Is There Anything Stopping Apple from only supporting Macs with a certain RAM amount?

[leman]

Yellow memory pressure on low-tab usage, Apple Mail, Preview, and Facebook Messenger, none of which are doing anything terribly crazy and on a base model M1 Air makes me think that you are at least incorrect about what I've quoted in bold.

Occasional yellow pressure is hardly a cause for concern. As long as there are no performance issues, if it works, it works, right? There are plenty of users who own an 8GB model on these forums, but I don't remember seeing widespread reports of performance issues when performing everyday tasks. There is an increasing umber of testimonies about 8GB causing problems with more intense workloads (e.g. photo editing), which is why I agree that the base configuration should be increased to, say, 12GB. But this won't instantly make 8GB models unusable.

There's nothing to suggest that they won't do this once macOS is entirely Apple Silicon-only too.

So far, you have shown a macOS example from an era where both RAM capacity and requirement was growing very rapidly. And it was not just Apple, but also the rest of the PC industry. I mean, we went from 256MB being standard to 4GB being standard in just a couple of years, which is a 16x jump in capacity! But that's just not the case today, where
both the demand and technological development have slowed down tremendously. Maybe AI will disrupt this trend again, but that will cause much more than changing required RAM capacity.

So yes, there is nothing to suggest that they won't do it. But equally, there is absolutely nothing to suggest that they will do it.

 

[Yebubbleman]

I’m saying memory size is not an issue as it makes no difference whatsoever on the programming effort. The OS might run a little slower on 8GB RAM, but that’s about it. Supporting both M-series and Intel CPUs however is a burden that causes many thousand additional programming hours. So Apple will stop supporting an Intel i9 with 64GB RAM before they drop support for the M1 with 8GB.

I'm not saying that they won't. Just that Apple DOES gate iPhones and iPads based on RAM. As needed, they do so with Macs too. Will they or won't they?

I'm not here to say they will.

I'm only asking if anyone knows enough about .ipsw files and how they're organized to determine whether or not Apple is even able to allow the same SoC on the same model identifier with a higher amount of RAM in while excluding that same SoC on that same model identifier with lower RAM.

Occasional yellow pressure is hardly a cause for concern.

I'm telling you that it's not occasional. Macs are built to sustain the occasional yellow or red memory pressures. When it happens on the regular, it's a problem. When I see the spinning beach ball on the regular, it's a problem. Admittedly, I see yellow pressure more often than I see red. I also see yellow pressure more often than I see the spinning beach ball. But I see yellow pressure pretty much all the time on 8GB of RAM.

As long as there are no performance issues, if it works, it works, right?

C'mon. We both know that regular yellow memory pressure leads to increased swap which leads to degraded SSD which leads to a logic board repair. We both also know enough about how computers work to know that "if it works, it works" doesn't negate the presence of a brewing problem.

I'm not saying I'm melting my logic board, but I know I don't have enough RAM in these machines. Maybe I did with Big Sur and Monterey. I definitely do not with Ventura and Sonoma.

There are plenty of users who own an 8GB model on these forums, but I don't remember seeing widespread reports of performance issues when performing everyday tasks.

You can live in yellow memory pressure and never notice that your memory pressure is yellow. That doesn't mean your Mac isn't doing its job or that it's malfunctioning. It just means you don't notice that it's being given a lot to juggle relative to its juggling capacity. Incidentally, one's mileage will vary greatly with this. I know several folks who won't notice the difference between an M1 with 8GB versus an M1 with 16GB, but that doesn't mean that those people shouldn't have 16GB instead. It just means that they don't notice that they should have 16GB.

There is an increasing umber of testimonies about 8GB causing problems with more intense workloads (e.g. photo editing), which is why I agree that the base configuration should be increased to, say, 12GB. But this won't instantly make 8GB models unusable.

I never used the word "unusable". I can use an M1 Mac with 8GB. But, clearly I'm not supposed to have much open concurrently on it.

So far, you have shown a macOS example from an era where both RAM capacity and requirement was growing very rapidly. And it was not just Apple, but also the rest of the PC industry. I mean, we went from 256MB being standard to 4GB being standard in just a couple of years, which is a 16x jump in capacity! But that's just not the case today, where
both the demand and technological development have slowed down tremendously. Maybe AI will disrupt this trend again, but that will cause much more than changing required RAM capacity.

So yes, there is nothing to suggest that they won't do it. But equally, there is absolutely nothing to suggest that they will do it.

Apple is upping RAM requirements regularly in iOS and iPadOS. While they may very well gate based on SoC when it comes to the M1 Macs, that's not to say that they won't split it down the middle of a subsequent SoC. They have done it before. I'd say there's more to suggest that they will do it (just likely not before naturally dropping all base M1 configurations).

 

[theorist9]

Snow Leopard required 1GB of RAM. If your Intel Mac had less, it wouldn't upgrade. This was correct-able on all of the Macs that this disparity applied to. Similarly, Lion upped that to 2GB where it stayed until Catalina upped it to 4GB.

Interesting! I looked into this and, through Lion, Apple does indeed list requirements by processor and RAM rather than model:

OS X Lion - Technical Specifications - Apple Support

support.apple.com

Starting with Mountain Lion, it lists by processor/RAM and model. And RAM is relevant since, for instance, the base RAM on the 2017 iMac is only 1 GB.

OS X Mountain Lion - Technical Specifications

Apple continues to do this through Catalina, which is the last OS that specifies a hardware requirement (subsequent OS's specify available storage only) However, I'm a bit confused by Catalina's 4 GB RAM requirement, since I looked at the lowest-end model in each category supported by Catalina, and none come with less that 4 GB RAM. So perhaps the requirement is needed only for models that have replaceable RAM, where the user has downgraded below the base configuration:

macOS Catalina is compatible with these computers - Apple Support

You can install macOS Catalina on any of these Mac models.

support.apple.com

BASE RAM FOR LOWEST-END MODELS COMPATIBLE WITH CATALINA
Mac mini (Late 2012) 4GB
iMac (21.5-inch, Late 2012) 8 GB
MacBook (Retina, 12-inch, Early 2015) 8 GB
MacBook Air (11-inch, Mid 2012) 4 GB
MacBook Pro (13-inch, Mid 2012) 4 GB
Mac Pro (Late 2013) 12 GB
[Ignore the different colors--I can't easily get rid of them.]

In any event, Apple's general trend for Macs has been to move away from OS restrictions based on hardware components, and to instead base it purely on model. Thus I would be surprised to see them reverse this. Note also that, for most of Apple's history, the RAM restriction, in particular, wasn't a hard restriction, since (unlike with current models) the user had the option to increase the RAM (and typically, as I recall, for not much money).

 

[Gudi]

I'm not saying that they won't. Just that Apple DOES gate iPhones and iPads based on RAM. As needed, they do so with Macs too. Will they or won't they?

Could you give an example? Because almost all iPhones and iPads of a given year have the same amount of RAM. And when an A5X has more RAM, then because a more powerful GPU needs more RAM. I can not remember any iOS update that required a certain amount of memory?

I'm not here to say they will. I'm only asking if anyone knows enough about .ipsw files and how they're organized to determine whether or not Apple is even able to allow the same SoC on the same model identifier with a higher amount of RAM in while excluding that same SoC on that same model identifier with lower RAM.

Well, of course they are able to do it. They build the entire system, so they can do whatever they want with it. They could even stop software support for space gray Macs but not silver ones. But that wouldn't serve the purpose of cultivating an install base, which serves as a healthy software platform for third party developers.

 

[Gudi]

Interesting! I looked into this and, through Lion, Apple does indeed list requirements by processor and RAM rather than model:

View attachment 2235793

The General Requirements of a piece of software, is the minimum hardware configuration needed for the software to run smoothly. For games it means, they won't work properly with anything less.

The Supported Models are those the OS installer will accept for starting with the installation. You'll need a patcher from some hacker to install the missing drivers to install an OS on unsupported hardware. The unsupported models were excluded because of some older hardware, not the amount of memory. Because as you said, RAM traditionally wasn't a fixed limited resource back in the days. You could always install more.

Going forward macOS will fit into 8GB indefinitely. Some other obsolete hardware will cause M1 Macs to become obsolete. Moore's Law alone will make CPUs so much faster within seven years, that it will seem ridiculous to keep supporting M1 by the time M8 is around.

 

[leman]

C'mon. We both know that regular yellow memory pressure leads to increased swap which leads to degraded SSD which leads to a logic board repair. We both also know enough about how computers work to know that "if it works, it works" doesn't negate the presence of a brewing problem.

I don't think this is a particularly strong argument. I mean, we had a thread that went on for hundreds of pages for this very issue, and yet there don't seem to be any M1 Macs failing in alarming quantities — years after they were first released. Naively, you are of course right — SSDs do have limited lifespans, but it only matters if the endurance is limited enough to actually cause a problem for typical usage of these machines.

Look at it from this perspective: Apple is actually willing to sell you extended warranty (which includes limited accident coverage!) beyond three years of purchase. Would they do it if there was a substantial risk of SSD failure?

I know several folks who won't notice the difference between an M1 with 8GB versus an M1 with 16GB, but that doesn't mean that those people shouldn't have 16GB instead. It just means that they don't notice that they should have 16GB.

The very notion of "should have 16GB" circles back to your original argument of "their machine will die faster with 8GB", which again, is not entirely convincing to me.

 

[theorist9]

The General Requirements of a piece of software, is the minimum hardware configuration needed for the software to run smoothly. For games it means, they won't work properly with anything less.

The Supported Models are those the OS installer will accept for starting with the installation. You'll need a patcher from some hacker to install the missing drivers to install an OS on unsupported hardware. The unsupported models were excluded because of some older hardware, not the amount of memory. Because as you said, RAM traditionally wasn't a fixed limited resource back in the days. You could always install more.

Going forward macOS will fit into 8GB indefinitely. Some other obsolete hardware will cause M1 Macs to become obsolete. Moore's Law alone will make CPUs so much faster within seven years, that it will seem ridiculous to keep supporting M1 by the time M8 is around.

Nope. That was the whole point of my post—for Lion and earlier, the installation requirements are based on General Requirements rather than Supported Models, which is the opposite of the overall claim you are making.

For instance, according to Lion's General Requirements, a Core Solo or Core Duo processor won't work:

You're saying that, since it's a General Requirment, it's a general operational recommendation rather than an installation requirement. Yet the opposite is the case here:
According to this youtube video, Lion can actually run fine on Core Duo; but if it has that processors, you won't be able to install it without a hack:

"Unfortunately, the latest version [of Lion] I was able to get working was 10.7.2, so the latest Lion updates cannot be installed. 10.7.2 seems to run just fine on these machines, however."

 

[Gudi]

Nope. That was the whole point of my post—for Lion and earlier, the installation requirements are based on General Requirements rather than Supported Models, which is the opposite of the overall claim you are making.

These terms are all closely related. Why would you even want/allow to install a software on hardware, which is generally viewed as too slow to run it properly? Why would you include drivers for hardware on which the software is not supposed to be installed anyway? And finally the supported models list is just an easier way to identify whether or not your machine fits the technical requirements. Just remember the year you've purchased it.

You're saying that, since it's a General Requirement, it's a general operational recommendation rather than an installation requirement. Yet the opposite is the case here:

It's both. The software development was not targeted for single-core processors and it doesn't provide the drivers for them either. You can hack the drivers back in and find that the performance is still good enough for our needs, but that doesn't change the recommendation.

According to this youtube video, Lion can actually run fine on Core Duo; but if it has that processors, you won't be able to install it without a hack.

Well, if you've ever heard about the concepts of Turing-completeness and Turing-equivalence, then any computer can emulate any other computer (just not at the same speed). If you're fine with a frame rate of two or three frames per hour, then you can probably run Crysis (with a few hacks) on an Intel 80486. The General requirements therefore set a minimum hardware and software configuration, which the programmer deems necessary to use his software with the intended performance. But you can always "run" the software on a Turing machine, which you've drawn yourself with pencil on a sheet of paper. Mathematically the result will be identical.

 

[Yebubbleman]

I don't think this is a particularly strong argument. I mean, we had a thread that went on for hundreds of pages for this very issue, and yet there don't seem to be any M1 Macs failing in alarming quantities — years after they were first released.

None that we know of. Apple doesn't make mention of those statistics until they put out a Repair Extension Program and even then, it's always "a small quantity of users".

Naively, you are of course right — SSDs do have limited lifespans, but it only matters if the endurance is limited enough to actually cause a problem for typical usage of these machines.

While endurance is the only factor that really matters when talking about failure, the fact of the matter remains that with less RAM, I am reliant on that endurance. I'm still taxing the SSD in ways that I, as a user, would probably rather not do, ideally.

Look at it from this perspective: Apple is actually willing to sell you extended warranty (which includes limited accident coverage!) beyond three years of purchase. Would they do it if there was a substantial risk of SSD failure?

Yes. They sold AppleCare+ throughout the supported lifecycles of several doomed products. See 2007/08 NVIDIA MacBook Pros, 2011 15-inch and 17-inch MacBook Pros, 2012 and early 2013 Retina 15-inch MacBook Pros, all butterfly keyboard Macs (especially 2016 and 2017 2-port 13-inch MacBook Pros, which, on top of a myriad of other Repair Extension Programs, also had one for SSD failures).

And, it's true that if you only have one or two programs running on an M1 MacBook Air at a time with only 8GB of RAM, your memory pressure will consistently be in the green. But, for how most people multi-task in 2023, it's not ideal operation.

The very notion of "should have 16GB" circles back to your original argument of "their machine will die faster with 8GB", which again, is not entirely convincing to me.

My argument is that machines with 8GB will rely more on the SSD's endurance rating than those that don't. I'm not really sure how someone who knows as much about computers as you genuinely do, are disputing that.

In any case, the 8GB vs. 16GB of RAM debate is entirely off-topic. The only thing I was curious about is Apple's ability to gate models of the same machine based on RAM. It's probably unlikely to happen with M1 Macs anyway. Though, I could see it happening with M2 Macs, especially given that every SoC (aside from the M2 Pro, for some reason) now has three different RAM options.

 

[leman]

Yes. They sold AppleCare+ throughout the supported lifecycles of several doomed products. See 2007/08 NVIDIA MacBook Pros, 2011 15-inch and 17-inch MacBook Pros, 2012 and early 2013 Retina 15-inch MacBook Pros, all butterfly keyboard Macs (especially 2016 and 2017 2-port 13-inch MacBook Pros, which, on top of a myriad of other Repair Extension Programs, also had one for SSD failures).

I am not talking about the old AppleCare+. I am talking about the new subscription that allows you to have extended warranty coverage after three years. This is a very recent offering (I think introduced in 2021?). This tells me that Apple is very confident that the SSD won't fail during the 5-6 years users tend to keep their computers.

My argument is that machines with 8GB will rely more on the SSD's endurance rating than those that don't. I'm not really sure how someone who knows as much about computers as you genuinely do, are disputing that.

And my argument that it doesn't matter. This is similar to the "high temperatures will damage your CPU". Yes, it will, but in the timeframes that are entirely irrelevant for normal ownership cycles. Nobody cares if operating a device in a certain way will reduce the ideal lifespan of a component from 60 years to 20 years, the computer will break down or be replaced long before that becomes relevant.

The only thing I was curious about is Apple's ability to gate models of the same machine based on RAM.

Any closed-source OS has that ability. MacOS, Windows, you name it. Especially if it mandates signed kernels that cannot be tampered with. But for as long as there is a way to disable the security so that a patched kernel can be loaded, this kind of limitation can be bypassed.

 

[Yebubbleman]

I am not talking about the old AppleCare+. I am talking about the new subscription that allows you to have extended warranty coverage after three years. This is a very recent offering (I think introduced in 2021?). This tells me that Apple is very confident that the SSD won't fail during the 5-6 years users tend to keep their computers.

I don't know why it gives you this confidence. It's not like Apple is asserting that these things are now more reliable than before. It's also not like Dell, HP, and Lenovo haven't been doing warranty extensions for several years prior to Apple offering AppleCare+ extensions. They were able to strike a different deal with a different insurance contractor that made this work in a way that it didn't before. That's all that happened here. Apple is not doing that to say "hey, look, our devices are so rock solid that we'll guarantee even longer reliable operation than the three years we promised before!". Furthermore, if it was to be worthy of an REP, it, statistically, wouldn't likely be known until the Mac has been on the market for two to four years; we're only at the beginning of that. I'm not saying that we'll ever see an REP for these SSDs; just that you're premature in saying that there isn't going to be an issue stemming from it.

And my argument that it doesn't matter. This is similar to the "high temperatures will damage your CPU". Yes, it will, but in the timeframes that are entirely irrelevant for normal ownership cycles. Nobody cares if operating a device in a certain way will reduce the ideal lifespan of a component from 60 years to 20 years, the computer will break down or be replaced long before that becomes relevant.

Whether it causes failure or even noticeable degradation is not of concern to me. What IS of concern to me is optimal operation. Cool that suboptimal operation isn't something that I'm likely to notice (incidentally, I DO notice it). I'm not about to buy or endorse that someone else buy a piece of hardware costing $850-1000 brand new that operates sub-optimally, even if that difference is never actually noticed. Clearly, there are several users that are okay with this and, honestly, good for them. But for my money, that's not acceptable. And that's seriously all that I want to say on this topic in this thread.

Any closed-source OS has that ability. MacOS, Windows, you name it. Especially if it mandates signed kernels that cannot be tampered with. But for as long as there is a way to disable the security so that a patched kernel can be loaded, this kind of limitation can be bypassed.

I would think this as well, speaking very base level about it. Though, I was more asking if anyone familiar with how ipsw files worked could assert one way or the other about how that would work with Configurator 2. Otherwise, for all other installation methods, I don't see how they couldn't do as they did before.

 

[NewUsername]

Interesting! I looked into this and, through Lion, Apple does indeed list requirements by processor and RAM rather than model:

OS X Lion - Technical Specifications - Apple Support

support.apple.com

View attachment 2235794

Starting with Mountain Lion, it lists by processor/RAM and model. And RAM is relevant since, for instance, the base RAM on the 2017 iMac is only 1 GB.

OS X Mountain Lion - Technical Specifications

View attachment 2235793

Apple continues to do this through Catalina, which is the last OS that specifies a hardware requirement (subsequent OS's specify available storage only) However, I'm a bit confused by Catalina's 4 GB RAM requirement, since I looked at the lowest-end model in each category supported by Catalina, and none come with less that 4 GB RAM. So perhaps the requirement is needed only for models that have replaceable RAM, where the user has downgraded below the base configuration:

macOS Catalina is compatible with these computers - Apple Support

You can install macOS Catalina on any of these Mac models.

support.apple.com

BASE RAM FOR LOWEST-END MODELS COMPATIBLE WITH CATALINA
Mac mini (Late 2012) 4GB
iMac (21.5-inch, Late 2012) 8 GB
MacBook (Retina, 12-inch, Early 2015) 8 GB
MacBook Air (11-inch, Mid 2012) 4 GB
MacBook Pro (13-inch, Mid 2012) 4 GB
Mac Pro (Late 2013) 12 GB
[Ignore the different colors--I can't easily get rid of them.]

In any event, Apple's general trend for Macs has been to move away from OS restrictions based on hardware components, and to instead base it purely on model. Thus I would be surprised to see them reverse this. Note also that, for most of Apple's history, the RAM restriction, in particular, wasn't a hard restriction, since (unlike with current models) the user had the option to increase the RAM (and typically, as I recall, for not much money).

About the Catalina requirement: actually, a ton of computers running Catalina had a base model with 4GB RAM, including all Mac minis before 2018, all MacBook Airs up to the 2015 model, etc. Even Monterey would run on some Macs with 4GB RAM.

Mojave only required 2GB because it supported the 2010 Mac Pro, which came with 3GB RAM.

Edit: never mind, I misread your post. Yes, Apple kept posting minimum RAM requirements even if all models had the RAM anyway, I also don’t know why.

 

[dmccloud]

While endurance is the only factor that really matters when talking about failure, the fact of the matter remains that with less RAM, I am reliant on that endurance. I'm still taxing the SSD in ways that I, as a user, would probably rather not do, ideally.

Even if one just applies the baseline TBW to a 256GB SSD (the TBW most 256GB SSDs are rated for), that's still 750 TB written over the lifespan of the drive. That means the entire disk could be filled to max and completely written over 3000 times before hitting that arbitrary limit. As the SSD capacities increase, so does the TBW rating. The other consideration is that more often than not those types of ratings err on the side of caution and actually are lower than the actual number.

The other number to look at is MTBF (mean time between failures). This is often measured in hundreds of thousands of hours, if not millions. For example, the Samsung 980 and 990 Pro have an MTBF of 1.5 million hours, which is far more time than anyone would be using their computers.

The point is that there are factors besides simply using swap space on an SSD that could have a more direct impact on SSD longevity. It would be more likely (yet still not probable) to lose an SSD due to an electrical fault in the controller, power surge, overheating of the system, or physical damage than simply due to writing too much data on the drive.

 

[theorist9]

These terms are all closely related. Why would you even want/allow to install a software on hardware, which is generally viewed as too slow to run it properly? Why would you include drivers for hardware on which the software is not supposed to be installed anyway? And finally the supported models list is just an easier way to identify whether or not your machine fits the technical requirements. Just remember the year you've purchased it.

It's both. The software development was not targeted for single-core processors and it doesn't provide the drivers for them either. You can hack the drivers back in and find that the performance is still good enough for our needs, but that doesn't change the recommendation.

Well, if you've ever heard about the concepts of Turing-completeness and Turing-equivalence, then any computer can emulate any other computer (just not at the same speed). If you're fine with a frame rate of two or three frames per hour, then you can probably run Crysis (with a few hacks) on an Intel 80486. The General requirements therefore set a minimum hardware and software configuration, which the programmer deems necessary to use his software with the intended performance. But you can always "run" the software on a Turing machine, which you've drawn yourself with pencil on a sheet of paper. Mathematically the result will be identical.

I don't think you understood my post. You're making claims about how software works generally. That's not what my post was about. It was specifically about the kinds of restrictions Apple has historically made, how they categorize them, and what they mean.

For instance, when Apple listed the required processors under "General Requirements for Lion", that referred to what the installer was looking for, not to strict operational requirements. You attempted to "correct" me by saying that, for software generally, the semantics are the opposite—but how other manufacturers may or may not label things is irrelevant for the question at hand here.

 

[Gudi]

For instance, when Apple listed the required processors under "General Requirements for Lion", that referred to what the installer was looking for, not to strict operational requirements.

And I said those terms are all closely related. It is an operational requirement, which is also an installer requirement. Because why install on a configuration, which doesn’t operate with the desired performance? Your nitpicking makes no sense. And there’s no installer which requires a certain amount of RAM, because memory is a variable.

 

[theorist9]

And I said those terms are all closely related. It is an operational requirement, which is also an installer requirement. Because why install on a configuration, which doesn’t operate with the desired performance? Your nitpicking makes no sense. And there’s no installer which requires a certain amount of RAM, because memory is a variable.

On the contrary, you're the one that's been nitpicking here. I posted an accurate summary of what Apple says it does, and you needlessly challenged it--based on semantics--with your post #30.

Honestly, a lot of your posts just seem like you're trying to bait people. And didn't you get banned for a derogatory sexist comment? If you don't like people rejecting your challenges, then you should stop making them.

 

[Gudi]

Honestly, a lot of your posts just seem like you're trying to bait people. And didn't you get banned for a derogatory sexist comment? If you don't like people rejecting your challenges, then you should stop making them.

No, I got banned a dozen or so times. Mostly for being political or otherwise controversial. Sometimes for making fun of people who insulted me first. Anglo-Saxons like their conversations excessively polite and harmless. If I didn't enjoy defending my opinions, I wouldn't do it. 😉

Back to the topic. Nothing is stopping Apple from dropping support based on RAM size, but they won't do it and haven't done so in the past. As the OP assumed, it was indeed a dumb question. Apple could as well drop support for 13-inch laptops but not for 15-inch laptops. This wouldn't make maintaining the code base any easier for software developers, which is the whole purpose of ending software support for old hardware. Small and large RAM age at the same speed. You can not buy yourself longer software support by upgrading the RAM.

This is the answer. If you don't like it, I can't help you. 🤷

 

[thenewperson]

On the contrary, you're the one that's been nitpicking here. I posted an accurate summary of what Apple says it does, and you needlessly challenged it--based on semantics--with your post #30.

Honestly, a lot of your posts just seem like you're trying to bait people. And didn't you get banned for a derogatory sexist comment? If you don't like people rejecting your challenges, then you should stop making them.

That user had absolutely nothing meaningful to say every time they showed up and I had to ignore them.

 

[dmccloud]

And I said those terms are all closely related. It is an operational requirement, which is also an installer requirement. Because why install on a configuration, which doesn’t operate with the desired performance? Your nitpicking makes no sense. And there’s no installer which requires a certain amount of RAM, because memory is a variable.

From where are you drawing this notion that operational requirements are the same as installer requirements? The standard Mac OS installer a lot of apps use has remained largely unchanged for years, so its requirements haven't changed much over the last 2-3 years. For apps that do not use any sort of installer (i.e., applications that are drag and drop into the Applications folder), only operational requirements would even be relevant. You also contradict yourself in this post when you claim that "there's no installer which requires a certain amount of RAM". That alone contradicts your argument that operational requirements = installer requirements.

This is not exclusive to Apple and Mac OS either. While the default Windows installer does check that a system meets minimum requirements for the software being installed, those requirements are not built into the installer itself. Rather, the developer tells the installer what the minimum requirements are, and the installer itself checks the system settings against that list. The downside there is that it's not unusual to see a Windows installer claim a system is not compatible with a given app even though the hardware clearly exceeds those minimum specs.

I also find it odd that you're accusing others of nitpicking when your own comments are much better examples of the practice.

 

[Gudi]

While the default Windows installer does check that a system meets minimum requirements for the software being installed, those requirements are not built into the installer itself. Rather, the developer tells the installer what the minimum requirements are, and the installer itself checks the system settings against that list.

Yeah, that’s what a requirement is. It’s an expectation set by the developer, not an intrinsic property of the software or of the installer itself. If it is printed on the box, then it is a general requirement a strong recommendation so that the buyer can make an informed purchase decision. And if it’s checked by the installer, then it’s an installation requirement to actually prevent the user from installing the software on unsupported systems.

German software developers have learnt a long time ago to prepare for the "dümmster anzunehmender User" or DAU, in English the 'dumbest assumable user'. Of course it’s not enough to tell users the general requirements of your software. You know they will install it anyway and later complain about poor performance all over the internet. So the general requirement is enforced by an installer requirement. But it’s all the same arbitrary minimum set by the developer.

 

[dmccloud]

You use the term "arbitrary" in reference to these minimum specs, but there's nothing arbitrary about them. The definition of arbitrary is "based on random choice or personal whim, rather than any reason or system."

Developers develop software to hit a minimum level of performance (playability on the gaming side). The minimum specs are what is needed to hit that minimum. Larger developers can have a large base of QA specialists and beta testers to help determine where that minimum is, while smaller developers more often use similar apps and their minimum specs as their baseline. In both cases, developers are using some level of testing and/or comparison to determine where those minimums lie, rather than randomly setting the specs as if they're creating a character for a D&D campaign.

 

[TechnoMonk]

Another thread about wanting 16 GB base model, just worded differently.