macOS Carbon vs. Cocoa: A structured debate

[GothicChess.Com]

Oh man… this is such an uncool thing to do!!!
b e n

Please be advised that my response was purely to counter the overly-sarcastic remark here:

If you could write something like this and were not working for an IB, then you are without a doubt the smartest person on the planet.

While his was latent with sarcasm, mine was presented without bias and was purely factual.

If I wanted to do something overly-outrageous, I would have cited the fact that I was the last human being to defeat a Computer World Champion in the games of chess (Deep Thought, 1989) and Checkers (Chinook, 1996) which you can see...

here:
http://www.chessgames.com/perl/chessgame?gid=1272214

... and here...
http://www.cs.ualberta.ca/~chinook/WallofHonor.php

.. though you might have to scroll down a ways on the checkers page.

Look man, I did not initiate any post that would have these people from the sidelines make marginal impugning remarks about my character.

I just wanted to know how definitive the statements were regarding Carbon being destined for the Apple Graveyard, and I asked for facts, not merely conjecture.

That having been said, I would appreciate the rest of the posts to be about this.

Thank you.

The Management

 

[garethlewis2]

Okay, you think that coding a chess program is easier than writing a simulation with partial diffs, fine. You have your opinion, I have mine.

For Carbon. Carbon was and is viewed by Apple as a transitory platform for Adobe and Microsoft. Nobody else. Apple don't need Adobe now. Adobe screwed up when they stated their video software won't be maintained on the Mac any more. Apple released FCP. With Aperture, Apple simply showed Adobe they had 1 year to release Photoshop to the Apple users. Adobe have complied, but with a 32bit version, which has pissed off so many users, they are looking for something better. Adobe don't have the position they used to enjoy. The instant the management at Adobe fart the wrong way, Apple will release their Cocoa version of a Photoshop killer.

I really can't see the problem with moving from Carbon to Cocoa. Yes, it is a different language. And no, it isn't difficult. I will reiterate my earlier, point about Wilco who are a subsidary of ADP and Calypso which is a company in SF. Wilco used to the software house to work for when it came to back-office processing, not front-office processing. All the major IBs used Glosspack and GlossHV. In 1998 I looked at what Wilco was doing and realised the management did not want to move to C++ on Unix, or Visual C++, or Java. They had some moron in who told them that moving would cost valuable time and money. Time in development and money in lost revenue. As all the developers would need to be retrained. They chose to listen to the consultant and not us the developers who told them, unless they move, the customers will look for something newer that is easier to use and maintain. Here is the main problem with Gloss. It only runs on Unix, and it only works through a JAM or Jyac Application Manager interface, which is a tty screen software package built on ncurses. 1998 and their software only has a text interface. Windows NT 4 is out. Other companies already have a team beavering away building a NT version of their software. Wilco has two choices. Move to a knew platform or buyout another company that already has a package. They chose the latter and bought an Indian company called DEShaw. 6 months later and god knows how many green cards, the indian develeopers had disappeared and let a Java system so rubbish and buggy, the develeopers at Wilco felt betrayed by management. 2001. Calypso a small startup in SF convert a C++ trading system that runs on Unix and Windows to Java, so it could run anywhere. They get their first big clients. Wilco are still trying to build a GUI version of their software and are using PowerBuilder. 2004. Wilco doesn't appear anymore in any of the financial papers. Paribas, ING, RBS have moved all trading off Gloss and onto Calypso. Gloss handled most products but could never process interest rate derivatives or any type of derivatives product. 2007. The last I heard, Wilco was a small part of ADP offering a brokerage service running on a Unix mainframe that allowed IBs and smaller hedge fund managers to trade without a IT unit. Calypso is the world leader in financial sofware.

What was the problem that Wilco had? They could not stomach having to recode the entire application in another language. It was just too much of a jump for them.

Carbon will be replaced, Apple don't maintain backwards compatability. Classic is dead. You can't 68k programs on an intel mac for OS 8 or 9. Carbon will most likely follow the same path. Adapt or Die.

 

[Krevnik]

Carbon will be replaced, Apple don't maintain backwards compatability. Classic is dead. You can't 68k programs on an intel mac for OS 8 or 9. Carbon will most likely follow the same path. Adapt or Die.

Carbon as we knew it as an OS 9 + OS X Library is already dead. APIs have already replaced QuickDraw and a bunch of other transitional technologies that Carbon dragged into OS X. A lot of what is available now is a lot cleaner, newer, and OS X-only. Why would Apple spend so much time adding OS X-only functionality (some of which are C-only APIs), if they just intended to whack it all? MLTE just got nuked in favor of a new C API in Leopard, BTW.

 

[kainjow]

For Carbon. Carbon was and is viewed by Apple as a transitory platform for Adobe and Microsoft. Nobody else. Apple don't need Adobe now. Adobe screwed up when they stated their video software won't be maintained on the Mac any more. Apple released FCP. With Aperture, Apple simply showed Adobe they had 1 year to release Photoshop to the Apple users. Adobe have complied, but with a 32bit version, which has pissed off so many users, they are looking for something better. Adobe don't have the position they used to enjoy. The instant the management at Adobe fart the wrong way, Apple will release their Cocoa version of a Photoshop killer.

Is this speculation, or do you know this as fact? It kind of makes sense, but a Cocoa Photoshop killer would be insane.

 

[GothicChess.Com]

Slight diversion...

Anyone here remember the SuperPaint program? (which was far superior to the free MacPaint that came with the original Mac). Is there a free OS X Paint-style program out there now?

I have a seen a few foreign language paint programs, and I have been half tempted to write my own OS X Superpaint application. I keep an OS 9 box running just for SuperPaint, it is sooooooo easy to use for quick-n-dirty graphics.

 

[MongoTheGeek]

Slight diversion...

Anyone here remember the SuperPaint program? (which was far superior to the free MacPaint that came with the original Mac). Is there a free OS X Paint-style program out there now?

I have a seen a few foreign language paint programs, and I have been half tempted to write my own OS X Superpaint application. I keep an OS 9 box running just for SuperPaint, it is sooooooo easy to use for quick-n-dirty graphics.

Have you looked at the gimp? http://www.apple.com/downloads/macosx/unix_open_source/gimpapp.html

 

[garethlewis2]

No, it's not speculation. It is well known in the Apple development community that Adobe realise Apple could release a Cocoa Photoshop replacement. Adobe tried to shake apple with their empty bluff of not supporting their equivalent of FCP on the Mac. Apple replied by releasing FCP. And you have to ask the question. Why would a Cocoa Photoshop style application be insane. It would be 64-bit. Photoshop is 32-bit, and as I mentioned previously, professional studios are pissed at this. They have machines with 16 gigs of Ram, and they are stuck with an app that can only access 2 gigs of Ram in a single data window.

As a side-note. Carbon was already updated to 64-bit. It is on the Apple development forums if you have access. Though of course, 64-bit Carbon is a step backwards. No GC, no support for NSOperation, etc. So it was killed.

Why would they kill a well written API? It is 2007 not 1997. 1997 C API's for application writing were the norm. Windows developers scoffed at this new API, MFC containing the newest features, they knew it would be backported into Win32 APIs. Didn't happen, and never will. Now MFC is under threat from CLR.

Coding practices change.

 

[Eraserhead]

No, it's not speculation. It is well known in the Apple development community that Adobe realise Apple could release a Cocoa Photoshop replacement. Adobe tried to shake apple with their empty bluff of not supporting their equivalent of FCP on the Mac. Apple replied by releasing FCP.

FWIW FCP came out before Adobe pulled Premiere .

For a free paint program their is Seashore

 

[garethlewis2]

And all the developers who worked on Premiere worked on FCP.

 

[Fukui]

Carbon as we knew it as an OS 9 + OS X Library is already dead. APIs have already replaced QuickDraw and a bunch of other transitional technologies that Carbon dragged into OS X. A lot of what is available now is a lot cleaner, newer, and OS X-only. Why would Apple spend so much time adding OS X-only functionality (some of which are C-only APIs), if they just intended to whack it all? MLTE just got nuked in favor of a new C API in Leopard, BTW.

Yes, but these new C-APIs are not carbon, they are CoreFoundation based. CoreFoundation != carbon.

 

[aleister55]

Apple forcing cocoa on all of you holdouts is like John Landen saying that he'll only fight if someone put's a particular kind of bottled water in his locker room.

There are no valid excuses left

 

[GothicChess.Com]

Apple forcing cocoa on all of you holdouts is like John Landen saying that he'll only fight if someone put's a particular kind of bottled water in his locker room.

There are no valid excuses left

Ummm, what?

 

[Soulstorm]

I began my OS X development by learning Carbon on OS X. Then, I decided to move to Cocoa, because everybody was telling me to do so. I think in order to see better the pros and cons of both APIs

Don't get me wrong here. I love Cocoa. I think the most complete and powerful API the world has ever seen. Class abstraction and manipulation, 'id' generic objects, and the way all this is integrated with Interface Builder continues to amaze me, even after so much time devoted to Cocoa development.

But I think that Carbon should stay. I have seen very little of Objective-C 2.0, but I don't think that it will ever be sufficient to replace another language such as C++. C++ has tons of libraries, and for that reason a developer should not be forced to throw away all those libraries, and to learn a totally new development method. I know, there is Objective C++, but that's not enough, as this is just a desperate move to support C++ with Cocoa.

Developers that develop for a specific platform need to have options. I hope Carbon continues to evolve. The way I see it, Carbon and Cocoa are here to stay. Carbon will evolve dramatically over the next versions of OS X, and will abandon many deprecated technologies and programming concepts that are now used, just as it did with Quickdraw. And, let's not forget one basic thing:

OpenGL, OpenAL, and libraries such as these in general, are not written for Objective C. I find it easier to write a program in OpenGL Carbon using classes, than developing an OpenGL application in Cocoa. Performance in OpenGL Carbon should be better, too, classes with OpenGL objects can be made cleaner and clearer.

The difference between Objective C and C++ is that Objective C is an open world programming language, and C++ is a closed world one. An open world programming language is just like designing the interior of a car. The driver seats, the wheel, etc. But C++ will be used to design the engine compartment of the car. It is more efficient, and provides better low-level management.

Concluding... I think that the world still needs Carbon. Just not in the form that exists in this time. I believe it should be upgraded, enhanced with better programming concepts, and re-designed to compete with Cocoa in areas where Cocoa lacks on functionality. I won't point out what those areas are, I believe many people in here have found some parts of Cocoa that need enhancements, or areas where C++ can perform better than Objective C.

 

[GothicChess.Com]

Concluding... I think that the world still needs Carbon. Just not in the form that exists in this time. I believe it should be upgraded, enhanced with better programming concepts, and re-designed to compete with Cocoa in areas where Cocoa lacks on functionality. I won't point out what those areas are, I believe many people in here have found some parts of Cocoa that need enhancements, or areas where C++ can perform better than Objective C.

I pray that you are correct! But it seems many who have posted here have whipped the Carbon-ites into submission.

I have spawned another Topic Group, https://forums.macrumors.com/threads/322246/ = Cocoa Topics for Carbon Ex-Patriots -- maybe you will find this of some interest.

 

[lazydog]

I know, there is Objective C++, but that's not enough, as this is just a desperate move to support C++ with Cocoa.

Hi

This is just my opinion but I can't see why Objective C++ isn't enough. What Apple has given us is a fantastic amount of flexibility in the way you can program the Mac. C++ classes, STL, OpenGL, Cocoa classes, standard Unix C, Carbon etc can all coexist happily in the same project. I would agree with you if Cocoa bound you to a 'closed box' environment… but it doesn't.

In the other thread about Cocoa, garethlewis2 mentioned that the dot method call for classes is being introduced. Along with garbage collection and integration with C++ I really think Objective-C is going to be an amazing language to program in.

b e n

 

[alFR]

<snip>
The Wright Brothers were bicycle mechanics, yet they invented the field of Avaiation.

Albert Einstein was a Swiss patent clerk, yet he published "Relativity" and demonstrated that the Newtonian Physics metaphor was lacking a great deal.
<snip>

Hmm, you've written some software and you're comparing yourself to Einstein and the Wright brothers. Hubris, thy name is GothicChess.Com....

 

[Eraserhead]

Hmm, you've written some software and you're comparing yourself to Einstein and the Wright brothers. Hubris, thy name is GothicChess.Com....

I don't think he really is, he's just using them as an example of people who weren't working in the field who became experts.

In the history of computer chess, it is well known that Ken Thompson's $600,000 (1982 prices) Belle program was the first ever artificial intelligence player to make Master in United States Chess Federation play.

What is not well known, is that when I was still 20, my own Macintosh chess program, The Sniper, was the first software program to break the 2200 Master barrier, just 4 years after Thompson. (Belle was a hardware program, a dedicated machine, funded by Bell Labs.)

So Gothic hasn't just written "some software".

 

[Fukui]

I have seen very little of Objective-C 2.0, but I don't think that it will ever be sufficient to replace another language such as C++. C++ has tons of libraries, and for that reason a developer should not be forced to throw away all those libraries, and to learn a totally new development method. I know, there is Objective C++, but that's not enough, as this is just a desperate move to support C++ with Cocoa..

Please don't take this the wrong way, but thats just wrong. Nobody has to throw away their C++ templates and libraries to use Cocoa. And if you haven't seen Objective-C 2.0, then well....

 

[GothicChess.Com]

Hmm, you've written some software and you're comparing yourself to Einstein and the Wright brothers. Hubris, thy name is GothicChess.Com....

You missed the point entirely.

Einstein was not well liked by some of his college teachers, and he was basically forced to take a job as a patent clerk to make ends meet. He couldn't get a job in his preferred field of work.

On the merit of his credentials, Albert Einstein was not highly recommended.

In 1902, the Wright Brothers were virtually unknown, unless you had a flat tire on your bicycle in the town of Kitty Hawke.

On the merit of their credentials, the Wright Brothers had no qualifications as aeronautical engineers.

That having been said, some pissant was posting on here that because I was not a slave worshipping the god of Cocoa, that I basically had no idea what I was doing as a programmer.

I merely replied, quite factually, that I had already something of merit before the age of 20. It's not relativity theory, and it won't get very far off the ground, but considering it ran on a 16 megahertz processor with no hard drive, and it fit on an 800 K floppy disk with the operating system also installed, my chess program broke the master barrier far before technological advances enabled another software program to follow suite. It took the Software Toolworks company 5 more years, on a machine that was over 6 times as fast, to finally overshadow it. I wrote it in my spare time while dual majoring. I think they had more manpower, more resources, and more time to throw at it than I did.

 

[FabL]

Core UI?

Someone claimed that Cocoa in Leopard will be founded on a new "Core UI" framework rather than HIToolbox (or whatever it uses today). If this is true, and the CoreUI framework is available to developers, wouldn't that be the solution to everyone's problems? Cross-platform developers could use pure C code and still be up-to-date with the latest UI stuff.

I've done some cross-platform development, and honestly: Wrapping C around Objective-C makes me want to wash my hands afterwards. An up-to-date C framework for UIs would be so very welcome.

 

[GothicChess.Com]

I've done some cross-platform development, and honestly: Wrapping C around Objective-C makes me want to wash my hands afterwards.

You and me both