Talk:Mac transition to Apple silicon

A Commons file used on this page or its Wikidata item has been nominated for speedy deletion

The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for speedy deletion:

• Apple apple-silicon 3rd-party 06222020.jpg

You can see the reason for deletion at the file description page linked above. —Community Tech bot (talk) 16:23, 23 June 2020 (UTC)

Improve the article's title by distinguishing the Mac product line from Apple as a whole

This article was originally created with the title Apple's transition to ARM processors, being modeled on the title of the existing article at Apple's transition to Intel processors. As many are aware, however, Apple Inc. is not at all new to ARM-based processors, having used them for over ten years in its iPhone and iPad product lines. Therefore it's not entirely accurate to say that "Apple" (as a company) is now "transitioning" to the ARM architecture; rather, it's the company's Macintosh product line that's now making that transition. (The same issue has, in fact, been raised and discussed on several previous occasions in regard to the "transition to Intel" article, as can be seen on its talk page.)
⋮
If this article is going to encompass the full history of Apple's adoption of ARM chips, going all the way from to circa-2006 up to the present, then the current title is fine. On the other hand, if this article is going to focus more narrowly on the transition from Intel x86 chips to ARM64 "Apple Silicon" of the Mac hardware line, and/or the macOS operating system (let's remember that the two, though related, are not the same), then it would probably be appropriate to refine the article's title accordingly. Do you agree?
⋮
If so, then choosing the best alternative would seem to be largely a matter of weighing clarity versus brevity. Options that come to mind fit any of the following patterns:

• ?{Apple|Apple's} {macOS|Macintosh|Mac} ?{platform} {transition to|migration to|adoption of} ARM ?{processors|architecture}
• ?{Apple|Apple's} {macOS|Macintosh|Mac}-on-ARM {transition|migration}
• {Transition|Migration} of ?{Apple|Apple's} {macOS|Macintosh|Mac} ?{platform} to {ARM|"Apple Silicon"} ?{processors|architecture}

(I'm hoping you can interpret my improvised grep-like syntax; basically, "{foo|bar|bazz}" indicates alternatives, and "?{xyz}" indicates optional omission.)
⋮
It's possible that, over time, there will come to be some vernacular term that is most commonly used to refer to the transition we're talking about here, and if that happens, the term would probably make for a good article title (per WP:COMMONNAME). Until then, it's up to us to decide, based on editorial judgment. What are your thoughts, fellow editors? ~ Thanks, — Jaydiem (talk) 06:08, 24 June 2020 (UTC)

Yeah. See also Apple Newton lol. Anyway, it'd be "Mac transition to ARM". "Apple's Mac" and "ARM processors" are each redundant, and both is even redundantly redundantlier. The older article's sloppy title should be changed to "Mac transition to Intel". — Smuckola(talk) 07:07, 24 June 2020 (UTC)

Yes, Apple's not undergoing a corporate transition to ARM, the Macintosh line is undergoing such a transition. "Mac transition to ARM" sounds OK to me. Guy Harris (talk) 07:43, 24 June 2020 (UTC)

I agree with these sentiments. I was similarly uncomfortable with the article title's suggestion that the entirety of Apple Inc.'s hardware products were being transitioned to ARM, when in fact what's ongoing exclusively refers to the Macintosh product line, and I'm glad this imprecision is being discussed. I'd be comfortable with most of the more precise titles suggested above -- maybe "Mac transition to {ARM, Apple silicon}", and also relatedly, "Mac transition to Intel x86 architecture." On a side note, as has been suggested, Apple first utilized ARM processors in the Newton line of computers (PDA's), which was introduced in 1993. Biblib (talk) 14:50, 24 June 2020 (UTC)

Agree with the above sentiments. Current title is misleading. --Resplendent (talk) 15:31, 24 June 2020 (UTC)

Clarity vs. brevity

I see that Gaelan has boldly changed the article title to "Mac transition to ARM". That takes care of the specificity to Mac rather than Apple, but it strikes me as excessively terse. My understanding of Wikipedia's general guidelines for article titles is that it's important to consider how easily a general-interest reader will be able to understand a title's meaning at first glance. I'm concerned that the words "Mac" and "ARM", without further context, are a bit too glib for a good article title. I would prefer adding a little more context, such as clarifying "Mac" by using one or more of the words "Apple", "Macintosh", and "platform", and clarifying "ARM" as "[the] ARM architecture", or perhaps using "Apple Silicon" or "Apple ARM processors" instead.
⋮
Having thought about it, I'd like to suggest an alternative title that's almost as brief as "Mac transition to ARM", but which reads better, and is (I believe) easier for a general reader to interpret: Apple's Mac-on-ARM transition. A similar wording could be used to improve the title of the current "Apple's transition to Intel processors" article, to Apple's Mac-on-Intel transition. This is the best combination of clarity and brevity that I’ve been able to come up with.
⋮
I suppose I could boldly just make these changes myself, but I'd prefer to "read the room" first. Thoughts? ~ Thanks, — Jaydiem (talk) 15:51, 25 June 2020 (UTC)

I guess we could turn the title into an essay, or maybe find a way to attach one below it. If only there was a way to explain a title! This makes no sense. WP:SLOP — Smuckola(talk) 17:01, 25 June 2020 (UTC)

I share Jaydiem's concern and thinking, though hesitate about his/her proposed solution. The word "transition" usually goes with the preposition "to", and so "Mac-on-XXX" doesn't quite sound right to me. Here's a suggestion, trying to balance brevity and precision: "Transition of Macintosh platform to XXX architecture", where XXX is either "ARM" or "Intel". "Macintosh" is more formal and I think more fitting than "Mac" in the context of a title, but perhaps the latter could be preferred over the former if brevity is a concern. Biblib (talk) 20:27, 25 June 2020 (UTC)

@Biblib: Thanks. Yes, "transition" calls for a from and a to, but either can be inferred from adequate context, and either can be omitted if it's irrelevant to the topic being discussed. Something of the form, "Jaydiem's transition to orange juice" is fine for an article about an important development in my preference of beverage, particularly when the emphasis is on "where" I ended up, not "where" I came from. The particular beverage(s) I may have preferred before that can be mentioned in the article, to the extent that it matters, but need not appear in the article title. Of course, one of the nice things about the English language is that it allows the use of nouns as functional adjectives, like this: "Jaydiem's orange-juice transition". Between these two specific examples, I think the former would be the better choice of title — but in the real-world case of this article, it's not so simple, because the concept of "Macintosh platform hardware based on the ARM architecture" is a lot harder to convey succinctly than "orange juice". However, the construction "Mac-on-Intel" or "Mac-on-ARM" gets the essential meaning across very efficiently, provided there's adequate context with which to interpret it.
⋮
Another thing is that it's probably more linguistically accurate to associate a willful or active transition — as opposed to a passive one, like a piece of iron rusting — with the actor driving the change, rather than the object being changed. In the present case, Apple (as a company) is the actor, while "Mac" (as a product line) is the object being changed. That's a point in favor of "Apple's" in the title: that single word serves the dual purposes of both naming the actor rather than the object, and of providing semantic context helpful in interpreting "Mac" and "ARM".
⋮
Although I don't have explicit citations at hand, my personal recollection from the circa-2005 period of Mac history is that the transition from PowerPC to Intel processors was frequently referred to in reliable sources at the time with the constructed compound adjective "Mac-on-Intel", as in "Mac-on-Intel strategy" or "Mac-on-Intel software". It's simply a shortening of "Mac (computers and operating system) on Intel (processors)". (I also remember "Macintel", though I think that was generally considered too informal for use in journalism.) Assuming my recollection is accurate, that would provide another point for using similar wording in the title of the "to Intel" article, and by extension, to this one.
⋮
Lastly, I could be wrong, but I feel pretty sure that no one would expect to open a conventional encyclopedia and find an article with title so extremely terse as "Mac transition to ARM". The full-flowing alternative, more in Biblib's style, would probably be "Transition of Apple's Macintosh platform to the ARM hardware architecture", but I think we all agree that's too long for Wikipedia (though it's a good short description). I still think "Apple's Mac-on-ARM transition" is an optimal compromise between the two extremes. — Jaydiem (talk) 16:05, 26 June 2020 (UTC)

My inclination is to leave things as they are, with Biblib's suggestion as second choice. Mac transition to ARM is succinct and clear enough to people with interest in the subject. We now have the short description which provides full context to general audiences. The Intel transition was not exactly parallel as it involved pretty much the entire Apple product line back then. The only exception was the iPod, which was not viewed as a significant software platform; the iPhone did not come out until 2007 and third party apps came later. By contrast, in 2020 most of Apple's products are ARM based as is most of Apple and third-party software development. This time the Mac is the last holdout.--agr (talk) 20:58, 25 June 2020 (UTC)

@ArnoldReinhold: Thanks. Well, as has been pointed out recently on this Talk page (see above), Apple's involvement with products based on ARM processors goes all the way back to the Newton, in 1993. The first iPod was introduced in 2001, and iPods ran on ARM-based chips, never on Intel. The iPod line was an important and profitable component of Apple's overall business as of 2005, so you can't dismiss it as irrelevant. Therefore, I would respectfully submit that to suggest that Apple as a whole "transitioned" to Intel processors in 2005 is just as inaccurate as it would be to suggest that Apple as a whole began "transitioning" to ARM processors in 2020. In this context, whether the iPod of 2005 supported third-party software development is beside the point. — Jaydiem (talk) 16:11, 26 June 2020 (UTC)

Boot Camp

I added this implication under Consumers and not Intel because it is the consumer's choice to buy a licensed Windows to run on a Mac, and it won't work on Apple Silicon unless there was an ARM version of Windows, which there is not. (Note: Apple's Boot Camp integrates better than a branded PC IMHO thus this is a significant feature.) Shencypeter (talk) 14:41, 28 June 2020 (UTC)

Describing the transition as "to ARM" vs. "to Apple Silicon"

Howard Oakley is a longtime writer of deep-dive technical articles about the Mac platform, most of which are published on his blog, The Eclectic Light Company. Oakley recently wrote something that struck me as salient to the editing of this Wikipedia article:

These aren’t ARM-powered Macs any more than iPads or iPhones are: officially they use Apple Silicon, a distinction which is more than a matter of branding. Although the CPU cores are at their heart, much of Apple’s success with them is the result of integrating them into a complete System on a Chip (SoC), in which Apple has invested many years of engineering effort and money. Had it not done so, none of these products would have succeeded.

— Howard Oakley, Big Sur and Apple Silicon: Interesting Times Ahead (2020-06-24)

Essentially, Oakley argues that what really matters about the Mac's latest hardware transition is not that the new chips will use the ARM architecture, per se, but rather that the chips will be completely custom-designed from scratch by Apple's own engineers to optimize their performance. This would suggest that the title and content of this article should portray the key distinguishing feature of the new chips as being the origin and goals of their design, rather than the singular detail of which instruction-set architecture they use. Thoughts? ~ Thanks, — Jaydiem (talk) 21:25, 29 June 2020 (UTC)
_

Another way to think about it is that the previous transition could be described in two ways:

• for business reasons, namely that IBM and Freescale weren't coming up with newer desktop/notebook processors fast enough for Apple (IBM probably mainly cared about servers and Freescale probably mainly cared about embedded processors), whereas Intel cared very much about desktops and notebooks, Apple decided to make an {IBM,Freescale} -> Intel transition;
• Intel were making x86 processors, not PowerPC processors, so, from a technical point of view, this was a PowerPC -> x86 transition, with all the requirements it imposed on Apple and developers.

This transition can also be described in two ways:

• for business reasons, namely that 1) Intel are having trouble with their roadmap and 2) Apple thinks (arguably correctly) that they can do a better job designing their own SoCs, Apple decided to make an Intel -> {Apple-designed processors,Apple Silicon,Apple silicon} transition;
• Intel were making x86 processors and probably wouldn't license the architecture to Apple, whereas Arm did license the architecture to Apple a while ago, so Apple had plenty of experience with ARM instruction sets, so, from a technical point of view, this is an x86 -> ARM transition, with all the requirements it imposes on Apple and developers.

I.e., the transitions weren't and aren't done because Apple wants to transition the Mac to a different instruction set (although there were benefits that came from both - the ability to dual-boot Macs with Windows and to run x86-based virtual machines in the first case, the ability to run iPadOS applications on the Mac without recompilation or translation in the second case), they were and are done for other reasons.

We're describing the first transition as a "transition to Intel", not a "transition to x86", so we could also describe this transition as a "transition to Apple SoCs", not a "transition to ARM". Guy Harris (talk) 22:14, 29 June 2020 (UTC)

• @Guy Harris: I think you're headed in the right direction. The term "SoC" or "SoCs" may be too arcane to use in an article title (it's even worse than "ARM"), but there are alternatives that are more or less synonymous: "chips"; "processors"; "silicon" (a term that has gained new notoriety lately, though it sounds a bit odd to some ears, including mine); "CPUs"; "hardware"; "circuitry"; or "chipset(s)". Accordingly, here are some suggestions:
  
  • Mac transition to Apple { hardware | chips | chipset | chipsets | processors | silicon }
  • Mac transition to Apple-designed { hardware | chips | chipset | chipsets | processors | silicon }
  • Mac transition to Apple Silicon { hardware | chips | chipset | chipsets | processors } (this treats "Apple Silicon" as a compound adjective, like a brand name, modifying the last-word noun)
  • Mac transition to Apple ARM {hardware | chips | chipset | chipsets | processors | silicon }
  • Mac transition to Apple–ARM { hardware | chips | chipset | chipsets | processors | silicon } (using an en dash)
  • Mac transition to Apple ARM-based { hardware | chips | chipset | chipsets | processors | silicon }
  • Mac transition to ARM-based { hardware | chips | chipset | chipsets | processors | silicon }
  Thoughts? — Jaydiem (talk) 20:47, 2 July 2020 (UTC)

I'd say "Apple" should be in the title - one reason for the transition is to bring CPU design for Macs inside Apple, to give them greater control over features, schedules, etc.. In this particular case, "Apple Silicon" would be one possibility, because we're only talking about Macs, where Apple's using the term "Apple {S,s}ilicon" for the SoCs they're designing (which they don't, at least yet, seem to be doing for the iPhone/iPad/iPod touch SoCs). Having ARM as well indicates that a consequence of the transition is an instruction set change, so there's stuff Apple and developers have to do, although I'm less tied to having "ARM" in the title than having "Apple" in the title. My inclination is towards "Mac transition to Apple-designed {something}". {something} should probably not just be "hardware"; my current MacBook Pro has "Apple-designed hardware", it just doesn't happen to have an Apple-designed CPU, although it might have some Apple-designed ASICs, and it definitely has an Apple-designed case, for example. As some Apple-designed "chips"/"chipset"/"chipsets" may already be in current Intel-processor Macs, I might go for "Apple-designed processors" - yes, there's more to it than just the CPU, but that might be the least confusing title for the layman. Guy Harris (talk) 21:59, 2 July 2020 (UTC)

At this point, I'm feeling inclined to support the title Mac transition to Apple silicon (lower-case "s"). I just checked over the past couple of hours, and found there are numerous reliable sources using that phrase, and after the first couple of days following the announcement, there seems to be a strong trend toward using the lower-case "silicon". ZDNet and TechSpot are sticking with "Apple Silicon" so far, but C|Net, Ars Technica, USA Today, and Thurrott are solidly in the "Apple silicon" camp. A certain AnandTech article skirts the issue by avoiding the two-word phrase altogether, writing out "Apple's own Arm-based custom silicon" instead (though that does end with the little-S "silicon"). Incidentally, I also did a quick search on the USPTO's website, and "apple silicon" is not presently registered as a trademark.
⋮
I could also be persuaded to support "Mac transition to Apple-designed processors", or "Mac transition from Intel to Apple processors", but "Mac transition to Apple silicon" seems just as clear, and more succinct. — Jaydiem (talk) 23:21, 2 July 2020 (UTC)

• I agree that calling it a transition to "Apple Silicon" would be a more representative title. --17jiangz1 (talk) 20:08, 2 July 2020 (UTC)

Fifth transition?

The article states that this is the third transition (m68k -> ppc -> x86 -> arm). I actually think it is the fifth; m68k -> ppc -> ppc64 -> x86 -> x86_64 -> arm64. The reason for wanting to state it that way is that ppc and ppc64 are not the same architecture, and some of the technologies (universal binaries) that apple designed for the ppc -> x86 transition were used again for the x86 -> x86_64 transition, too. Versions of macOS that support 32-bit x86 programs and/or processors also ship with universal binaries for the two subarchitectures of x86, so in that regard it is a transition too.

Thoughts? — Preceding unsigned comment added by 196.251.239.242 (talk) 16:25, 30 October 2020 (UTC)

That sounds good to me. It demonstrates Apple keeping their platform transfer code in use even between the more famous transitions. --Octavo (talk) 04:02, 17 March 2021 (UTC)

To what "platform transfer code" are you referring?

For the m68k -> ppc transition, the "platform transition code" was a combination of a 68k emulator running on PPC and, perhaps, some of the classic Mac OS code being written in high-level languages that could be compiled into 68k or PPC code.

For the ppc -> x86 transition, the "platform transition code" was a combination of Rosetta, based on Transitive Technology's binary-to-binary translation code, rewriting or changing low-level platform code, and "it's a BSD UNIX plus C, C++, and Objective-C frameworks running atop it, so the bulk of it can be recompiled - yes, there could be byte order issues, but the code was being maintained in parallel to try to keep byte-order issues from sneaking in". There's nothing in common between that and the classic Mac OS transition support.

For the ppc -> ppc64 transition, the "platform transition code" consisted of rewriting or changing low-level platform code plus "make sure the code is 64-bit clean". That involved some significant work (and the help of gcc and clang having the -Wshorten-64-to-32 flag added to catch some issues at compile time).

For the x86 -> x86_64 transition, the "platform transition code" was rewriting or changing low-level platform code plus "hey, we already did the 64-bit work, yay!"

For the x86_64 -> arm64 transition, the "platform transition code" was a combination of Rosetta 2 translation code - which may have been written from scratch by Apple - rewriting or changing low-level platform code, and "it's a BSD UNIX plus C, C++, and Objective-C frameworks running atop it, so the bulk of it can be recompiled, and *this* time we don't have byte-order issues, although we may have memory ordering issues that require memory barries".

So I'm not sure what "platform transfer code" was in use here - especially not in common use between the one classic Mac OS transition and the transitions done with macOS. Guy Harris (talk) 05:34, 17 March 2021 (UTC)

Inconsistent capitalization of “silicon” in “Apple silicon”

This article (including its title) does not consistently capitalize the word “silicon” in “Apple silicon.” Apple’s official style appears to be “Apple silicon” (example) and the lowercase version is used on the Apple-designed processors page, so this page should probably be updated to reflect that. 67.6.15.171 (talk) 17:12, 15 January 2021 (UTC)

Mac transition to PowerPC?

I think this would be a good article to have to complete the transition timeline. This also occurred when System 7's future was uncertain with projects like Star Trek and Copeland were being explored. 77.103.193.166 (talk) 21:19, 27 March 2024 (UTC)