If I could make Apple do one thing, it would be...
Work on better x86 integration.
This would involve:
1. Buying the VirtualPC team off of Connectix. This would probably involve several million dollars (no problem, Apple has cash on hand). VPC is a well developed and robust product.
2. Optimize VirtualPC to work better on the G4, and OS X (I am not aware of how well it uses multiple processors).
3. Making Motorola add big to little endian conversion to future PPC chips (and vice versa). This is pretty straightforward and is not technically difficult. This would make a good deal of the number crunching in VPC go faster.
4. Negotiating with Microsoft to obtain a decent price on WinXP software liscenses. I would guess that MS would be mostly ok with this because it is new sales for them (and upgrade sales as well). The goal would be to try to sell VirtualPC + XP for the same cost as the windows liscense alone (or as close as possible).
5. Hack the DirectX dlls to contain PPC specific code. This is the real coding nightmare. COM will let you do this (hide native PPC code inside a language and platform independent object), but the emulator would have to be able to switch between emulated x86 and real PPC code. This is a large coding effort (and an ongoing effort as new versions of DirectX come out), but the result is playable game performance. Apple would have to work with MS on this one (or try to work out what the DirectX APIs are doing which isn't impossible because they're documented).
6. Sell the Apple branded Virtual PC through the Apple Store as an accessory. It adds $200 to the cost of your Mac, but that is actually a pretty reasonable price to pay for the equivalent PC computing power you are getting.
[ 05-09-2002: Message edited by: Yevgeny ]</p>
This would involve:
1. Buying the VirtualPC team off of Connectix. This would probably involve several million dollars (no problem, Apple has cash on hand). VPC is a well developed and robust product.
2. Optimize VirtualPC to work better on the G4, and OS X (I am not aware of how well it uses multiple processors).
3. Making Motorola add big to little endian conversion to future PPC chips (and vice versa). This is pretty straightforward and is not technically difficult. This would make a good deal of the number crunching in VPC go faster.
4. Negotiating with Microsoft to obtain a decent price on WinXP software liscenses. I would guess that MS would be mostly ok with this because it is new sales for them (and upgrade sales as well). The goal would be to try to sell VirtualPC + XP for the same cost as the windows liscense alone (or as close as possible).
5. Hack the DirectX dlls to contain PPC specific code. This is the real coding nightmare. COM will let you do this (hide native PPC code inside a language and platform independent object), but the emulator would have to be able to switch between emulated x86 and real PPC code. This is a large coding effort (and an ongoing effort as new versions of DirectX come out), but the result is playable game performance. Apple would have to work with MS on this one (or try to work out what the DirectX APIs are doing which isn't impossible because they're documented).
6. Sell the Apple branded Virtual PC through the Apple Store as an accessory. It adds $200 to the cost of your Mac, but that is actually a pretty reasonable price to pay for the equivalent PC computing power you are getting.
[ 05-09-2002: Message edited by: Yevgeny ]</p>
Comments
I'd be pretty content then.
-val1s
As to the reply, valis, that would be a good idea, only if I were to recieve the lifetime supply of free new mac computers. One of each new product or update that Apple ships every year. Clearly, giving the free macs to you would not be such a good idea when they could be mine all mine.
<strong>Two equally bad posts so far. Yevgeny, why would Apple want to waste money licensing the competitions product??? Or even getting too good at emulation? Why would anybody write OSX apps if they could get onboth platforms using windows?
</strong><hr></blockquote>
These are common objections to the idea of better cross platform integration. I'll address them.
Apple would liscense money on a competitor's product because it would make for a better Mac user experience. IF a user wants to use PC software that does not exist for the Mac, THEN there is a viable alternative. Such emulation software should not be standard on all Mac's. The existence of a better emulator would make a Mac a better buy, like getting a second computer for a few hundred dollars. Apple's selling of this emulator would add legitimacy to Macs as a solution (and add $ to Apple in software sales).
Apple would compete on the basis of design, OS X and Altivec. Apple's industrial design is unquestionably the best. OS X is a good OS and certainly will get better as time goes on. G4's are slow for stanrard "for loop" programming, but are excellent for anything that can can be programmed using SIMD instructions. Each of these offer compelling advantages to various users and developers.
If the only reason developers make Apple Software is "just" to support the Mac platform, then Apple is doomed. If Apple can not offer a compelling case for developers to make software for the platform, then they are not going to get new development. Fortunately, this is not the case- there are compelling reasons to make custom Mac applications.
Matsu, I think that your reasoning is somewhat faulty. I think that this is demonstrated by looking at WINE on Linux. Why is it that Linux has WINE? Wouldn't developers want to develop for windows instead? How badly is Linux hurt by the presence of WINE? How many Linux users secretly fear that if WINE is completed that Linux developement will grind to a halt? The reason why they do not fear this is that Linux as an OS offers advantages over Win32. WINE enhances Linux- it allows you to run programs that do not exist for Linux. (yes, I know that Wine is not an emulator and I am discussing emulators, but the idea is the same).
What Apple would need to do is to avoid the situation that OS/2 found itself in: the ability to run Win16/Win32 code well, but with no ability to sell its compelling feaures (OS/2 could not sell on design or on CPU features, and only barely on the OS GUI, and it had to deal with MS's restrictive liscensing policies).
There are PC users who like the Mac platform, but who worry that they won't be able to run a particular piece of software. There are alot of Mac users who would also like to be able to run particular pieces of PC software (read: games) who do not have that ability. Apple can retain users who might drop the platform and can grab users who are lured in by the possibility of running Mac software.
Likewise, M$ is regularly circulating only half standard stuff into the internet. Not good, not Apple's fault, but still Apple's (and the DOJ's) problem. So a team to exclusively look after web/network interoperability is a must. Apple seems to be doing a good job here. They need to look after I/O standards and compatibility with Windows. They're doing a decent job, but they need to do even more, and stay on top of it at all times (cause things can change quickly)
I submitt that it's far more detrimental to the mac experience everytime a disc doesn't mount, or a PC network rejects your mac, or a web service doesn't work on your mac -- but all work on a PC for whatever reasons -- than not being able to run some app (that probably has a mac equivalent)
The emulator should be a solution of last resort, or some quick cross platform testing. Everything that can be done in the Mac environment needs to be done in the mac environment.
personal rant:
The next step in software evolution will come as computers become so powerful that write once read anywhere becomes a reality, and platforms become irrelevant. It's already possible to design programs to run in a browser that will run on any platform with a compliant browser. THIS IS why M$ wanted (and still wants) to break internet standards/protocols into something proprietary. Of course any sofware deployed in this fashion will be a lot simpler/slower than software written to live in the OS itself. But the first real candidate already exists: the office suite. A company could write an office suite with a decent feature set to run in an html browser. One version would run on any platform -- Mac, PC, Amiga, Unix, Linux, etc -- sporting a compliant browser. You can't do video editing or 3-d, but as computers advance... who knows what kind of power your iMac will have in 20 years time?
In that world, mac may be better positioned than M$. All the basic stuff will run anywhere flawlessly and instantaneously. The other stuff WILL come down to style, a logical tool design, a system that enables work and creativity with a minimum of hinderance, and pleasant hassel free interoperability with the digital substrate penetrating all media and communication, somethng you use without appreciating how easily you use it. It may not be possible to own the digital substrate, and if it is not, and M$ pushes too hard, they will fail. Maybe they're too big to ever go away, but they won't be the giant they are today.
<strong>I understand, but I still disagree. A very good emulator will only encourage people NOT to write mac specific apps. That isn't good. What I would much rather see is the increased control and regular updating of cross platform file compatibility. People have been complaining that macs are having trouble mounting data CD's that mount with no problem on their PC's. It may be the PC's fault; it may not -- Windows is notorious for turning even 'standards' into 'semi-propreitary' standards. That becomes a mac problem because regardless of what caused the glitch, your mac can't open the disc!
Likewise, M$ is regularly circulating only half standard stuff into the internet. Not good, not Apple's fault, but still Apple's (and the DOJ's) problem. So a team to exclusively look after web/network interoperability is a must. Apple seems to be doing a good job here. They need to look after I/O standards and compatibility with Windows. They're doing a decent job, but they need to do even more, and stay on top of it at all times (cause things can change quickly)
I submitt that it's far more detrimental to the mac experience everytime a disc doesn't mount, or a PC network rejects your mac, or a web service doesn't work on your mac -- but all work on a PC for whatever reasons -- than not being able to run some app (that probably has a mac equivalent)
The emulator should be a solution of last resort, or some quick cross platform testing. Everything that can be done in the Mac environment needs to be done in the mac environment.
personal rant:
The next step in software evolution will come as computers become so powerful that write once read anywhere becomes a reality, and platforms become irrelevant. It's already possible to design programs to run in a browser that will run on any platform with a compliant browser. THIS IS why M$ wanted (and still wants) to break internet standards/protocols into something proprietary. Of course any sofware deployed in this fashion will be a lot simpler/slower than software written to live in the OS itself. But the first real candidate already exists: the office suite. A company could write an office suite with a decent feature set to run in an html browser. One version would run on any platform -- Mac, PC, Amiga, Unix, Linux, etc -- sporting a compliant browser. You can't do video editing or 3-d, but as computers advance... who knows what kind of power your iMac will have in 20 years time?
In that world, mac may be better positioned than M$. All the basic stuff will run anywhere flawlessly and instantaneously. The other stuff WILL come down to style, a logical tool design, a system that enables work and creativity with a minimum of hinderance, and pleasant hassel free interoperability with the digital substrate penetrating all media and communication, somethng you use without appreciating how easily you use it. It may not be possible to own the digital substrate, and if it is not, and M$ pushes too hard, they will fail. Maybe they're too big to ever go away, but they won't be the giant they are today.</strong><hr></blockquote>
Considering that the Apple computers are slower than the PC , i see no reasons to fear that the x86 emulation will be to good
<img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
<strong>
Considering that the Apple computers are slower than the PC , i see no reasons to fear that the x86 emulation will be to good
<img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> </strong><hr></blockquote>
The purpose is not to get the best emulation, but to get acceptable emulation. Particularly for games where much of DirectX is direct hardware calls to the same video cards that are present in Macs. If the emulator runs at 50-70% the speed of a PIII at the same MHz, then this is good enough. Adding little endian to big endian conversion will help even more.
Games generally aren't written to require state of the art CPU's (GPU's maybe, CPU's, no). Most productivity software requires even less horsepower. The proposed emulation scheme would directly throw the truly intensive stuff where it is meant to go: the graphics card. VPC is rather poor at DirectX.
Obviously, you would not run anything that was CPU intensive on such a system (it can be debated if WinXP itself is CPU intensive). If you wanted to do something CPU intensive, then you should write it in native code and take advantage of Altivec. For example, only a fool would rip MP3's in a windows emulator, and only a foolish company would refuse to sell software to rip MP3's on a Mac.
For all the other more run of the mill stuff (non games), the big to little endian conversion would help quite a bit.
The point is to make your Windows experience better than what you would get for a PC at the same cost. How many Mac users would pay $200 for the ability to run Windows code reasonably quickly? If I could spend $200 to make a new Ti book operate as fast as a PIII 500 with a good video card, then Apple has just ADDED value to my machine and to the Mac OS.
<strong>
The next step in software evolution will come as computers become so powerful that write once read anywhere becomes a reality, and platforms become irrelevant. It's already possible to design programs to run in a browser that will run on any platform with a compliant browser. THIS IS why M$ wanted (and still wants) to break internet standards/protocols into something proprietary. Of course any sofware deployed in this fashion will be a lot simpler/slower than software written to live in the OS itself. But the first real candidate already exists: the office suite. A company could write an office suite with a decent feature set to run in an html browser. One version would run on any platform -- Mac, PC, Amiga, Unix, Linux, etc -- sporting a compliant browser. You can't do video editing or 3-d, but as computers advance... who knows what kind of power your iMac will have in 20 years time?
In that world, mac may be better positioned than M$. All the basic stuff will run anywhere flawlessly and instantaneously. The other stuff WILL come down to style, a logical tool design, a system that enables work and creativity with a minimum of hinderance, and pleasant hassel free interoperability with the digital substrate penetrating all media and communication, somethng you use without appreciating how easily you use it. It may not be possible to own the digital substrate, and if it is not, and M$ pushes too hard, they will fail. Maybe they're too big to ever go away, but they won't be the giant they are today.</strong><hr></blockquote>
This is a good rant to pick up on. I agree that the general trend in computing is to move to write once, run anywhere and that this is a good thing, (esp. for Mac users). Apple's strategy of supporting Java is particularly nice, and I am glad to see Apple make a move for Java developers. Kudos to Apple for making a great JVM.
The bad news is that the real world is sadly dominated by MS. It stinks to hear it, but it is true. Even worse, there is quite a bit of nice software that is written for MS only. I think that Apple could pursue a two pronged strategy of continuing to aggressively support Java and offering a MS emulator for extra cash. Is this contradictory? Maybe. It certainly is pragmatic, and it probably would sell well. Apple's current strategy works better in a world where developers are actually writing Java apps.
What are examples of LARGE cross platform Java apps? I am not aware of many (although I like to think that there are quite a few).
One problem with the argument "that computers will become ever faster" is that there exist certain application domains which need every CPU cycle they can get. Video editing is one such domain. It will never be advantageous for some software to be written in a language like Java until Java runs as quickly as C++. Future CPU designs may be able to address this, but barring this wildcard, some appliction domains will never be written in Java code. Computers will not become infinitely powerful and thus make design decisions irrelevant. There will be no 100% Java version of Quake.
So in summary, I agree with you that if everything was written in a platform independent language, then there would be no need for emulators. I look forward to the day when this is predominantly the case. Unfortunately, very little is written in such languages and so I think that there is a valid case that can be made for an emulator.
I remember 15+ years ago that there was a fair amount of finger tapping involved just waiting for a Commodore 64 word processor to perform certain functions. Our computers are magic relative to those machines, can you imagine where we will be in another 20 years!
Oh, and find a way to permanently hide any trace of Unix on X unless the user deliberately made it otherwise
So, I guess I'll be disappointed on both counts then . . .
Write their own software for mission critical apps: Web browser, Pro word processor, image editing, (they already have video editing), music editing, television/radio, games that would work well on the Mac, database, and on and on and on...
<strong>3. Making Motorola add big to little endian conversion to future PPC chips (and vice versa).</strong><hr></blockquote>
Small note: I read somewhere that current PPC processors can already switch from big to little endian (and back again
[ 05-09-2002: Message edited by: spotbug ]</p>
If anything, Apple still has the best laptops and I am one to admit this, even though I use a PC every day and have stuck to it. For me, I probably won't purchase a powermac or tower based computer running MacOS just because my PC can do everything it can and faster. Whereas, on portable side, I simply want battery life and less weight while maintaining decent performance.
That said, however, if I could make Apple do one thing in particular...
I'd have Apple devote more attention to the growing Chinese PC market. It's still early enough in the game for the Mac platform to make a big splash. If its own hardware is a tough sell, perhaps Apple could grant Mac OS licenses to a handful of clonemakers? There's no shortage of investment capital in China these days. We're talking about getting in on the ground floor of a potentially huge market - and its attendant influence in the region.
Remember when the Berlin Wall fell, and growing (formerly) East German middle classes suddenly had cars to spend their money on? Automakers fell over themselves serving those new buyers. Well, the Chinese market may turn out to be exponentially greater; imagine a Japanese level of fervor for the Mac in a country of a billion people! :eek: It could be hugely lucrative and influential.
Just an idea; long term thinking points away from areas already saturated by PCs.
[ 05-11-2002: Message edited by: Fischer ]</p>
I know the trend is for smaller and smaller computers, but I've always wanted an over-sized PowerBook.
I want a PowerBook with an 18 inch wide aspect screen, a full sized extended keyboard, dual G4s, a 64 mb graphic card, superdrive, and the usual USB, firewire and audio ports. Battery life should be at least 2-3hours. The weight should be as light as possible, but I know that it will be heavy.
Basically I want my desktop mac squashed into a 17" x 13" x 2" briefcase.
Call it the Portfolio -- the perfect computer for presentations.
I don't know if anyone else in the world would buy one, but this is my dream computer and it would be perfect my work and lifestyle.
The best way i see to do this is for a PPC decoder in the Hammer instead of the x86. I don't care if Apple's offerings are then equal to PC, as development there is good, sustained and vitually guaranteed. Hammers should be fast enough for almost everyone. The g5, may come out and be faster than the hammer at launch, but give it 6 months and the g5 will be behind again.
With guaranteed hardware parity with the PC at ALL times, Apple then has the software to work in its favour. This almost seems like a no brainer. The other option is to have no clear idea where the platform is going, no guaranteed development from moto/ibm and slowly declining sales for ever.
This is just a joke, don't take it seriously.