open port 3689).
so if i made a internet radio and streamed stuff to my friends i wouldnt need to worry about anything, and i could play any song i wanted without retribution?
iTunes does limit the number of clients that can connect (to 5 I believe), so I imagine this has already been taken into consideration. I find it hard to imagine that streaming between two arbitrary IP addresses (rather than just the local subnet) would get into a product as significant as iTunes now is for Apple if they weren't 100% sure that the record industry was OK with it.
I suspect the client limit was specifically to turn this from a "internet radio" situation into "play your CDs to a couple of your friends, just as if they'd come over to your house".
It took about 5 minutes after iTunes was released to people start sharing stuff across the net, and about 10 minutes before people were writing apps/php scripts to list active servers, so this has to have been cleared beforehand.
No western philosophy discusses in too great of detail whether this world is real or not.
Er, not quite - that was exactly what Descartes (rationalism) and Berkeley/Lock (empiricism) were all about (as well as nutbars like Heidegger and Kant to a degree).
This was probably the most significant development in "western" (i.e., european) philosophy at the time, and arguably a lot more profound than Nietzsche's semi-sociological leanings (which are more about his obsession with control than anything else - "if you're going to see a woman, remember to take a whip!", etc).
If the Matrix takes anything from philosophy, it's the Cartesian method - how do you know that what you perceive is real, and not just put there to trick you? Descarte's answer was that you can't, and the only thing you can ever know for sure is that you, the individual doing the questioning, exists.
Neo takes it for granted that the reality shown to him by Morpheus is really real, but of course he has no more way of testing that than he did in the original simulation. Ultimately he takes the route of Empiricism, and choses to act in a way that seems appropriate to what he perceives - which of course is exactly what he was doing before.
Slashdot Mirror
Yes it is, but it's not apple's idea. ZeroConf is an open standard invented by the IETF.
Actually, it was "invented" by Stuart Cheshire, when he was working at Apple (I put invented in quotes since this wasn't just pulled out of the air in an afternoon - but he was instrumental in setting up the IETF group that hammered out the standard, and did a lot of the groundwork to get to that point).
"minimum": the least possible quantity or degree.
"Humor": That which is intended to induce laughter or amusement.
Of course, I don't remember Microsoft throwing money into Apple in the late 90's
Yeah, you probably don't remember it - because it didn't happen. Microsoft bought $150 million of non-voting stock, which for a company with over $4 billion in the bank was meaningless. Apple gained mindshare out of the deal, not cash.
It is those executives at Apple who get to decide what the native API is, not you or I.
If "native" is just a synonym for "Cocoa", sure, call it what you like. If "native" is meant to imply there's some kind of performance or memory penalty for other APIs, I'm afraid not.
My Mac programming experience is becoming somewhat dated since I still use a Mac OS 9 system but I believe there are some features that are only available when using Cocoa. If not now this will surely become the case since Apple bills Carbon merely as something to make the transistion from 9 to X easier.
Your Mac programming experience is pretty dated - the line you're taking here is what Apple first pushed for Rhapsody, which was basically NeXT-on-PowerPC. This was a monumental flop in trying to win Mac developer mindshare, which was the reason for Carbon. The fact is that Carbon is around for the duration, and isn't a "transitional API". If you've been to any of the WWDCs in the last couple of years, you have seen there's a phenomenal amount of stuff being added to Carbon (Carbon Events, HIObject, HIView, drawers, sheets, services, etc) and to APIs that can be used from any app framework (Quartz, CFNetwork, Keychain, etc).
The message for the last couple of years has been pretty consistent: use whatever app framework you like, and if you find something that's possible in one that's not possible in the other - log a bug.
Thankfully it looks like T17 felt the same way - this is much more like a traditional Worms games than the bizarre spin-offs.
You're quite correct in that stdout/stdin-style Unix apps can be ported to Mac OS X, but Mac OS X GUI-style apps can't be ported to other platforms (since most of the APIs they use - Carbon, Cocoa, Core Audio, HID, etc, aren't available).
But I have to take issue with your "one that is considered the native API". This is complete tripe, and something that gets regurgitated by ex-NeXT executives at Apple that really should know better.
Cocoa and Carbon are both equally "native" - they both sit on top of the lower-level APIs like Core Graphics/Core Foundation, and quite happily talk to each other (e.g., the Cocoa menu system used to call over to the Carbon Menu Manager - may still be the case). With 10.2 onwards you can place Carbon windows into a Cocoa app, and vice versa.
The term "native" on the Mac means something quite specific - it was introduced for the 68K/PowerPC transition, as a way to distinguish between apps that used the native PPC ISA vs apps that were being run in the 68K emulator. The term was misappropriated when NeXT were purchased, as an attempt to paint older Mac APIs as somehow less worthy (thankfully most people at Apple have moved past this now, but it still gets trotted out once in a while).
FWIW, 99% of the games ported to the Mac (i.e., ported by anyone other than the Omni folks, who are ex-NeXT developers) are written to Carbon. Cocoa is useful for putting UIs together quickly, but really doesn't buy you anything for a game (given that almost all of the Mac-specific code will be talking to C based APIs like OpenGL, HID, or the Sound Manager).
Although these figures are just for North America, American/Canadian sports games (hockey, basketball, American football, nascar, etc) always sell badly outside of North America.
Everywhere else the "global" sports like F1 or soccer tend to do a lot better (e.g., soccer games like Championship Manager sell by the truckload everywhere but the US).
I think this goes some way to explaining the Xbox's lacklustre sales in the other two big video game territories (Europe and Asia: just behind the GC in Europe, behind just about everything in Asia). Different cultures prefer different things, and the Xbox is very much a product of the US (the PS2 did well everywhere because it was out there first, and had an installed base of existing Playstation owners to sell into).
You may also have seen it here a fortnight ago. Check it out - it's a site with loads of geek stories like this. Oh, wait...
If you have evidence that these instructions are not supported on the PPC970, please post it.
I believe they're supported, so you won't crash, but they invoke an exception handler on the 970 (as per misaligned loads/stores on the 603 or later).
This probably makes them too slow for something like VPC, and if you use them in performance critical code you would probably be better off using a vector permute on larger blocks of data. I can't find any docs at IBM or Apple to back this up unfortunately, this was based on conversations at WWDC (so may be hearsay, but would seem to explain the issue with VPC).
It is, because the G5 DOES support pseudo little-endian mode. It must be a stupid fuck-up on MS's side (as if that'd suprise anyone).
VirtualPC does not use the PowerPC's ability to boot in big or little endian - it uses the lwbrx/stwbrx instructions, which will automatically endian-swap during a load or store. This allows them to keep data in memory in little endian form, have it swapped automatically when it's brought into a register for processing, and have it swapped back when it's written out to memory.
This is the feature which isn't present on the G5, and was responsible for the big speedup in the latest rev of VPC - and the reason it now requires a G3 or G4 (since the previous PPC chips didn't support these instructions).
Since the G5 doesn't support this feature either, they'll need to go back and resurrect some of their previous code - they will doubtless take a performance hit for having to do the swapping themselves, but the massive bandwidth in the new systems will probably help cancel some of that out.
Or ECC even, too much time in Europe...
How do you know that Apple isn't putting ECC memory into these boxes? Is the motherboard chipset incapable of handling ECC memory?
The current G5 motherboards do not support EEC memory. I asked about support for EEC at WWDC, and the reply was that they were definitely thinking about that and similar high-reliability features.
I'm quite sure that things like EEC and G5 Xserves are on the drawing boards, but realistically they can't do everything at once - and for Apple, the first rev of these machines is a replacement for their existing desktop models (big sales like this are impressive for PR, but don't sustain a company the size of Apple).
A somewhat more informative link for more info. Would it really kill submitters to put a link to the actual project in their submission...
Some tests have already proven that the G5 is not overwhelmingly superior to the G4 when using Altivec code (just a linear increase with the clock rates)
I would find that quite surprising - each G5 CPU has two Altivec units within it (i.e., four total in a 2xG5 system).
Obviously you're unlikely to get perfect parallelism out of them, but I doubt a second unit would have been added if it was of no benefit in typical code - it sounds more likely that the test you quote was a particularly poorly performing example.
Didn't Voyager and Galileo take advantage of the solar wind to get way out there in a short time?
No (human) spacecraft to date has used the solar wind for propulsion - the solar sail is the only realistic mechanism for doing so, and that's never actually been tried (there was to have been a test of the Cosmos 1 couple of years ago but it suffered a launch failure).
F1 is like soccer - it's absolutely massive everywhere but the US, and the US equivalents are pretty much ignored everywhere else.
The US may/may not be the single biggest market in the world, but realistically it's just one of the top three territories for videogames (North America, Europe, and Asia). Games that sell well in the US (e.g., Madden) can easily bomb everywhere else (e.g., Madden) - games that sell well everywhere need to hit at least 2 of the 3 big markets.
Check your local TV Guide - that should have them.
Actually they changed the name of it as part of a product relaunch - SoftWindows hadn't been updated for while, and was not really regarded as offering great performance.
Connectix, who by then had built up a name for themselves as very proficient low-level developers with hacks like RamDoubler, brought out VirtualPC with a lot of PR about how much faster it was than SoftWindows.
RealPC was a direct response to that, if a little lame in the naming department...
Do they have some sort of loch
You mean "lock" - a loch is a lake or a body of water like a fjord. I believe the plan is that ships can go directly up to the various towns en route, not up to the dam.
Of course, he did have something of a point - the three machines were code-named Piltdown Man, Cold Fusion, and Carl Sagan.
:-)
He first wrote an annoyed sounding letter to MacWEEK (who had revealed the code names) complaining about being associated with two well-known scientific frauds, which caused the whole BHA renaming thing in retaliation, which sparked off the lawsuit.
You can't fault the judge for dismissing his case though - "One does not seriously attack the expertise of a scientist using the undefined phrase 'butt-head.' "...
AltiVec is important for Apple marketing because it lets them claim impressive performance figures without actually needing to push the state of the art in terms of processor design further than Intel.
No, AltiVec is important for Apple full stop - in the short term to make up for the anemic bus speeds allowed by the G4, and in the longer term because a SIMD unit is now as expected a component of modern desktop CPUs as an FPU is.
And even something like a hand-coded vectorized BLAS library doesn't help because most scientific software still doesn't use such libraries
The only thing you can really sure about "most" scientific software is that it needs an FPU. Scientists and engineers do a huge variety of simulations, some of which are vectorizable and some of which aren't.
If AltiVec has a weakness in the scientific field, it's the lack of support for double precision. And there's nothing in the instruction set which precludes this, so I wouldn't be surprised to see it appear in some future CPU.
Imagine how much better it would be if Apple could ship systems based on the 970 today, rather than after a few months additional delay due to AltiVec.
If it didn't have AltiVec, it wouldn't be what Apple needs in a desktop CPU - not much point in getting what you don't need a few months early (not like that would happen anyway: this isn't lego: you can't unplug "the AltiVec bits" without any impact on the rest of the design).
And every dollar and watt that is shaved off the AltiVec price makes it a much more viable processor for servers and blades, which would get volume up and prices down.
Except that Apple aren't currently in the blade market at all, and have a fairly small presence in the more general server market. If they can sell a few boxes there, fine, but getting the volume up means targetting consumers - not server farms.
No, the International Federation of the Pronographic Industry - what do you think he is, some kind of weirdo?
The shared file feature only works for INTRANET LAN's using Rendezvous.
Not quite - the discovery of servers on the local net is done with Rendezvous, but you can "share" (i.e., stream) music between any two IP addresses (if you're behind a firewall, you need to open port 3689).
so if i made a internet radio and streamed stuff to my friends i wouldnt need to worry about anything, and i could play any song i wanted without retribution?
iTunes does limit the number of clients that can connect (to 5 I believe), so I imagine this has already been taken into consideration. I find it hard to imagine that streaming between two arbitrary IP addresses (rather than just the local subnet) would get into a product as significant as iTunes now is for Apple if they weren't 100% sure that the record industry was OK with it.
I suspect the client limit was specifically to turn this from a "internet radio" situation into "play your CDs to a couple of your friends, just as if they'd come over to your house".
It took about 5 minutes after iTunes was released to people start sharing stuff across the net, and about 10 minutes before people were writing apps/php scripts to list active servers, so this has to have been cleared beforehand.
No western philosophy discusses in too great of detail whether this world is real or not.
Er, not quite - that was exactly what Descartes (rationalism) and Berkeley/Lock (empiricism) were all about (as well as nutbars like Heidegger and Kant to a degree).
This was probably the most significant development in "western" (i.e., european) philosophy at the time, and arguably a lot more profound than Nietzsche's semi-sociological leanings (which are more about his obsession with control than anything else - "if you're going to see a woman, remember to take a whip!", etc).
If the Matrix takes anything from philosophy, it's the Cartesian method - how do you know that what you perceive is real, and not just put there to trick you? Descarte's answer was that you can't, and the only thing you can ever know for sure is that you, the individual doing the questioning, exists.
Neo takes it for granted that the reality shown to him by Morpheus is really real, but of course he has no more way of testing that than he did in the original simulation. Ultimately he takes the route of Empiricism, and choses to act in a way that seems appropriate to what he perceives - which of course is exactly what he was doing before.