Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Apple considering Marklar?

Not sure I get your point blue. . but it will never happen regardless. . .


I'd just like Apple to be closer to the performance curve. I am an ex PC speed junkie. . it's a hard habit to kick. . . :p

I love everything else about the platform though. Give me the option to be on the bleeding edge of performance and I won't crave anything else. . . Being a recent switcher I have to say that if Apple could offer that under the hood regardless of whatever processor is driving it. . . Well let's just say that I fully believe Apple would get a LOT more of the PC user base to move over.
 
Re: apps

Originally posted by arn
Cocoa and probably Carbon apps would likely be a simple recompile to run on x86 architecture.
Click to expand...

Yeah, but what about programs that are hundreds of megabytes (or even gigabytes) that ship on 3 CDs. Do you think companies are just going to manufacutre "upgrades" and give them away to current customers for free? No way! You'll have to wait for the next version cycle for each app. And Carbon wouldn't be so easy, due to it's ties with Classic Mac OS and PowerPC

NeXTStep was multiplatform for some time before it became Mac OS X. In fact, Apple had originally promised cross-platform development capability with Rhapsody (OS X).
Click to expand...
Yes, but these were all Cocoa apps, which are very easy to port (just recompile).. But Carbon apps are a whole different thing. Carbon was created to get developers to stick with the Mac, since they DID NOT (and still DO NOT) want to rewrite their entire code over to Cocoa. It'd cost way to much and take way too much effort.

OS 9 Classic would be dead.... but it would open up the possilibty of Windows "Classic".
Click to expand...

Quark users would *LOVE* this :rolleyes:... and I don't think Microsoft would take too kindly to Apple emulating Windows on MS's home turf. They'd probably sue Apple under the DMCA for reverse engineering code to make windows OS/apps run.
 
Apple could take the route of only catering to OEMs with their OS (like Lindows is doing) so they would not have to worry about hardware compadiblity....the OEMs (Compaq, Sony, Gateway) would sell configurations that were compadible with OS X(86) and pre-install OS X(86). Driver problems are solved.
 
If I were a product manager at Apple and this Marklar Project was being taken seriously as a seperate project that was actually going to be released as an alternative to Pallidium (Longhorn is the first step towards that, reports say that the final version will require new hardware, i.e. secured hardware) I would release it before Pallidium is to be a harsh reality.

I for one would like to see Pallidium fail before it's even released.
 
Re: Re: apps

Originally posted by dricci


Yeah, but what about programs that are hundreds of megabytes (or even gigabytes) that ship on 3 CDs. Do you think companies are just going to manufacutre "upgrades" and give them away to current customers for free? No way! You'll have to wait for the next version cycle for each app. And Carbon wouldn't be so easy, due to it's ties with Classic Mac OS and PowerPC
Click to expand...

The actual code in software is usually very small. Most of the megs and megs of data is from just that - data... graphics, sound etc...

"FAT" Binaries can hold code for both platforms and share data.

arn
 
I'm unaware of what purpose Apple's Marklar development is designed to serve. However, wouldn't the soon to be tightened iron grip of M$ be enough to entice switchers in droves to OS X? Why make a whole new OS in that case? If things like Palladium (or whatever) is expected to bring love of M$ to an all time low, it would make more sense to encourage people to switch to Mac and spend $$$ on Mac hardware (Apple's "bread 'n butter).
 
Re: Re: apps

Originally posted by dricci


Yeah, but what about programs that are hundreds of megabytes (or even gigabytes) that ship on 3 CDs. Do you think companies are just going to manufacutre "upgrades" and give them away to current customers for free? No way! You'll have to wait for the next version cycle for each app. And Carbon wouldn't be so easy, due to it's ties with Classic Mac OS and PowerPC
Click to expand...

Usually executalbe code is only a few megabytes. You can't honestly believe that the code of an application would stretch over one CD let alone three. Most of the stuff is data.


Yes, but these were all Cocoa apps, which are very easy to port (just recompile).. But Carbon apps are a whole different thing. Carbon was created to get developers to stick with the Mac, since they DID NOT (and still DO NOT) want to rewrite their entire code over to Cocoa. It'd cost way to much and take way too much effort.
Click to expand...

Carbon apps are a different thing? This is a stupid myth with no foundation whatsoever. Carbon is an API written in C. Its simply a rather large set of C functions that wrap to the lower level APIs on OS X. These are (unsurprisingly) exactly the same APIs that Cocoa wraps to!

Getting an app to run from PPC to X86 is nowhere near as big a task as Carbonizing it was. The time this would take would be measured in days or weeks depending on the level of assembler --not much-- or the amount of byte reversing that needs implemented (if any). Fortunately this stuff may be directly reusable from the Windows version.
 
Marklar or OpenStep?

as a developer I hate it enough having to test on windows - we have separate test suites for windows 2000, windows 98, windows xp and windows 95. every now and then the differences between the systems shows up in my code, and it's no use explaining to the customer that it's microsoft's fault - all they know is that MY software doesn't work. and that's WITHOUT a recompile.

so all those people saying that a recompile is simple are missing the whole picture (it takes two people five days to run our test suites, and we still miss platform specific bugs - and this is a reasonably small project compared to the photoshops of this world).

however, it would be sensible for apple to release Cocoa for windows (like OpenStep). At the moment, I'm learning Cocoa in my spare time, but it's really quite useless to me. All my customers use windows, so they don't want Mac apps. But if there was Cocoa for Windows, I could write in that, and then say "but it also works on a mac". yes I still have lots of cross-platform testing to do, but it's an easy solution for me, it gets apple more software titles and most importantly doesn't threaten hardware sales - in fact it should increase them (obviously, you would want to develop in projectbuilder on your 970, or 8-way G4, and test and deploy on Cocoa/Win).

developers are key - a platform lives or dies on its software. one of the things often mentioned about OS/2 was its ability to run Win3.1 apps. So developers said to themselves, why write two versions, when we can write a Win3.1 version and hedge our bets on the platform. If they play their cards right, Apple could repeat the exact same situation. However, now is the time to do it, before .NET becomes widespread (there is a lot of indecision about it) and windows developers get used to that OO toolkit.

B.
 
Originally posted by idkew
i think that this would only be a good idea if people were truely sick and tired of microshaft and its practices.

if people are not ready to switch, the only people running marklar would be previous apple users who wanted a "faster" computer for less than an apple branded computer.

that would obviously hurt apple a ton since it would lose sales to dell...
Click to expand...

this is exactly my thinking... also what current PC owner would want to spend more on Apple h/w when they could stick with their cheapo PC? Apple would loose money and sell no Mac h/w...??? Doesn't make sense 2me.
Another duff rumour?

Steve
 
Originally posted by Sedulous
I'm unaware of what purpose Apple's Marklar development is designed to serve.
Click to expand...

I reckon Marklar is mainly there to ensure that Apple developers don't accidentally stick PPC-specific code into OSX. Unless you are INCREDIBLY vigilant, it's easy to drop platform specific code in, and, as good software engineering practice, you want to reduce your dependencies, as it gives you much greater flexibility in the future.

The fact that it could also be used as an X86 OS, and a testing ground for Cocoa/Win, as well as a big stick to threaten microsoft with, is probably an added bonus.

B.
 
Drivers...

You are all going on about drivers for X86, but you seem to forget that OS X is backed by Darwin, doesn't this mean that most drivers would be supported anyway as unix is supposed to support everything or am I thinking stupidly?
 
Jobs does not rule out Intel switch

Apple CEO Steve Jobs did not rule out switching the company's computers to chips made by Intel, rather than the Motorola and IBM Power PC processors currently employed.

Questioned by financial analysts about the possibility of making the switch, which would mean that Apple machines could match the megahertz speeds of rival PCs, Jobs replied that the company first had to complete the transition to OS X, 'Then,' he said, 'we'll have options, and we like to have options

Despite Apple's efforts to persuade computer buyers that megahertz is only one measure of a computer's performance, and not the best measure at that, Apple machines, which currently top out at 1.25GHz, are still widely perceived as slower than Intel-based PCs which exceed 2GHz.

Switching Macs to Intel or alternatively AMD processors would not necessarily mean that you would be able to run OS X on any Intel- or AMD-equipped machine. Apple could incorporate ROM chips in Macs without which OS X would not be able to run.
Click to expand...
 
Re: Marklar or OpenStep?

Originally posted by X-Baz
as a developer I hate it enough having to test on windows - we have separate test suites for windows 2000, windows 98, windows xp and windows 95. every now and then the differences between the systems shows up in my code, and it's no use explaining to the customer that it's microsoft's fault - all they know is that MY software doesn't work. and that's WITHOUT a recompile.

so all those people saying that a recompile is simple are missing the whole picture (it takes two people five days to run our test suites, and we still miss platform specific bugs - and this is a reasonably small project compared to the photoshops of this world).
Click to expand...

Its far simpler recompiling for the same OS on two different CPUs than it is compiling for different operating systems such as Win95/98 and 2K/XP. I'm not a Win32 developer but I've been told that there are a large number of API calls in 2K that do not exist in 98. However it is quite amazing how easily the open-source people move apps across platforms in so little time.

I'm currently developing a support system (with a small team) that will have client and administrator apps on Win 32, PocketPC (which I've never even used), various unix flavours (including Linux) and Mac OS X. They will also be available in web based and command line form. The server will run on a Solaris or Linux and will use mysql or oracle for its data store. Initial development should give us a working Java based prototype client and server within a few weeks. By final release all interface calls will be native. This is what I would call porting overload and is far far more than would have to be done to bring the client between the same OS.
 
Re: Drivers...

Originally posted by littlerich
You are all going on about drivers for X86, but you seem to forget that OS X is backed by Darwin, doesn't this mean that most drivers would be supported anyway as unix is supposed to support everything or am I thinking stupidly?
Click to expand...

There is no such thing as a Unix driver. There are countless different Unix versions, each with their own driver models and platform support. OSX needs mach drivers for low-level access to hardware (graphics cards, motherboards,networking and the like) and uses CUPS for printers. This means that you cannot simply use the drivers for Free BSD or Linux (which is a Unix like OS, not a Unix OS) in OSX, even if you are running on the same hardware.
 
DRM

When no one can download music or movies from the net without encountering a hassle with DRM from Palladium, they would switch to OSX. That's just about the only way.
 
I want a subnotebook w/ Marklar!

I love my iBook. But it's 2 lbs overweight!

Give me Marklar so I don't have to wait for an Apple subnotebook any longer!

I would gladly pay even inflated-MS-OS-prices to be able to run OS X on an Intel or Transmeta subnotebook.

Escher
 
Originally posted by bbyrdhouse
.... Apple could incorporate ROM chips in Macs without which OS X would not be able to run.
Click to expand...

Wrong, wrong, wrong!

Apple abandoned the ToolBox ROM because of the expense in money and time. Each new hardware release required a revision of the ToolBox ROM. That takes time and money to debug. Each new software release had to be tested on every supported version of the ToolBox ROM. That cost a lot of money and time.

OpenSTEP ran on a subset of standard Intel-based PCs. Rhapsody runs on a subset of standard Intel-based PCs. For Apple to sell a MacOS X-exclusive ROM-enabled Intel-based Macintosh would mean that the company would have to develop a brand new computer that would not run neither existing Intel-based software nor PPC- or 68k-based Mac software. This computer would not have the economies of scale enjoyed by Dell, Gateway, or HP/Compaq. As such, the computer would be more expensive than any of those. It would be more expensive even than existing PPC-based Macintoshes.

Help! Somebody, please explain this business model!
 
Re: Re: Marklar or OpenStep?

Originally posted by iJed


Its far simpler recompiling for the same OS on two different CPUs than it is compiling for different operating systems such as Win95/98 and 2K/XP.
...snip...
I'm currently developing a support system (with a small team) that will have client and administrator apps on Win 32, PocketPC (which I've never even used), various unix flavours (including Linux) and Mac OS X.
...snip...
This is what I would call porting overload and is far far more than would have to be done to bring the client between the same OS.
Click to expand...

All I am saying is that there is significantly more to do than simply hitting the recompile button - and commercial users who have paid thousands of pounds for the software will expect it to work with no glitches (unlike unsupported open source - but that's a different issue). If I was doing your project I'd scream - what are you using btw? Java and SWT sounds like it will fit the bill (but as I say, there's lots of testing to be done :))

As for the API differences between the various versions of windows, I have to say that microsoft has done a good job here (the APIs are horrible, but the way the two OSs share them works very well). The only areas I've had problems are threading and, obviously, security - even linking against the security apis on windows98 seems to bring my app down!
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back