User talk:Guy Harris/Archives/2020/06

Big- and Little-end of a Computer word and not of an egg

You reverted my revert which reverted A_D_Monroe_III's revert of a contrib of Dscotese. So I won't revert your double revert.
You say: big and little in the case of an egg and of a multi-octet quantity are two ENDs. But a multi-octet quantity does NOT have two ends. It HAS a start (and an end). And this is IMPORTANT. Because it's not either way:

• Big-endianess STARTS big (high-order) and ends little.
• Little-endianess STARTS little (low-order) and ends big.

(I'm unable to perceive how you would define: A big-endian multi-octet quantity ends at its high-order [your words] byte and a little-endian multi-octet quantity ends at its low-order byte. Isn't this kind of tautological?)
Btw, in almost all computers (do you know of an exception?) a multi-octet quantity is addressed by its start address and not by its end address. And the start address is obviously the “end” under consideration. - Nomen4Omen (talk) 06:29, 29 May 2020 (UTC)

@Nomen4Omen: "But a multi-octet quantity does NOT have two ends. It has a start and an end." OK, so what defines the "start" and "end" of a multi-octet quantity? Is the start the octet with the low-order bits, or the octet with the numerically lowest address?

(Note that the document that introduced the terms "big-endian" and "little-endian" refers to two ends of a data word.) Guy Harris (talk) 07:08, 29 May 2020 (UTC)

OK, I thought I talked about “start address” and that this is the one which is the crucial point here. (The start address of a multi-octet entity is the address of the octet with the lowest address and the one by which it is usually pointed to by all pointers and logical references.) Of course, I know that one is able to speak of the start as another (a second?) end. But in our context it is important that this second end is distinguishable from the first one by “numerically lowest address” [as you say] and NOT by “low-order bits” which indeed would help absolutely nothing (tautological as I said above).

In terms of defining something, Danny Cohen's introduction of Little-Endians and Big-Endians are a nice and funny story, but the choice of the terms “little-endian” and “big-endian” is a mistake. Of course we have: little-endian == low-order-at-some-end and big-endian == high-order-at-some-end. But at which end? It is important to distinguish the two ends by other means (than low- or high-order). And Danny Cohen knows it and says: «which bit should travel FIRST» (here: which == low- or high-ORDER and first == low ADDRESS). Nevertheless, his nice, pretty and really funny allusion to Jonathan Swift's Gulliver's Travels (which maybe brought him on the walk of fame) and his subsequent introduction of the words “little-endian” and “big-endian” destroys the point he is really trying to make — for so many many people. - Nomen4Omen (talk) 08:48, 29 May 2020 (UTC)

@Nomen4Omen: In your edit comment, you said "But "end"ianness in place of "start"ianess", which is using the first definition of "end" from this dictionary. In that definition, there's a "start" and an "end". In "endianness", "end" is being used according to the second definition, in which there can be two "end"s (two extreme points); the endianness is a choice of which end will be transmitted first on a network or will have the lower storage address.

Hopefully this edit by Kbrose will make things clearer. Guy Harris (talk) 16:44, 29 May 2020 (UTC)

Application ARM-based chips

Thanks for the clarification over on Apple A7 as to the A-series chips being Template:Application ARM-based chips and that being a better fit than Template:ARM-based chips. Wanted to let you know that I'm going to run through all the A-series articles to make that change… Ehler (talk) 23:10, 22 June 2020 (UTC)

@Ehler: Thanks. While you're at it, fix any links to Apple-designed processors, etc., to go to Apple Silicon and to note that at least some of the future processors will be going into machines that will not be easy to move (unless you order the wheels for the Mac Pro :-)), so they just say "processors", not "mobile processors". Guy Harris (talk) 23:58, 22 June 2020 (UTC)

@Guy Harris: Got a couple of them while I was swapping those out. As far as them being considered "mobile processors," it most certainly doesn't need to change right now on anything other than the A12Z; for the future, we don't know yet whether the chips going into Macs will be branded as part of the A-series or something else. That said, if there's a chip utilized only in a Macbook, I still don't think it's appropriate to call that a "mobile" chip, because connotationally, mobile refers to phones and maybe tablets. Ehler (talk) 00:03, 23 June 2020 (UTC)

@Ehler: Well, Apple's press release says that "This transition will also establish a common architecture across all Apple products", so either 1) Apple will stop offering desktop products or 2) at least one "Apple Silicon" chip will appear in a desktop machine. Apple Silicon doesn't just cover the A-series chips, so even if Apple has a "B1" in the first AArch64 desktop, it'll belong on that page by virtue of being an Apple-designed processor, so that page won't only be for mobile processors. It's OK to refer to a particular A-series processor as being mobile, if it's used only in iPhones, iPod touches, and iPads (although that raises the question of processors used in Apple TVs - they're light, but not exactly "mobile", given that they aren't battery-powered), but Apple Silicon, when mentioned in "See also", shouldn't speak of mobile processors. Guy Harris (talk) 00:41, 23 June 2020 (UTC)

@Guy Harris: Point taken as to where you're seeing that "mobile" reference that needs to be taken out! The Silicon page is definitely referring to both the current and future chip plans (including the upcoming non-mobile). Apple TVs do run on A-series chips, so maybe it's inappropriate to call the A8 or A10X mobile…

@Ehler: That raises the question of what makes a processor a "mobile" processor. It's probably not weight, except for very small devices - I could easily carry a z15 chip, even though I'd probably need a hand truck to move the system cabinet it's in. It's probably a combination of "how much power does it need?" (i.e., how long can it be battery-powered) and "how much heat does it put out?" Those criteria probably divide chips into multiple categories, given that Apple puts different chips into phones and tablets, the former having lower power and heat budgets. Some x86 processors are called "mobile", which probably means "you could put these into a laptop", although "mobile" processors might sometimes be used in desktops as well. Guy Harris (talk) 01:00, 23 June 2020 (UTC)

Thanks

Wanted to say thanks for helping out with the article on Apple's transition to ARM. I don't have as much tech background as you and your edits really helped out! Kiteinthewind ^{Leave a message!} 08:15, 23 June 2020 (UTC)