The End of Apple?

[Adrith]

Sorry for the sensationalist title

As I'm sure many (some?) of you have seen, Apple is switching from IBM's PowerPC architecture to an Intel x86 processor and, appearantly, have been planning the move for the last few years.

About as many people that say that this is a good move and a step forward say that it's a step in the wrong direction. I'm not really convinced either way.

What's interesting, is that they're moving to x86 now, the 32 bit architechture as opposed to an AMD64 or Itanium core. I would think that at this point, it would be a step down from their G5s. I know that Apple has been having heat and noise issues with the G5s and that this should help them make a smaller, quieter, higher power laptop, but I think that there's still plenty of room for development on the PPC platform.

Appearantly, they plan on the move taking ~2 years and it being much less bothersome than the move to the carbon framework; i'm not sure if i buy that. I know that they have an XCode compiler just about ready to go that is able to compile on both the x86 and the PPC platforms, but I still think that there's going to be more hiccoughs than Jobs forsees.

I guess I'm just curious what other people think about the whole thing. Maybe I'm just biased against Intel.


(edited for typos)
_________________
People don't tase people, I tase people.

[Jack_Ian]

It's a major pain for me.
I've been programming on the Apple since System5 when the processor was 68K.
First there was the changeover from Volume References to Working directories to cater for hierarchical disk structures. This was quite simple but was a bit of a hack.
Then there was the changeover to FSSpec, which was a pain but was worth it, since it got rid of the need for the working directory hack (well mostly, it was still used in the StandardFileDialog calls)
Then there was the change to PowerPC. This was a pain but was helped by a very efficient 68K emulator, so that the user was unaware that some of the code was actually being emulated. It involved the introduction of UPP's (Universal Procedure Pointers) wherever there was a callback routine. The worst part was rewriting the assembly code, which we used to free up computational bottlenecks. For the most part, due to the new faster processors, we were able to re-write them in PASCAL. Over time the task was completed with some 68K assembly still needed as glue code for system traps. The _Trap system remained 68K, even when running on a PowerPC processor.
Then came the Navigation Services, which was a major pain for me, as we had code that patched the StandardFile calls to replace them with a database interface that handled document workflow. so now I had to develop similar patches for the Navigation Services too. After a few pints of blood, sweat and tears the job was done.
After that, life was pretty easy for quite some time, until...
PASCAL was no longer to be the language of choice for Macintosh programming. The MPW Pascal compiler would no longer be supported and no bugs would be fixed. This was a major pain. We had vast stores of PASCAL code that was extremely well tested and just worked. I experimented a little with p2c, which could convert Pascal to C, but it could only do about 75% of the work and produced ugly, bug-prone, C code. There seemed no reason to convert to C, a process which would ultimately result in the introduction of bugs. Luckily the MPW Pascal compiler was very sturdy and produced extremely efficient executable code, although it was a pain to watch it grind away during a build. Also there was the MetroWerks Codewarrior Pascal compiler which could also compile our code and was used already by us on some projects. New projects were then created mostly in C, although they depended heavily on the existing libraries generated from the PASCAL legacy code. Now we are beginning to dig ourselves a hole and we realise it but it still seems ludicrous to invest serious time and resources into a process that will introduce bugs into our systems with no visible benefit to the customer. Life goes on. Over time we move all our projects to Codewarrior. We dig deeper.
MetroWerks announce that they will no longer support Pascal. They will bring out a final build with most outstanding bugs fixed and then no longer support further Pascal development. Well this final build CodeWarrior6 was a disaster that crashed regularly so we just reverted to our old Codewarrior4 for anything to do with Pascal since it was the last stable build.
Again life was pretty easy for a while...
Then came OSX.
At this point, this particular range of products were approaching the end of their life cycle and we seriously thought of dropping them altogether. The market for them was rapidly diminishing and the products we produced were being overtaken by our web-based and PC-based solutions. After a lot of pleading from our customers, we decided to invest the resources to get our main products to run natively on OSX.
This was a nightmare. OSX was in complete flux and the environment changed rapidly, totally unravelling the work that was already done. Eventually, the product was transferred and was very stable (although not as stable as our classic version, which was rock-solid). We got around the Pascal issue by just generating a CFM library from our old trusty CodeWarrior4 and then loading it dynamically during runtime. The transfer was extremely hard work was only worth it because enough of our customers were willing to pay for it, although they did grumble a lot about why they needed to pay so much for an upgrade that had less functionality and was likely to be less stable then the one they already had. It was hard to resist saying "I told you so", but somehow I managed it.
"At least let us have the metal-look windows", they implored.
This of course was impossible. In order to do this we would need to drop all support for our OS9 customers and write a version that will run only on OSX, since we would have to convert all Quickdraw calls to Quartz calls and re-organise our event handling throughout. Eventually we agree to change the user interface and the customers are happy that they now really do have a new product for their money.
Time to relax again. Wrong...
WHAT!!! Change the processor to Intel! Are you f@©king mad?
OK, let's see if this is doable.
Byte-Swapping issues: Should not be too difficult, since we have to interface with our PC product range already and have code already that swaps bytes around. It will just mean making sure we take the endian compiler variable into account in all cases when deciding "to swap or not to swap".
Must be compiled in XCode: This might be a problem. No Pascal compiler to compile the legacy libraries. Wait! Maybe we can use the CFM solution that worked already and just load these routines as necessary and have them run under the new Rosetta PowerPC emulator using some new version of UPP's. Nope! All or nothing. The complete process, including sub-processes, must use one binary format.
Well it appears as though we will be doomed to run under the Rosetta emulator, since it would take about two man-years to convert the Pascal legacy code. It seems this horse is finally dead.

------------------
Is it a good idea for Apple?

For the company it probably is. The processor under the hood has largely become irrelevant (unless, like me, you need to cater for legacy systems) and Apple have been itching for a reason to dump support for Classic Applications. Also, it will allow Apple to compete more fully in the digital entertainment market segment, which is where they are making the most money at present, since Intel chips have Digital Rights Management built in. I suspect also that they intend to get into the digital video market in the near future, hoping to match their current success in Audio entertainment. They probably intend to make use of Wimax for this purpose, the new high-bandwidth wireless protocol, and this is likely to be much easier if they are on an Intel platform.

It is pretty much a death-knell for their desktop products though, not to mention a kick in the teeth to those who have loyally supported them over the years. I don't expect that they will be investing much effort in this sector in future and will be content to keep a toehold in it just for prestige. I'd say windows is a little worried though. Not about their market share, but they cannot afford for Apple to leave the desktop OS market. At present, they are the only credible alternative Desktop OS for the average home user and are the only thing keeping the anti-trust and anti-monopoly hounds at bay.

[Moose*]

I don't speak computer but can you guys tell me if this will effect me at all (I am a loyal apple user...) what does this all meaannnn!?

fanks

[wordcross]

It pretty much means that you'll have to upgrade to an intel-chip computer or make due with what you've already got. Software will stop being written for your current OSX platform (or if you have something else, it most likely already has been) and you'll be stuck with a computer that will not only be obsolete physically, it will not be able to install new mac software either, since it'll be written for new computers with a different hardware setup.

So if you like what you have, you should be good to go for years yet. If you like to have the most up-to-date software, though, expect it to be necessary to upgrade to a brand-new computer.
_________________
Has anyone really been far even as decided to use even go want to do look more like?

[Moose]

that's sort of how it's been with macs anyway...no?

anyway...i'm in love with my computer. the only complaint i have is that i don't have a cd burner but otherwise I guess I'll just be using it for years to come. either that or I'll go the cheap route and buy a pc (shh...don't tell my computer that)

[Dan]

Man, I really like OSX too...oh well, at least my G4 is a couple years old. By the time this really takes effect I'll be needin' a new compy.

Jack_Ian:

Are you saying that Apple will become the Sega of computers and just make software like iTunes, Final Cut and DVD Studio Pro? Well, that and the iPod of course.

Hmmm...anyway, I think this might mean I won't be upgrading my Final Cut 3 software. :-p
_________________
Rokk. uuddlrlrbastart

[wordcross]

Haha, looking for an old thread in SAC and came across this. Oh the irony
_________________
Has anyone really been far even as decided to use even go want to do look more like?

[Jack_Ian]

Wow! That was fun to read again.

Since then, we have developed replacements for these legacy systems (completely from scratch) that fit with our new design philosophy and while many of our customers have made the leap, we still have customers happily using the old reliables and they seem to be working well still, by all accounts.

I still have an old PowerPC mac on my desk in case I need to support these customers, but lately it only gets used for emails. I must admit, I'll feel a little sentimental about finally getting rid of it when all customers finally adopt the new systems. It will close the door on an era of programming for me where I needed to regularly examine binary code and where sometimes I needed to use a soldering iron to fix my problems.

On the upside, I'm getting involved in writing apps for the iPhone. I was a little reluctant to adopt yet another language, but now that I'm using it, I must say that I'm beginning to like Objective-C. It kind of brings be back to University when I had to write some apps in SmallTalk and Ada. Back when object-orientation was a little purer.

[Zag]

Umm, Ada? Object-oriented? Isn't Ada lacking polymorphism? That would mean it is merely Object-based.

[jesternl]

If only I had bought some AAPL back then

[Jack_Ian]

Back when I qualified, object orientation was just in its infancy and wasn't as clearly defined as it is now. Ada had the ability to create generic code which, although not polymorphic, was pretty cool for its day. I guess that's why I intuitively lump it in with the OO languages, although it's probably a bit of a stretch. It certainly was a major paradigm shift from the procedural programming I was used to and parallel processing raised all sorts of interesting problems for an enthusiastic student.