You might easily be using a Mackster Air.
September 19, 2015 10:10 AM   Subscribe

 
meMac
posted by Sys Rq at 10:18 AM on September 19, 2015 [1 favorite]


The key selling point of the iMac was that it was the first computer with an internal modem.

I'm sure that I was aware of this at the time but I had completely forgotten it. It seems worth remembering.
posted by shakespeherian at 10:29 AM on September 19, 2015 [1 favorite]


Actually, Apple introduced the internal 2400 data/fax modem card for its Macintosh Portable in 1989.
posted by HuronBob at 10:30 AM on September 19, 2015 [12 favorites]


#Macsplaining
posted by Flashman at 10:38 AM on September 19, 2015 [25 favorites]


I'm sure that I was aware of this at the time but I had completely forgotten it.

It's funny to think that internet connectivity was a selling point.
posted by ChurchHatesTucker at 10:39 AM on September 19, 2015


Am I misremembering, or wasn't it also the first computer with USB? I remember that being a big selling point - that you could plug peripherals in or remove them with the computer running.
posted by wanderingmind at 10:44 AM on September 19, 2015 [5 favorites]


The key selling point of the iMac was that it was the first computer with an internal modem.

I guess that could be true for certain definitions of "computer" that ignore most other definitions of the word...
posted by ElDiabloConQueso at 10:44 AM on September 19, 2015 [14 favorites]


It wasn't due to the inclusion of an internal modem, it was because the iMac was the first Apple desktop to lack a floppy drive

I have no floppy drive!
posted by Aznable at 10:54 AM on September 19, 2015 [6 favorites]


It was odd, then, that the tech genius advocated "MacMan", a name that simultaneously evoked the period's iconic portable device, the Walkman, and its most popular computer game, Pac Man .

WTF
posted by snuffleupagus at 11:12 AM on September 19, 2015 [2 favorites]


This was both poorly written and poorly researched. The Walkman and Pac Man were iconic for the late Nineties?

Good job!
posted by rock swoon has no past at 11:24 AM on September 19, 2015 [22 favorites]


I just recently watched the iMac introduction presentation. It's worth a watch if you're into this kind of Apple/Steve history.

Also released at the same time: PowerBook G3. The beauty shots of those are sort of hilarious today, though the hot-swappable bays still seem pretty cool. Also Phil Schiller's entire look is amazing.
posted by wemayfreeze at 11:43 AM on September 19, 2015 [1 favorite]


Am I misremembering, or wasn't it also the first computer with USB?

It was the first one that only had USB (no serial, parallel, ADB, etc).
posted by dirigibleman at 12:09 PM on September 19, 2015 [4 favorites]


> The key selling point of the iMac was that it was the first computer with an internal modem.

What? No. That wasn't true of personal computers. That wasn't even true of Apple products. The Performas that Apple was making when the iMac was announced had optional built-in modems.

The main advantage of the iMac was that it was a personal computer you could use straight out of the box: You didn't need to buy any peripherals to use it, because the CD drive, monitor, and speakers were all built-in. At most you had to buy a printer and start an ISP account to get online.

And the key selling point was that it was cheap -- $1k was cheap for Apple hardware, not to mention decent-quality PC hardware. And you could buy one without Circuit City nickle-and-diming you with accessories or upgrades you're not certain you might need.

When the iMac was announced, I was working at a shop that specialized in used Mac sales and off-warranty repairs. It was immediately obvious the shop's days were numbered. First, because the Performas were mostly awful and there was no point in buying a used Performa for $899 when you could get a new, better iMac for $999 -- and we couldn't afford to drop our prices much while keeping a sustainable margin. Second, nobody was going to want a beige Mac any more.
posted by ardgedee at 12:11 PM on September 19, 2015 [10 favorites]


Two fun quotes:

"We said what is the largest display you'd ever want in a consumer computer? and the answer to us was a 15" running 1024x768." — Bill Gates Steve Jobs

"The back of this thing looks better than the front of the other guys, BTW"
posted by furtive at 12:14 PM on September 19, 2015 [3 favorites]


I thought the i was for, you know, the first person. Or "individuality" or whatever. Because the iMac came in colours to express yourself with. (Assuming your self was made of sherbet or jelly or something.)
posted by Sys Rq at 12:24 PM on September 19, 2015 [2 favorites]


"We said what is the largest display you'd ever want in a consumer computer? and the answer to us was a 15" running 1024x768." — Bill Gates Steve Jobs

Laugh but a lot of web content is based on a 1024x768 minimum resolution and cheap laptops have almost universally settled on widescreen 768p for 15" panels.

For less discerning consumers 768p is still an every day reality in the year 2015.
posted by Talez at 12:36 PM on September 19, 2015 [4 favorites]


I bought an iMac for my Mom that August. I suspected the round mouse was silly so I went to CompUSA to get a proper USB mouse, but they didn't have any.

Windows had supported USB since OSR2.1, sorta, but since everyone had serial stuff nobody in Wintel Land was that enthusiastic about actually trying it.

Us Apple people had been using ADB for almost 10 years, which was USB-ish, at least for input peripherals, and USB looked to be a good upgrade in the scheme of things, making Macs less gratuitously incompatible with the Wintel world.

Two decades on I'm typing this on a Gigabyte Z97 LGA 1150 motherboard running a very slightly modified 10.11 GM (an extra unsigned kext for the system controller and a kext to support the onboard ethernet).

Apple's come a long way, now I just wish they'd officially support DIY installs like Microsoft does.
posted by Heywood Mogroot III at 12:40 PM on September 19, 2015


It was a terrible, terrible machine. It crashed like every 10 minutes, right out of the box; it was basically useless for any real work. And it wasn't really all that cheap. But it looked cool and was sold as a fun consumer product, so they sold millions and millions, which made Apple the money to do OS X and iPods.
posted by miyabo at 12:48 PM on September 19, 2015 [2 favorites]


Ken Segall has a great blog with lots of commentary on more recent Apple marketing ideas.
posted by Lanark at 1:02 PM on September 19, 2015


eh, the first iMac was a basic PPC750 "G3" machine. It wasn't any more underpowered than anything else in 1998.

Its main weakness was the decision to go with a pretty sketchy ATI Rage IIc graphics w/ 2MB of VRAM, the Rev B later that year was a silent upgrade to Rage Pro w/ 6GB.

The iMac boom of 1999-2000 did give the company a lot more runway to get OS X out, along with the cash for the board to give Jobs that Gulfstream. These iMacs, like today's, were pretty decent machines overall, if AIOs are your thing.

Apple didn't need a lot of money to do iPods, anyone could have done that product. Just takes putting Toshiba's 1.8" drive in a small plastic box with PortalPlayer OS and SoundJam as the management software, plus swinging the deals with LA to sell digital downloads at $1 a pop.
posted by Heywood Mogroot III at 1:03 PM on September 19, 2015 [3 favorites]


> The key selling point of the iMac was that it was the first computer with an internal modem.

1983 is calling. It wants it's facts back.

(I used to sell a lot of these at Radio Shack In L.A. - Reporters loved them)
posted by AGameOfMoans at 1:11 PM on September 19, 2015 [8 favorites]


Yeah, My first entry on to the bbs world was a Kaypro 2x (or 4x - I forget, the difference was mostly academic iirc) with an internal 300 baud modem. It was a sort of predecessor of the iMac - an all in one device with everything you need to compute with in one bundle, and only a couple of cables to make it work. Of course, CP/M isn't MacOS by a long shot.
posted by wotsac at 1:28 PM on September 19, 2015


The key selling point of the iMac was that it was the first computer with an internal modem.

While the article is interesting and worthwhile due to the first-person nature of the anecdotes, it's poorly edited. The incorrect statement cited above serves as the flagship example. "Mackster" and "Mackintosh" are clearly misspellings of "Macster" and "Macintosh." I found it distracting enough that I didn't finish reading the piece. Too bad, a bit of editorial polish would have cleaned that all right up.
posted by mwhybark at 1:29 PM on September 19, 2015 [3 favorites]


I have a theory that the iMac was particularly successful because it came in the run up to y2k, which if you're young you don't remember as a sort of mass millennial hysteria, and the iMac was literally transparent. The very opposite of black box technology. I think that was very comforting. Not sure if that was an explicit part of th design but it caught the technophobic zeitgeist beautifully.
posted by Rumple at 1:48 PM on September 19, 2015 [3 favorites]


Two memorable ads from the early iMacs.
posted by persona au gratin at 2:02 PM on September 19, 2015


The main advantage of the iMac was that it was a personal computer you could use straight out of the box: You didn't need to buy any peripherals to use it, because the CD drive, monitor, and speakers were all built-in.

That's funny, because I remember there being tons and tons of peripherals for the iMac, of different levels of indispensability, and all sold in colorful plastic shells to match the "flavor" of iMac you had. For example, if you had a computer before and wanted to get your ADB or serial or SCSI devices to work with the iMac, that would be a handful of adapters. The iMac didn't have a floppy disk drive, but there was still plenty of software that still came on floppies, so you better get an external one.

While it's true that the monitor, computer and speakers were all-in-one, that was also true of the original Macintosh, as well as certain models of the Performa.

They did make it kinda sorta easier to add RAM (at least compared to other Mac models, but not compared to Windows desktop PCs).
posted by overeducated_alligator at 2:04 PM on September 19, 2015 [1 favorite]


I'd forgotten the scourge of frequently-crashing computers.

I've been a MacOS user since 1984. I nearly was talked into GeoWorks in the early 1990s. Glad I wasn't.
posted by persona au gratin at 2:08 PM on September 19, 2015


Here, let Drunk Jeff Goldblum explain.
posted by mmoncur at 3:05 PM on September 19, 2015 [5 favorites]


Eh, the original Compaq portable had everything you needed right of the box. So did CBM's SX64. that's back in the early 80s.
posted by rfs at 3:25 PM on September 19, 2015 [1 favorite]


folks can't we all just agree that every computer comes with everything you need, if you just believe in yourself
posted by DoctorFedora at 3:51 PM on September 19, 2015 [16 favorites]


They did make it kinda sorta easier to add RAM (at least compared to other Mac models, but not compared to Windows desktop PCs).

How much easier could it have been?
posted by fairmettle at 3:55 PM on September 19, 2015


It was a terrible, terrible machine. It crashed like every 10 minutes, right out of the box; it was basically useless for any real work. And it wasn't really all that cheap. But it looked cool and was sold as a fun consumer product, so they sold millions and millions,

Did You Know: Every product ever made by Apple has been a worthless fraud sold to gullible rubes under mass hypnosis? That's what I gather from Apple threads on MetaFilter anyway.
posted by Horace Rumpole at 4:04 PM on September 19, 2015 [15 favorites]


Heywood Mogroot III: The iMac boom of 1999-2000 did give the company a lot more runway to get OS X out ... Apple didn't need a lot of money to do iPods, anyone could have done that product. Just takes putting Toshiba's 1.8" drive in a small plastic box

Not really. The iMac boom had puttered out within a few years, because there really was nothing new to add except those lame paisley cases. And the Cube was a disaster because Apple then did not have enough mindshare to convince people to pay way over commodity prices for fancy design. Back in 2000/2001, people were writing stories like:
Steve Jobs - The Graying Prince Of a Shrinking Kingdom

Even in 2004, people were still doubting that Apple could use its new-found digital music niche to muscle back into mainstream computing, because there had yet been no "halo effect" seen on Mac sales:
If He's So Smart...Steve Jobs, Apple, and the Limits of Innovation

Apple was seriously down to its last few hundred million in the bank, even after getting the Microsoft emergency cash injection to stay afloat (which MS benefitted from because it was in the middle of its criminal monopoly antitrust shenanigans, and keeping its main "competitor" afloat seemed like a win). But Apple somehow managed to grab an exclusive buy for all of Toshiba's 1.8" drives for nearly two years (and until Hitachi later brought out small runs, the only 1.8" supplier), precluding any competitor from launching a similarly small magnetic drive digital player. That was an incredibly significant contract win from a position of relative weakness.

In later years, when it had much more money in the bank and wanted to launch and then upgrade the memory on its solid-state digital players (where Jobs had spent several years dissing competitors' models, because Apple until then had no interest in controlling flash RAM prices), Apple locked in a similar exclusive buys from (mainly) Samsung for its newest, larger flash RAM chips. This had the effect of driving up the price for available chips on the spot market, impairing competitors' ability to launch much cheaper players and cutting into their margins. Apple still uses this tactic today, although given the size of the market, its adverse pricing power wrt competitors has relatively declined. But it still routinely gets to hurt its competitors' margins by issuing basically saying "dibs" over so many essential components.
posted by meehawl at 4:12 PM on September 19, 2015 [7 favorites]


shakespeherian: "It seems worth remembering."

As a big ol' lie. At a minimum many laptops had built in modems well before the iMac.
posted by Mitheral at 4:16 PM on September 19, 2015 [2 favorites]


Apple certainly marketed the built-in modem as a feature, but they definitely weren't claiming it was a first or unique feature. That seems to be a construction of either the article's author or the interview subject.
posted by ardgedee at 4:39 PM on September 19, 2015 [1 favorite]


I didn't read any of the comments in this thread, has anyone pointed out that many computers had internal modems before the iMac yet? Because they did, chumps. They totally did.
posted by Kwine at 5:06 PM on September 19, 2015 [4 favorites]


The most fascinating part is that the name iMac didn't come from Jobs and that he totally hated it when it was proposed. But then he came around and blessed the name, after testing how it looked and played on this "inner circle". The man wasn't perfect, not be a long shot, but when he was good, he was great.
posted by Brandon Blatcher at 5:17 PM on September 19, 2015 [1 favorite]


Apple didn't need a lot of money to do iPods, anyone could have done that product. Just takes putting Toshiba's 1.8" drive in a small plastic box with PortalPlayer OS and SoundJam as the management software, plus swinging the deals with LA to sell digital downloads at $1 a pop.

"Just."

This is one of my red-flag, no-hire, you've-failed-this-interview words. Anyone who says "just" about putting new hardware into a new enclosure and writing new on-device and client software that makes consumers happy while facilitating third-party contractual and business relationships has no idea whatsoever, none, how hard it is to ship a meaningful product.
posted by mhoye at 5:44 PM on September 19, 2015 [29 favorites]


How much easier could it have been?

Oh, hm. I remember for some reason a situation where you had to release the whole outer casing from the top seam? But I'm probably thinking of another model or a hardware upgrade other than just installing RAM... Does anyone else remember what I'm thinking of?
posted by overeducated_alligator at 6:47 PM on September 19, 2015


Does anyone else remember what I'm thinking of?

The Mac pro towers (G3-G5 and later) were a joy to work with, and you could fold them open from the side with one latch; everything in it was laid out and beautiful and easy to access, no screws, no tools required.

That was a standard feature on beige-box Dells from the previous century.
posted by mhoye at 6:53 PM on September 19, 2015 [2 favorites]


It was a terrible, terrible machine. It crashed like every 10 minutes, right out of the box; it was basically useless for any real work. And it wasn't really all that cheap.

Damn there is some serious revisionist history in here.

It wasn't an amazing deal, but the rest is basically untrue. It had a normal or even good spec of everything else. Middle of the road drive size, solid amount of ram, excellent display.

Did everyone forget how bad and low res monitors were at the time? 800x600 was still totally normal. Many applications and even websites still displayed correctly without much or even any scrolling at 640x480! 1024x768 was the equivalent of 1080p now. Machines were available with less, but it definitely slotted in as "solidly upper midrange" with the maximum resolutions available being what, 1600x1200 on truly highend workstations and maybe 1280xSomething for TOTL consumer stuff?

Go look at a Best Buy catalog from that era, they're all online. It was an upper middle of the road price for the specs, and other companies were rooking people plenty hard.

And yea, classic Mac OS sucked at that point, but the iMac ran it just as well as anything else available at the time. It wasn't a sideways garbage can compared to the grey(or b&w) G3.

I think that year range is blending together in people's heads because compitrs changed so fast then. A "decent" amount of ram changed basically every year. And now we've been at 8gb being upper midrange or even somewhat high end for what, almost 10 years?
posted by emptythought at 7:15 PM on September 19, 2015 [5 favorites]


man oh man I forgot about infrared. That 1998 video is giving me serious bondi-blue-colored macworld expo flashbacks

yes i was a cool kid
posted by you're a kitty! at 7:44 PM on September 19, 2015


It's not really the specs I had a problem with, just the crashiness. I was on the school newspaper and responsible for keeping about 10 of them running. Even compared to contemporary machines running 98 or whatever, they crashed all the frigging time, usually taking important files with them. It was just a dark, dark time for Apple -- they were very clearly patching their OS out of old releases and chewing gum or something. We upgraded to OS X Beta 1 I think and at least the crashes were constrained to the old OS emulator.
posted by miyabo at 9:02 PM on September 19, 2015


At a minimum many laptops had built in modems well before the iMac

Funny thing was the iMac was basically a laptop + CRT in a very pretty case.

But laptops were niche in 1998.

Apple was seriously down to its last few hundred million in the bank, even after getting the Microsoft emergency cash injection to stay afloat

FY1997 was horrible for Apple, it's retained earnings went from $1.6B to $600M:

http://investor.apple.com/secfiling.cfm?filingid=1047469-97-6960&cik=320193

but Gates' $150M was neither here nor there for that, just one middling quarter's profit (or loss).

That was a 'kitchen sink' year --its Q2 especially -- and Apple turned a $300M profit on much lower sales in FY98 (before the iMac had any impact):

http://investor.apple.com/secfiling.cfm?filingid=1047469-98-44981&cik=320193

R&D was cut in half from 1996, down to $300M.

In FY99, iMac boosted Apple's unit sales 25% and put Apple's retained earnings back to its pre-crisis position of $1.5B.

FY00 retained earnings were at $2.3B, tho Apple hit an earnings wall in its 4Q -- FY01 was flat for the company earnings-wise but its gross margin fell by $1B as unit sales fell 1M from FY00.

iMac was tired and the "Flower Power" skins weren't going to do it.

2002 was treading water for the company, unit sales were stuck at 3.1M but retained earnings were still $2.3B and the stock with its $5B market cap was trading for not much more than its enterprise value.

If memory serves the December 2002 ESPP shares went for something like $12, so an employee putting $5000 into the plan received ~400 shares, worth $600,000 yesterday.

Anyhoo, in 2002 Apple's three tent poles were PowerMacs at $1.4B (down 17% YOY), LCD iMacs (ever notice how the 2002 iMac's LCD is basically an iPad??) also at $1.4B (up 30%) and "Other", $1.2B, up 25%.

2003 was more of the same, flat unit sales (down to 3.012M), no earnings to speak of.

Powerbooks started eclipsing PowerMacs, drawing almost even in revenue. Apple did $1B in "peripherals" tho, up 60% YOY, as the sweet iPod money rafted in -- $345M in 2003.

2004, Mac sales were flat still rising PowerMac G5 ($1.4B) and Powerbook ($1.6B) sales offset by tanking iMacs. iPod revenue tripled to $1.3B, putting it on par with the Mac categories . . . Apple sold 4.4M iPods that year.

2005 was a magical FY for Apple, a billion-dollar profit year, actually a billion more than 2004, which was OK at $300M. Sales revenue up 70% -- retail doubling to $2.3B.

Apple's music products were $5.4B that year as they sold 22.5M iPods @ $200 apiece . . . retained earnings went up to $4B, up from $2.6B.

Apple was out of the woods, LOL. Note that in the 2000-2004 doldrums they were never really "in" the woods, as they aggressively cut costs in 2001-2002 to keep losses to a minimum.
posted by Heywood Mogroot III at 9:14 PM on September 19, 2015 [7 favorites]


Ah, much of 1997's loss of cash was due to the ~$400M that went to NeXT that year, not operating losses per se.
posted by Heywood Mogroot III at 9:36 PM on September 19, 2015


It's not really the specs I had a problem with, just the crashiness. I was on the school newspaper and responsible for keeping about 10 of them running.

This was a Mac OS problem, not an iMac-specific problem. I too supported many Macs during those years. The iMacs were not less stable than any of the others (if anything, they might have been a little better, especially since there was no SCSI to deal with). The OS was unstable, and a bug in any app could bring the whole thing down. Windows of the era wasn't very stable either. But if yours were literally crashing every 10 minutes, you probably had some other badly behaving hardware or software involved.
posted by primethyme at 10:32 PM on September 19, 2015 [4 favorites]


This was a Mac OS problem, not an iMac-specific problem.

This is also my recollection. iMac hardware, after the initial GPU revision, was adequate for that years OS release. Which may be faint praise, but still.
posted by snuffleupagus at 11:17 PM on September 19, 2015


Here is a video of someone upgrading the RAM in the original Mac. They certainly didn't make it easy, and according to the video, possibly fatal.
posted by adept256 at 11:20 PM on September 19, 2015


I remember one of OSX's selling points was that a crash in one program wouldn't (necessarily!) bring the whole thing down.
posted by persona au gratin at 11:24 PM on September 19, 2015


This was basically parity with Windows (actually better, because OS X is built on BSD, but that was extraneous detail for marketing) in contradistinction to the original Mac OS in which multitasking was cooperative and memory protection was lacking—thus the ability for any misbehaving executable to bring down the system.
posted by snuffleupagus at 11:28 PM on September 19, 2015


Windows 95 ~ Me had memory protection but not true preemptive multitasking, as the typical 32-bit process often had to thunk down into legacy 16-bit code (of which there was a lot), taking a lock along the way. Should the process crash with that lock, the machine was locked up (symptom sometimes being a black Start button).

Windows NT didn't have this issue. Technically, layering the Macintosh GUI over NT would have been similar to going with NeXT. But the Windows way of doing things turned out to be off the mainstream thanks to the industry largely sticking with Unixy workflows.
posted by Heywood Mogroot III at 12:12 AM on September 20, 2015 [1 favorite]


This was a Mac OS problem, not an iMac-specific problem.

Yep. I have one of the original colored iMacs here, which has been upgraded to one of the early versions of OS X (the one that still works with non-Intel CPUs, I forget which) and it's perfectly stable and works fine to this day. We just sold another one in the same condition. Nothing wrong with the hardware at all...
posted by mmoncur at 4:09 AM on September 20, 2015


Apple had been trying to replace 'classic' OS since the late 80s, but all that came out of the efforts was a lot of flailing. Taligent and Copland were large, money-gobbling projects that died before birth.* By early 1997, smart money was on Apple buying either NeXT or Be to replace an OS that was struggling under its own weight.
* And, as an aside, institutional memory of working with IBM on Taligent, not to mention IBM and Motorola on PPC, probably contributed a lot of reinforcement to Apple's 'not invented here' isolationism.
posted by ardgedee at 4:47 AM on September 20, 2015


This is one of my red-flag, no-hire, you've-failed-this-interview words. Anyone who says "just" about putting new hardware into a new enclosure and writing new on-device and client software that makes consumers happy while facilitating third-party contractual and business relationships has no idea whatsoever, none, how hard it is to ship a meaningful product.

Rapidly prototyping DIY != actual product you can ship.
posted by Talez at 6:02 AM on September 20, 2015


And, as an aside, institutional memory of working with IBM on Taligent, not to mention IBM and Motorola on PPC, probably contributed a lot of reinforcement to Apple's 'not invented here' isolationism.

I think that dismissing it as isolationism is shortsighted. Apple realized very early that owning the the entire stack gave them an enormous amount of operational flexibility and ability to drive change that other companies were willingly throwing away to save costs. That investment has paid enormous dividends for them, and for their ecosystem - if you were building your software with XCode and using the Apple-blessed system libs when they revealed the move from PowerPC to x86, for one notable example among many, that transition was one checkbox. If you'd gone third party there, like Borland or something homebrewed, you were in for a bad few years. Likewise owning their own chips, design, supply chain and even the procurement processes for their manufacturing; Tim Cook has basically solved first-mover disadvantage by leveraging their entire production stack, and that's been a huge strategic and logistical advantage for Apple.

For what it's worth, they've been gently telegraphing another major shift in the ppc->x86 vein for a while now in a few quiet ways, like Grand Central Dispatch and Swift implicitly aiming for the strongly-parallel architectures of the future. If you want to start future-proofing your application (or your career) right now is a good time to start writing functional, parallelizable code.
posted by mhoye at 6:48 AM on September 20, 2015 [4 favorites]


Windows 95 ~ Me had memory protection but not true preemptive multitasking, as the typical 32-bit process often had to thunk down into legacy 16-bit code...Windows NT didn't have this issue.

Quite right. I don't think I would've grasped that at the time. But still, it was better than MultiFinder.
posted by snuffleupagus at 7:25 AM on September 20, 2015


Oh yeah, Copland. There sure was a lot of flailing at Apple in the late 90s.
posted by persona au gratin at 8:17 AM on September 20, 2015 [1 favorite]


mhoye: can you translate your last paragraph for me? My CS knowledge consists in one class in Pascal and awareness of basic Unix commands.
posted by persona au gratin at 8:20 AM on September 20, 2015 [1 favorite]


mhoye: can you translate your last paragraph for me? My CS knowledge consists in one class in Pascal and awareness of basic Unix commands.

Apple spends money on their internal development tools and gives them away for free instead of it of Microsoft's approach of seeing it as a profitable line. Because everyone uses Xcode it means that when new technologies come in it's a case of Apple making their libraries use that new technology as much as possible implicitly. You don't even need to recompile the app in some cases.

Take for example the new App Store. It uses an intermediate format now and they call it Bitcode. This means that when Apple releases an A11 or whatever in the future that might have 512-bit vector registers, instead of limiting itself to 128-bit registers it's possible that the backend compiler can try to revector the code to use 512-bit packed representations. This would be a massive increase in performance for any application that uses vectorized code.

This can only really happen because Apple controls the CPU, Apple controls the library, Apple controls the compiler, etc.
posted by Talez at 8:43 AM on September 20, 2015 [1 favorite]


>Apple spends money on their internal development tools and gives them away for free

They give the tools away for free but if you want to take advantage of their App Store - the only way to mass distribute your IOS app then that costs $99. And another $99 if you care to do the same for their OSX app store.

They gave the razors away for free but are charging for the blades now.
posted by AGameOfMoans at 9:37 AM on September 20, 2015


They give the tools away for free but if you want to take advantage of their App Store - the only way to mass distribute your IOS app then that costs $99. And another $99 if you care to do the same for their OSX app store.

$198 for a single developer account for your company. You know how much MSDN costs these days? It's almost impossible to quantify the difference between VS Community and XCode because the VS feature matrix is literally a product marketer's vomit on a webpage and it still costs $99 for a "full" Windows store developer account.
posted by Talez at 10:25 AM on September 20, 2015 [3 favorites]


$198 for a single developer account for your company.

Even that changed this summer; iOS and OS X developer accounts were consolidated, so it's only $99 per year.
posted by bonje at 11:08 AM on September 20, 2015 [2 favorites]


They gave the razors away for free but are charging for the blades now. They don't break this number out, but at their financial scale that $99 per developer fee almost certainly adds up to less than a rounding error. It has the same role as the Metafilter signup fee; a nominal charge that acts as a small barrier to participation, that goes a long way towards keeping out the deluge.
posted by mhoye at 12:02 PM on September 20, 2015 [4 favorites]


mhoye: can you translate your last paragraph for me? My CS knowledge consists in one class in Pascal and awareness of basic Unix commands.

Apple has bought a few companies, released some software and designed some processes in ways that hint at a future where computers feature a large number of lightweight cores, rather than a small number of fast cores. The free-running multithreaded code that can take advantage of those processors, though, has historically been considered brutally difficult to write. So rather than do that, we relied on faster and faster processors to run single-threaded code. The difficulty is that we've stopped getting away with that, for the most part - processors are not speeding up fast enough to cover for our industry's multitude of sins, so another approach is needed.

What we've discovered recently - well, what CS language nerds have been yelling at us for some time, but which we now have a bunch of powerful real-world tooling to actually use - is that the biggest reason contention-free multithreaded code was hard to write was because the tools we've been trying to do that with are antediluvian junk, and that there are in fact way better ways to tell computers what to do than imperative assembler with flames painted on the side that C and its descendants represent. Swift, Rust (a personal favorite) and Go (I hear) will let you take advantage of a large number of CPU cores with little mental overhead on your part as a programmer. Grand Central Dispatch is another tool meant to make it easier to take advantage of multicore CPUs, and as noted above Apple's move to an intermediary data format suggests they'll want to recompile your stuff for you to take better advantage of new processors without having to go back to ask you to do that first.

Between that, and seeing how ARM and Intel CPU design has moved in the last few years - ARM towards a large number of cores per die (sometimes a seemingly ridiculous number) and Intel towards bigger caches, faster I/O lanes and pipelining - and you're left with the impression that Apple's long game is to at least least keep the option of jumping off Intel and on to their ARM-based A-series processors for their entire line, depending on which one gives them better performance and battery life on any given generation. Relatively speaking, it won't be difficult for them to do that ; relative is a huge word there, but compared to Microsoft's abortive attempt to make Windows work on ARM, it's something they can get right. They've done it successfully before, which is something very, very few companies can say. Their control of the developer tools and intermediary representation puts them in a position where backwards compatibility for older applications and developer support for a whole new architecture are solvable, well-understood problems.

Does that make any sense?
posted by mhoye at 12:53 PM on September 20, 2015 [9 favorites]


Yes! Thanks to all.
posted by persona au gratin at 1:27 PM on September 20, 2015


I'm sure Apple has amazing, incredible new system architectures internally -- probably many different ones. Tech companies always invest in trying to anticipate the future. It doesn't necessarily mean things will go according to plan though. Bytecode buys them flexibility without locking them in to any particular architecture. The only thing that's certain is change.
posted by miyabo at 3:48 PM on September 20, 2015


I apologize if asking if "that make any sense" came across as patronizing, which was not my intention at all. I'm always a bit concerned that I sound like a rambling crazy person after I go on a long tear like that.
posted by mhoye at 7:17 AM on September 21, 2015


No, you're fine. As a non-programmer that comment was very interesting.
posted by Brandon Blatcher at 7:44 AM on September 21, 2015 [2 favorites]


Apple's move to an intermediary data format suggests they'll want to recompile your stuff for you to take better advantage of new processors without having to go back to ask you to do that first.

That's makes perfect sense given the number of transitions Apple has made over the past couple decades. (Would have loved to have seen this for Carbon apps.)
posted by ChurchHatesTucker at 11:07 AM on September 21, 2015


« Older "Even the middle class deserves to eat solid food...   |   Custom excavator Newer »


This thread has been archived and is closed to new comments