As I recall I had no issue during PowerPC and Intel Mac transitions. It has been over a year since M1 Mac came out, why there are still lots of software, including those from well established big companies, that have M1 version still underdevelopment with no timeframe on availability? Didn’t Apple provide them hardware long enough for development?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
macOS Why the M1 transition takes so long softwarewise?
- Thread starter hajime
- Start date
- Sort by reaction score
No business need to do so.
Apple user base is comparatively tiny, they’re not going to sell the software again so there’s it’s all cost to do so, and Rosetta 2 is pretty damn good.
I may be wrong but it looks like the increased number of people buying Mac the past few years are students and household consumers who make videos to post on youtube. So they are satisfied with using M1 Mac with Final Cut and basic productivity software and M1 Mac continue to be in big demand.
MacOS accounts for just 15% of the desktop market share in 2021. Apple don’t release the sales figures but not all of that 15% share will be running apple silicone.
Arguably only a small percentage, considering the huge number of Intel macs out there.
Arguably only a small percentage, considering the huge number of Intel macs out there.
It took a long time for some PPC software to move to Intel as well, and of course some never did.
For lots of software it might just be a recompile so, if it's under active development, an ARM version will come out fairly quickly - but for others it can be more complex.
* It might use third party libraries for example, and there are restrictions on mixing code, so the app has to wait on someone else who may have different priorities.
* IIRC Autodesk said they had to move their OpenGL code to Metal as part of the transition. I'm not sure why as I thought that, while deprecated, OpenGL was still available, but I assume they know what they are doing. The more general point though is if you were relying on a library or facility that's not even available anymore then of course rewriting your code to use something else can be a lot of work.
* If you have an old code base you might also find there are lots of hardware assumptions built in that need sorting, in parts of the code no one has looked at in years. I imagine unaligned accesses might be problematic on the M1 but you can usually get away with them on Intel.
These things are all difficulties, and then, as others have said, the economics needs to make sense.
Unless its something that will boot sales. There is no reason to rush to M1.
**** most companies still use Objective C for their code base and have yet upgraded to Swift.
Because the development cost is not worth the rewrite.
With Companies it's all dollars and cents.
So it is more about money than they are obligated to make an updated version optimized for M1 as soon as possible because they have already been selling a version for the Mac for years? So to these companies, it is like Windows version, Linux version, Intel Mac version, and the smaller market as well as less profitable M1 Mac version that does not motivate them to devote more human resource to make the update?
Last edited:
Of course no one is _obligated_ to do the port (excluding the unlikely possibility of some contractual issues that some companies may have with large customers). For a company its always cost/benefit (even if they don't always calculate correctly). The point I tried to make above is the cost might be higher than you think. The point others have made is the benefit might not be as high as you think.
Without wishing to end up in a politics thread you will find that most problems in the world come down to the money as well - though of course it still needs to be spent in the right way.
Speaking for myself, there are a lot of third-party libraries that need to be ported before the software I'm working on could be built for M1. Those third-party libraries are not available as Xcode projects, but use a hodgepodge of different build systems and compilers, not all of which can compile for ARM out of the box.
The existing software runs fine under Rosetta, so there's not a compelling need to move to M1. In addition to all this, the new software would need to be extensively tested to ensure that no bugs were introduced during the transition.
If it were as easy as checking a checkbox and rebuilding, it would have been done by now. In the real world, not every piece of software is built using Apple's toolchain.
Rosetta 2 provides a nice baseline and hence many developers are taking longer to update apps to M1 which means slower adoption of the new M1. Also, as others have stated, install base is so small that many developers are focusing on the bigger Intel install base for the moment.