I agree with you completely that $129 US for OS X should not be a problem. Yet, why do people run Linux rather than OS X? I think there are two reasons. First, OS X only runs on the Macintosh. Most buyers are attracted by lower prices and greater selection of generic, or open, hardware platforms. Second, hardware vendors can all offer Linux Unix, whereas only Apple can build and sell an OS X platform.
HP, IBM and Dell are pushing Linux and I think that that's the main reason for the growth in the Linux market. But what are HP, IBM and Dell pushing? Linux on their hardware and with their support options - not much different than Apple and Mac OS X.
Quote:
Originally posted by snoopy
As market conditions appear today, Linux market share will likely continue to grow. It may be a safe bet, therefore, that a good GUI will be available several years from now.
Unix is over 30 years now and every Unix maker has their own interface and their own version of Unix - why would this be different on Linux? One of the problems with OSS is that people are doing what they want and are essentially running in different directions. We have 10+ window managers, a couple of desktop environments and a gazillion themes.
Quote:
Originally posted by snoopy
When this happens, such a GUI could become the de facto standard for software development.
If it happens.
Quote:
Originally posted by snoopy
Mac OS X may easily run Linux binaries, but what happens when most applications are written for a unique Linux API and graphic subsystem?
Uhm, since Linux is open source, no one is stopping no one from porting that API to Mac OS X. Today you can run the previous mentioned window managers, desktop environments and themes on Mac OS X with little or no problems.
Quote:
Originally posted by snoopy
Based on limited software knowledge, my guess is that OS X applications will have to be ported from Linux, much like they are ported from Windows today.
Since Fink ports are free and done on a voluntary basis, I don't think that porting an X11 app from Linux to Mac OS X is that hard.
Apple should release Cocoa for Linux and other Unixes. I really think that developers would like to have a great API and toolbox for doing great apps, apps that through a mere recompile will look and work great in OSX and vice versa.
This is an interesting proposal, especially if one says that Apple could release Cocoa APIs for *nixes and not for Windows.
. . . Apple should release Cocoa for Linux and other Unixes. I really think that developers would like to have a great API and toolbox for doing great apps, apps that through a mere recompile will look and work great in OSX and vice versa. . .
I'm trying to understand the key differences between this idea of yours, and my original proposal, which Amorph shot down. I proposed licensing Cocoa API and Quartz, thinking that OS X applications might run on any PPC OS that licensed these features. The goal was to have truly common software applications, which was too ambitious apparently.
Now, if I understand what you propose, Cocoa would be open source so anyone could use it. Your goal is to make it easy to port applications from any OS using Cocoa to OS X, yes? I can see where this goal, being less ambitious, might be easy to achieve. My concern is that Apple may be giving away too much by making Cocoa open source, and that easy porting might be accomplished in other ways. I'm sure Cocoa would be a big boost to Linux, but would bring with it fewer reasons for someone to choose OS X. And I thought I was going out on a limb by suggesting that Apple license Cocoa and Quartz. At least through a license, Apple could limit and select those it grants permission to use their technologies. I had in mind IBM, who might integrate Cocoa and Quartz into their private version of Linux.
Developers do like to have a great API and toolbox for doing great applications. However, the idea should be to get those developers working on the Mac, not Linux. It could happen if Xcode makes it easier to port applications to a popular open source Unix variant. If open source developers are working on a Mac with Xcode, OS X gets an application too. Whether a developer chooses to market the OS X version is another matter. I think most would.
Thank you for your detailed analysis. Possibly I may conclude the following:
You haven't found anything technically wrong with Apple promoting an open source distribution of Darwin BSD that includes a reasonably Apple GUI, better than what Linux has today. Your conclusion appears to be that such an undertaking by Apple will not bring any major gains to OS X. Essentially, I may be trying to solve problems that don't exist, and may never come to pass.
I can accept that as a real possibility. I'm not ready to concede that this is true, but it will make me think about it a lot more.
The question, snoopy, is: how do you find the proper place of such a Darwin between "an OS X for poor people and second-class citizens" (as perceived) and "an OS X that has everyone the 'real' OS X does, without the price tag and the marketing stuff"?
Such a question never occurred to me. An open source distribution of Darwin BSD would be seen as just another open source Unix, a little like Linux. I don't think it would be considered a poor people's OS X any more than X86 Linux is considered a poor people's Windows. The GUI that Apple would make for open source PPC Darwin need not look anything like Aqua. It's likely Apple would distribute it through another company, either as a partner or subsidiary.
Now, if I understand what you propose, Cocoa would be open source so anyone could use it. Your goal is to make it easy to port applications from any OS using Cocoa to OS X, yes?
That was my idea yes. I didn't mean porting of Quarz though so a ported Cocoa would have to be adapted to whatever screen drawing technique the other OSs might use. OSX will still have an edge in that department and even if Quartz is nice it would be quite a feat to introduce a entirely new screen drawing technique to another platform, and even if it only was meant to worl for Cocoa apps, such an app would look and behave very differently from other apps on that platform so I don't think that would be such a great idea.
If Apple went out on a limb to make a Quartz/X11-merged WM that'd retain all of the coolness of X11 and Quartz, that'd be another matter entirely,. Such a thing would see the ligt of day on OSX first, and that'd be years away if ever.
Quote:
Originally posted by snoopy
However, the idea should be to get those developers working on the Mac, not Linux. It could happen if Xcode makes it easier to port applications to a popular open source Unix variant.
Why use OSX and Xcode to do general UN*X-development if Cocoa wasn't available on those other platforms? I can't really imagine people using Xcode for doing X11 based apps in GTK, Qt or something like that, perhaps Java though. Porting Cocoa to Linux, AIX, Solaris (any other major UN*X alive an kicking?) wouldn't nessecairily include Xcode and Interface Builder on those plattforms so OSX would be the preferred development platform for Cocoa apps. But If Cocoa would succeed porting of Xcode and IF would be a big help.
If Apple went out on a limb to make a Quartz/X11-merged WM that'd retain all of the coolness of X11 and Quartz, that'd be another matter entirely,. Such a thing would see the ligt of day on OSX first, and that'd be years away if ever.
Apple could start by becoming a member of X.org (if they aren't already). Then they could actively help forming the future of X11 and perhaps shoehorn Quartz into X11 to make a unified modern WM for *nixes.
But I still se the various desktop environments as a problem, and I don't think that Apple wants a unified DE* since Mac OS X then wouldn't act differently from Red Hat with KDE installed.
*) By a unified DE I don't mean a unified look but a set of UI guidelines upon which Apple can put Aqua, others can choose Gnome, KDE or something else.
On the other hand, a unified DE would mean that customers can choose what they want in the land of *nixes without any fear of not being able to run an app and with a minimum of relearning your habits as every *nix functions the same way.
That would make a standard OS in an open market just like we have standards with regards to TVs, CDs, DVDs, AM/FM and so on.
I would love to see the day where you buy computers as you buy TVs today. All work in basically the same way and have the same basic features, but you choose IBM, HP, Dell, Apple, Sun or SGI based on what else they offer that fits your needs.
*nix makers should start to talk to each other and look at that big competitor from Redmond which is close to achieve what I discribe above - but with one OS supplier.
. . . Why use OSX and Xcode to do general UN*X-development if Cocoa wasn't available on those other platforms? . . .
Oops, sorry. My mind jumped back to another post, and I got off track. With such good feedback, I'm rethinking parts of my proposal. Right now, however, I want to point out some of my concerns about making Cocoa and Xcode open source.
Open source systems will always cost less than equivalent Macs, and have a larger selection of hardware. The only way Apple can stay in business and have a reasonable market share is to offer better and more desirable systems than open source. If Linux eventually improves in ways that make it very much like OS X, there will be few if any reasons to buy a Mac. Generic PPC hardware running Linux might become tomorrow's overwhelming standard. I think many people would like that, but it would put an end to much of the innovation we see today. Do we want a standard OS, which is just refined gradually as time goes by, or do we want to keep new things coming? The best way I know to stoke up the furnace of innovation is let it be profitable.
Would making Cocoa and Xcode open source be giving away too much? That is a critical question. If Cocoa gives Linux a big boost toward a better user interface, than I believe it is too much. If Xcode significantly improves code development for Linux, then OS X loses an advantage and it is also too much to give away.
I believe Apple does need a strategy to deal with the growth of open source systems, which will soon be coming in force to the PPC world if IBM has their way. Apple needs a way to happily coexist and not be damaged by this increasing trend. Making it easier to port applications between open source and OS X would help, but it should not be at the expense of taking away the Mac's advantages.
...."Yes but the software's Windows® ... why get a Mac?".
Interesting thread Snoopy.
no truer words ever spoken: of course there are zillions of greatapps for mac but picture the senario: you have adobe photoshop, illistrator, after effects, M$ office, and tons of games for windows, you realy want mac but it just isnt practical, the price is right but who wants to shell out 2000$ or more to re-buy what you have already purchased, if people could "trade horizontaly" (i.e. trade a windows license and media of photoshop 7 for a mac license and media of the same) if this were instituted and advertised, ms would go mental as they loose installed base so rapidly.
Don't forget that Apple was an early adopter of Linux. They had MkLinux (Linux ported to the Mach kernel) up and going on Apple hardware before OS X. They ended up dropping it when they purchased NeXT, but not after releasing all the hardware enabling code that allowed other Linux distros to run on Macs - LinuxPPC and Yellow Dog and company owe their existences to the MkLinux effort.
Note also that Apple is adding a lot of Linux compatibility to their core OS. After all, you don't have to adopt the whole codebase, you only need to adopt enough to make your codebase look like Linux to applications.
A few other comments on the thread:
There's a lot more involved in "porting Cocoa" then there was in the NeXTStep days. For one thing, Cocoa used to be almost completely self-contained when it was on NeXTStep/OpenStep. Now, much of its capabilities have been moved to Foundation, the very low-level system libraries that also service QuickTime, Java, Carbon, and much of the operating system itself. You could port a Cocoa lookalike, sort of like GNUStep, but it wouldn't be a transparent way to run OS X apps. It would be an 80% solution at best, and it would involve a lot of work hauling all that code back out of Foundation. There would still be huge gaps like the lack of HFS+, VFS, QuickTime, etc. Not gonna happen. If you want to write UNIX-agnostic apps, Java, Mozilla/XUL, Python/Tk, or some other language with TrollTech's Qt (not QuickTime!) commercial framework will all do. They won't create anything that looks as purely OS X as, say, OmniWeb is, but that's because OS X is light-years ahead of other UNIX interfaces.
As to the quantity of interfaces on Linux, I really don't think this is a problem at all. (Their quality, on the other hand, is!) Linux could plausibly take over the market for a horizontal base operating system - I'd much rather see it there than Windows, although I'd prefer one of the BSDs there instead. This is already starting to happen: People are starting with Linux, porting it to some hardware, and building a custom interface for whatever their particular piece of hardware requires. This is an immensely valuable thing, given that computers are becoming both specialized and ubiquitous. The common baseline makes it that much easier for them all to communicate, and it spares the developers a great deal of effort writing system code.
In fact, if you were so inclined, you could think of the Mac itself as one such project - on a huge scale, but nevertheless. Rather than reinvent wheels that have long since been invented, just take the work that has been done, adapt it to your purposes, and do the work where it really makes a difference - talking to the hardware, and talking to the user.
You could port a Cocoa lookalike, sort of like GNUStep, but it wouldn't be a transparent way to run OS X apps. It would be an 80% solution at best, and it would involve a lot of work hauling all that code back out of Foundation. There would still be huge gaps like the lack of HFS+, VFS, QuickTime, etc. Not gonna happen. If you want to write UNIX-agnostic apps, Java, Mozilla/XUL, Python/Tk, or some other language with TrollTech's Qt (not QuickTime!) commercial framework will all do. They won't create anything that looks as purely OS X as, say, OmniWeb is, but that's because OS X is light-years ahead of other UNIX interfaces.
That's my point!
If all UN*X development is made in Java, Mozilla/XUL, Python/Tk, or Qt then those apps won't be adopted by OSX users because the look and behave like crap. Certainly good enough for the Linux or even Windows crowd but not to us, so all that development will be of no or litte use to us. If one eould make a decent OSX app out of lets say OpenOffice, then someone have to do some serious rewriting and adjustments just to make it look good in OSX.
If Apple released Cocoa into the wild that might change and porting to a really nice looking OSX app would be a breeze. The lack of HFS+, VFS, QuickTime, etc in non OSX Cocoa wouldn't be a serious issue since that stuff isn't a part of Java, Mozilla/XUL, Python/Tk, or Qt either.
If Apple released Cocoa into the wild that might change and porting to a really nice looking OSX app would be a breeze. The lack of HFS+, VFS, QuickTime, etc in non OSX Cocoa wouldn't be a serious issue since that stuff isn't a part of Java, Mozilla/XUL, Python/Tk, or Qt either.
And my point is that that's the reason you can't write a really good OS X app in any of those environments.
Porting Cocoa would be a tremendous amount of effort that would a) dilute the appeal of OS X, and b) fail to reproduce the appeal of OS X. You'd end up with a lot of superficially native-looking, but lacking, applications.
Basically, operating systems are significant. If you want to do transparently cross-platform work, you have two choices: Code to the lowest common denominator, or bring your own OS (this is basically Java's approach). Neither one will produce a native-looking OS X app, and porting Cocoa to other platforms would not change this basic fact. Cocoa is basically a (currently incomplete) means to access the underlying functionality of OS X. It's not a standalone environment.
Comments
Originally posted by snoopy
I agree with you completely that $129 US for OS X should not be a problem. Yet, why do people run Linux rather than OS X? I think there are two reasons. First, OS X only runs on the Macintosh. Most buyers are attracted by lower prices and greater selection of generic, or open, hardware platforms. Second, hardware vendors can all offer Linux Unix, whereas only Apple can build and sell an OS X platform.
HP, IBM and Dell are pushing Linux and I think that that's the main reason for the growth in the Linux market. But what are HP, IBM and Dell pushing? Linux on their hardware and with their support options - not much different than Apple and Mac OS X.
Originally posted by snoopy
As market conditions appear today, Linux market share will likely continue to grow. It may be a safe bet, therefore, that a good GUI will be available several years from now.
Unix is over 30 years now and every Unix maker has their own interface and their own version of Unix - why would this be different on Linux? One of the problems with OSS is that people are doing what they want and are essentially running in different directions. We have 10+ window managers, a couple of desktop environments and a gazillion themes.
Originally posted by snoopy
When this happens, such a GUI could become the de facto standard for software development.
If it happens.
Originally posted by snoopy
Mac OS X may easily run Linux binaries, but what happens when most applications are written for a unique Linux API and graphic subsystem?
Uhm, since Linux is open source, no one is stopping no one from porting that API to Mac OS X. Today you can run the previous mentioned window managers, desktop environments and themes on Mac OS X with little or no problems.
Originally posted by snoopy
Based on limited software knowledge, my guess is that OS X applications will have to be ported from Linux, much like they are ported from Windows today.
Since Fink ports are free and done on a voluntary basis, I don't think that porting an X11 app from Linux to Mac OS X is that hard.
Originally posted by Henriok
Apple should release Cocoa for Linux and other Unixes. I really think that developers would like to have a great API and toolbox for doing great apps, apps that through a mere recompile will look and work great in OSX and vice versa.
This is an interesting proposal, especially if one says that Apple could release Cocoa APIs for *nixes and not for Windows.
Originally posted by Henriok
. . . Apple should release Cocoa for Linux and other Unixes. I really think that developers would like to have a great API and toolbox for doing great apps, apps that through a mere recompile will look and work great in OSX and vice versa. . .
I'm trying to understand the key differences between this idea of yours, and my original proposal, which Amorph shot down. I proposed licensing Cocoa API and Quartz, thinking that OS X applications might run on any PPC OS that licensed these features. The goal was to have truly common software applications, which was too ambitious apparently.
Now, if I understand what you propose, Cocoa would be open source so anyone could use it. Your goal is to make it easy to port applications from any OS using Cocoa to OS X, yes? I can see where this goal, being less ambitious, might be easy to achieve. My concern is that Apple may be giving away too much by making Cocoa open source, and that easy porting might be accomplished in other ways. I'm sure Cocoa would be a big boost to Linux, but would bring with it fewer reasons for someone to choose OS X. And I thought I was going out on a limb by suggesting that Apple license Cocoa and Quartz. At least through a license, Apple could limit and select those it grants permission to use their technologies. I had in mind IBM, who might integrate Cocoa and Quartz into their private version of Linux.
Developers do like to have a great API and toolbox for doing great applications. However, the idea should be to get those developers working on the Mac, not Linux. It could happen if Xcode makes it easier to port applications to a popular open source Unix variant. If open source developers are working on a Mac with Xcode, OS X gets an application too. Whether a developer chooses to market the OS X version is another matter. I think most would.
Thank you for your detailed analysis. Possibly I may conclude the following:
You haven't found anything technically wrong with Apple promoting an open source distribution of Darwin BSD that includes a reasonably Apple GUI, better than what Linux has today. Your conclusion appears to be that such an undertaking by Apple will not bring any major gains to OS X. Essentially, I may be trying to solve problems that don't exist, and may never come to pass.
I can accept that as a real possibility. I'm not ready to concede that this is true, but it will make me think about it a lot more.
Originally posted by Chucker
The question, snoopy, is: how do you find the proper place of such a Darwin between "an OS X for poor people and second-class citizens" (as perceived) and "an OS X that has everyone the 'real' OS X does, without the price tag and the marketing stuff"?
Such a question never occurred to me. An open source distribution of Darwin BSD would be seen as just another open source Unix, a little like Linux. I don't think it would be considered a poor people's OS X any more than X86 Linux is considered a poor people's Windows. The GUI that Apple would make for open source PPC Darwin need not look anything like Aqua. It's likely Apple would distribute it through another company, either as a partner or subsidiary.
Originally posted by snoopy
Now, if I understand what you propose, Cocoa would be open source so anyone could use it. Your goal is to make it easy to port applications from any OS using Cocoa to OS X, yes?
That was my idea yes. I didn't mean porting of Quarz though so a ported Cocoa would have to be adapted to whatever screen drawing technique the other OSs might use. OSX will still have an edge in that department and even if Quartz is nice it would be quite a feat to introduce a entirely new screen drawing technique to another platform, and even if it only was meant to worl for Cocoa apps, such an app would look and behave very differently from other apps on that platform so I don't think that would be such a great idea.
If Apple went out on a limb to make a Quartz/X11-merged WM that'd retain all of the coolness of X11 and Quartz, that'd be another matter entirely,. Such a thing would see the ligt of day on OSX first, and that'd be years away if ever.
Originally posted by snoopy
However, the idea should be to get those developers working on the Mac, not Linux. It could happen if Xcode makes it easier to port applications to a popular open source Unix variant.
Why use OSX and Xcode to do general UN*X-development if Cocoa wasn't available on those other platforms? I can't really imagine people using Xcode for doing X11 based apps in GTK, Qt or something like that, perhaps Java though. Porting Cocoa to Linux, AIX, Solaris (any other major UN*X alive an kicking?) wouldn't nessecairily include Xcode and Interface Builder on those plattforms so OSX would be the preferred development platform for Cocoa apps. But If Cocoa would succeed porting of Xcode and IF would be a big help.
Originally posted by Henriok
If Apple went out on a limb to make a Quartz/X11-merged WM that'd retain all of the coolness of X11 and Quartz, that'd be another matter entirely,. Such a thing would see the ligt of day on OSX first, and that'd be years away if ever.
Apple could start by becoming a member of X.org (if they aren't already). Then they could actively help forming the future of X11 and perhaps shoehorn Quartz into X11 to make a unified modern WM for *nixes.
But I still se the various desktop environments as a problem, and I don't think that Apple wants a unified DE* since Mac OS X then wouldn't act differently from Red Hat with KDE installed.
*) By a unified DE I don't mean a unified look but a set of UI guidelines upon which Apple can put Aqua, others can choose Gnome, KDE or something else.
On the other hand, a unified DE would mean that customers can choose what they want in the land of *nixes without any fear of not being able to run an app and with a minimum of relearning your habits as every *nix functions the same way.
That would make a standard OS in an open market just like we have standards with regards to TVs, CDs, DVDs, AM/FM and so on.
I would love to see the day where you buy computers as you buy TVs today. All work in basically the same way and have the same basic features, but you choose IBM, HP, Dell, Apple, Sun or SGI based on what else they offer that fits your needs.
*nix makers should start to talk to each other and look at that big competitor from Redmond which is close to achieve what I discribe above - but with one OS supplier.
Originally posted by Henriok
. . . Why use OSX and Xcode to do general UN*X-development if Cocoa wasn't available on those other platforms? . . .
Oops, sorry. My mind jumped back to another post, and I got off track. With such good feedback, I'm rethinking parts of my proposal. Right now, however, I want to point out some of my concerns about making Cocoa and Xcode open source.
Open source systems will always cost less than equivalent Macs, and have a larger selection of hardware. The only way Apple can stay in business and have a reasonable market share is to offer better and more desirable systems than open source. If Linux eventually improves in ways that make it very much like OS X, there will be few if any reasons to buy a Mac. Generic PPC hardware running Linux might become tomorrow's overwhelming standard. I think many people would like that, but it would put an end to much of the innovation we see today. Do we want a standard OS, which is just refined gradually as time goes by, or do we want to keep new things coming? The best way I know to stoke up the furnace of innovation is let it be profitable.
Would making Cocoa and Xcode open source be giving away too much? That is a critical question. If Cocoa gives Linux a big boost toward a better user interface, than I believe it is too much. If Xcode significantly improves code development for Linux, then OS X loses an advantage and it is also too much to give away.
I believe Apple does need a strategy to deal with the growth of open source systems, which will soon be coming in force to the PPC world if IBM has their way. Apple needs a way to happily coexist and not be damaged by this increasing trend. Making it easier to port applications between open source and OS X would help, but it should not be at the expense of taking away the Mac's advantages.
Originally posted by Harald
...."Yes but the software's Windows® ... why get a Mac?".
Interesting thread Snoopy.
no truer words ever spoken: of course there are zillions of greatapps for mac but picture the senario: you have adobe photoshop, illistrator, after effects, M$ office, and tons of games for windows, you realy want mac but it just isnt practical, the price is right but who wants to shell out 2000$ or more to re-buy what you have already purchased, if people could "trade horizontaly" (i.e. trade a windows license and media of photoshop 7 for a mac license and media of the same) if this were instituted and advertised, ms would go mental as they loose installed base so rapidly.
Note also that Apple is adding a lot of Linux compatibility to their core OS. After all, you don't have to adopt the whole codebase, you only need to adopt enough to make your codebase look like Linux to applications.
A few other comments on the thread:
There's a lot more involved in "porting Cocoa" then there was in the NeXTStep days. For one thing, Cocoa used to be almost completely self-contained when it was on NeXTStep/OpenStep. Now, much of its capabilities have been moved to Foundation, the very low-level system libraries that also service QuickTime, Java, Carbon, and much of the operating system itself. You could port a Cocoa lookalike, sort of like GNUStep, but it wouldn't be a transparent way to run OS X apps. It would be an 80% solution at best, and it would involve a lot of work hauling all that code back out of Foundation. There would still be huge gaps like the lack of HFS+, VFS, QuickTime, etc. Not gonna happen. If you want to write UNIX-agnostic apps, Java, Mozilla/XUL, Python/Tk, or some other language with TrollTech's Qt (not QuickTime!) commercial framework will all do. They won't create anything that looks as purely OS X as, say, OmniWeb is, but that's because OS X is light-years ahead of other UNIX interfaces.
As to the quantity of interfaces on Linux, I really don't think this is a problem at all. (Their quality, on the other hand, is!) Linux could plausibly take over the market for a horizontal base operating system - I'd much rather see it there than Windows, although I'd prefer one of the BSDs there instead. This is already starting to happen: People are starting with Linux, porting it to some hardware, and building a custom interface for whatever their particular piece of hardware requires. This is an immensely valuable thing, given that computers are becoming both specialized and ubiquitous. The common baseline makes it that much easier for them all to communicate, and it spares the developers a great deal of effort writing system code.
In fact, if you were so inclined, you could think of the Mac itself as one such project - on a huge scale, but nevertheless. Rather than reinvent wheels that have long since been invented, just take the work that has been done, adapt it to your purposes, and do the work where it really makes a difference - talking to the hardware, and talking to the user.
Originally posted by Amorph
You could port a Cocoa lookalike, sort of like GNUStep, but it wouldn't be a transparent way to run OS X apps. It would be an 80% solution at best, and it would involve a lot of work hauling all that code back out of Foundation. There would still be huge gaps like the lack of HFS+, VFS, QuickTime, etc. Not gonna happen. If you want to write UNIX-agnostic apps, Java, Mozilla/XUL, Python/Tk, or some other language with TrollTech's Qt (not QuickTime!) commercial framework will all do. They won't create anything that looks as purely OS X as, say, OmniWeb is, but that's because OS X is light-years ahead of other UNIX interfaces.
That's my point!
If all UN*X development is made in Java, Mozilla/XUL, Python/Tk, or Qt then those apps won't be adopted by OSX users because the look and behave like crap. Certainly good enough for the Linux or even Windows crowd but not to us, so all that development will be of no or litte use to us. If one eould make a decent OSX app out of lets say OpenOffice, then someone have to do some serious rewriting and adjustments just to make it look good in OSX.
If Apple released Cocoa into the wild that might change and porting to a really nice looking OSX app would be a breeze. The lack of HFS+, VFS, QuickTime, etc in non OSX Cocoa wouldn't be a serious issue since that stuff isn't a part of Java, Mozilla/XUL, Python/Tk, or Qt either.
Originally posted by Henriok
If Apple released Cocoa into the wild that might change and porting to a really nice looking OSX app would be a breeze. The lack of HFS+, VFS, QuickTime, etc in non OSX Cocoa wouldn't be a serious issue since that stuff isn't a part of Java, Mozilla/XUL, Python/Tk, or Qt either.
And my point is that that's the reason you can't write a really good OS X app in any of those environments.
Porting Cocoa would be a tremendous amount of effort that would a) dilute the appeal of OS X, and b) fail to reproduce the appeal of OS X. You'd end up with a lot of superficially native-looking, but lacking, applications.
Basically, operating systems are significant. If you want to do transparently cross-platform work, you have two choices: Code to the lowest common denominator, or bring your own OS (this is basically Java's approach). Neither one will produce a native-looking OS X app, and porting Cocoa to other platforms would not change this basic fact. Cocoa is basically a (currently incomplete) means to access the underlying functionality of OS X. It's not a standalone environment.