Hacker News new | past | comments | ask | show | jobs | submit login
Valve drops VR support for macOS (techcrunch.com)
193 points by BSVogler 23 days ago | hide | past | web | favorite | 282 comments



Mac is a horrible platform to support for video game developers. Speaking from experience the issues are:

1. No cross platform builds allowed. We can compile to Windows, Linux, Android, Switch, PS4, Xbox One all from one windows build machine.

2. Shifting sand of non-backwards compatibility. Notice how of the above list everyone except Linux and Android have solid back compat stories. Video games are in development for 2-3 years, and stay on the market for 5+ years with no planned recompiles. So a 7 year period of support after inital build machine setup is expected, anything less is going to get your platform side lined.

3. Small market. Mac users make up a share of sales similar to Linux. "What are you talking about! Linux is 1% while Mac is a whole 4%!", nope! Those numbers are the same insignificance per a video game business plan. Platforms with such low sales, like Stadia, might be support provided the platform owner pays porting costs.

4. Culture. Steve Jobs disliked games, it shows in Apple's support.

5. Horrible hardware. Supposing you've decided to ignore all the downsides, and the small upside still appeals to you. Now you have to tune your game to run on hardware which would put cheap Wal-Mart machines to the test. Apple's install base is almost all integrated older intel GPUs. If you do not support Metal, then you are forced to support these old intel integrated GPUs with bonus hacked up OpenGL drivers.

All points are ones I am speaking from experience on. We shipped TINY METAL on MacOS, the lack of cross build delayed the release. The horrible sales numbers meant carrying forward support was a net drag on the game. All further games I work on do not get MacOS ports for these reasons.


> 1. No cross platform builds allowed.

This is SO annoying. I can cross compile to Windows from GNU/Linux quite easily. In fact with Wine I can even get MSVC running. But Apple, even though they employ many of the Clang/LLVM developers, believes that you must use their OS to target any Apple product.

Your only reprieve outside of buying their extremely expensive hardware is using CI services like Travis or Github actions.


Even if you DO buy a lot of expensive hardware, it’s a giant pain in the ass. There are no good rackmount solutions and so we ended up with a room full of expensive “trashcan” macs and they were STILL the biggest drag on our build infra.

This has been an issue for decades. Twenty years ago I was doing grid/HPC computing at Sun and there were a lot of bioinformatics firms that used macs and wanted compute clusters of them. It was a huge problem even then.


It's possible to mount the trashcan mac pros with unofficial mounts, its just awkward and definitely not intended [0]. You can also mount iMac Pros [1].

[0] https://twitter.com/brianstucki/status/800778773998145536

[1] https://www.macstadium.com/imacpro


It's also possible to outright buy Mac Pro Rack, a $6500 full depth 4U server, that is 1CPU all the way up, that offer neither any redundancy nor any remote management, if that kind of trash matters at all.


The lack of dual power supplies and remote power on/off and good PXE support is crazy for a rack mount computer (won’t even dignify it by calling it a server).


The rack mounted Mac Pro isn’t targeted at the server market, it’s targeted at creative industries (i.e, film/tv production, audio recording, etc) where you’d have it rackmounted into a flight case as part of a mobile post-production unit.


That's a fat machine


What? I can use online CI services to compile for MacOS? TIL.


I think Unity does it (sort of). But yea...


I thought it just popped out an Xcode project for mac/ios builds.


Unity does sort of do it, but even then, the result is not the same as compiling on OSX - in particular, I've noticed differences in the shaders emitted, around HDR/bloom.


You could build a hackintosh instead. Pain to get software set up just right but when it is it works great.


It is not a technical problem, but a legal one. It is that they do not want to expose themselves by violating agreements with Apple.

Do you want to be the genius that saved a few dollars on hardware with a Hackintosh only to get the company sued?


Exactly. If you don't care about Apple ToS, you can cross-compile on GNU/Linux just fine without ever touching macOS.


I hadn’t considered this at all mainly because I haven’t heard of apple taking legal actions against people promoting hackintosh related things and methods. Though I guess it could become a problem for a company.



It's only because most hackintoshers don't make money by selling them.

It also takes a lot of time to setup so hackintoshing will never gain mainstream numbers.


It's too time consuming and risky for a company.


There is a large rate of failure and larger fix times for issues, as well.


I forget if there’s a better way to describe it but usually if there’s going to be a failure, it’ll be at the beginning (unless you’re constantly updating software which does have some risk)


If it’s too risky I’d assume the company would have enough money to not have to resort to building hackintoshes


Of you are thinking of a hackintosh just to compile, then it would probably be easier to just virtualize it.

https://github.com/kholia/OSX-KVM


Sorry, since when is a few hundred dollars “extremely expensive”?


A few? I was just looking at getting MacOS hardware so I could do iPhone builds of a Flutter app. The bottom-end Mac Mini is $800 with 8 GB of RAM, and $1400 with 32 GB, which is what people on forums were recommending for running XCode properly. Is that what you'd call a few?

Looking at System76's equivalent mini, it's $555 for the low end, which has the same RAM and a faster, newer processor. Going to 32 GB takes it all the way to $700, or half what Apple's charging. So yes, "extremely expensive" is what it seems like to me.

Especially given that it should be entirely unnecessary. Just let me do a cross-platform build!


I came here to say more or less the same thing... but for people who aren't native english speakers: a "few hundred" means $300, perhaps $350. Apple doesnt sell anything this cheap, and if you're comparing used Apple hardware to new PC hardware, thats a bit unfair.


well, maybe $400, $500, and maybe you could stretch it to $600 but I agree $800 is well beyond the reach of "a few hundred"


For me $500 and up becomes 'several hundred'.


The Mac mini amazingly has user expandable RAM. Crucial sells kits for up to 64GB. It is still too expensive a machine got what you get.


Only the old Mac Minis from 2012 and before have user expandable RAM, the newer ones have the RAM and storage soldered on.

https://support.apple.com/en-us/HT205041


uhm.. no... the 2018 went back to expandable RAM https://www.ifixit.com/Teardown/Mac+mini+Late+2018+Teardown/...

It is a bit of a pain and took me about 45 minutes, but I was going slow.


Compared to IBM computers, they are overpriced.

Of course, if you want to test your software on Macs, you'll have to buy their hardware sooner or later, but even then you have to set up the build on the system instead of keeping it as pure test system.

If you do it for your day job, it's probably justified, but if you are just a hobbyist, then even hundreds of dollars are expensive. Especially if you live in a poorer area or are still a student.


The insult is that the mac you "you have to buy" is just the same Intel CPU, AMD GPU, Micron RAM and Samsung SSD you already own which can build and virtualise the 10+ other platforms.


IBM computers!? Have not heard for quite awhile. Thought they quit the PC world to Lenovo for quite sometimes la.


The word PC existed long before IBM called one specific model "IBM PC" and since that specific model is no longer in widespread use it makes no sense to refer to regular non mac computers as IBM computers.


I've been part of the Apple cult for a while now. I've even started to bulk at the price / benefit ratio ever since 2017.


(Balk not bulk. Just FYI.)


Autospell on mobile

Maybe you should ask "where", not "when".


Agree with everything except

> 4. Culture. Steve Jobs disliked games, it shows in Apple's support.

IMO their lack of support has less to do with hate and more that they just didn't find value in that demographic. No one is making money off gaming hardware. On the other hand Apple has put in a LOT of effort into iOS gaming. Steve Jobs himself had multiple games showcased in his keynotes.


>No one is making money off gaming hardware.

Huh? I mean, surely nobody is making Apple-money from gaming hardware. But plenty of companies of various sizes are decently profitable and making money in this space.


It wasn't that long ago that gaming consoles were being sold at an astonishing loss, to be made up on services subscription revenue. They're mostly not loss leaders anymore, but the hardware sales are definitely not responsible for significant profits. Video game console hardware has lower margins than gaming PC components, which are already a very low-margin market segment (except where companies can get away with a 30% surcharge by adding RGB LEDs). If Apple tried to compete in that space, they'd face an instant shareholder lawsuit.


>Video game console hardware has lower margins than gaming PC components, which are already a very low-margin market segment

The high end has quite generous margins and the volumes are absolutely insane, which is why we're now doing HPC and ML on what is basically gaming hardware. Nvidia have a broad spectrum of markets, from the high margin and low volume Quadro to the low margin and high volume GTX - in a business that's dominated by NRE costs, that's a good place to be.

Apple providing support for gaming wouldn't necessarily be profitable in itself, but it'd potentially provide the volumes to improve their offerings for professional rendering applications. Apple refuse to do business with Nvidia and they've been hamstrung by AMD's weak offerings, but a gaming-focused Apple could acquire some IP and tape out their own GPUs.


They branded their Apple TV as a game platform.

It has a controller that looks like a console gamepad and is one of the target platform of Apple Arcade.


That's obviously not in the same league. The gaming capabilities of Apple TV are still pretty much an afterthought. The games available are more likely to be ports from smartphones, not Xbox/PS platforms. The processing power is a small fraction of what a video game console or gaming PC requires, which is why Apple can use leftover old smartphone SoCs instead of having to invest in actual high-end chips for the device.

There's the potential for the Apple TV to become a serious gaming platform in the future, but it would require a major shift in product strategy from Apple. That shift would have to include putting much more expensive hardware inside the Apple TV, increasing the BOM by more than the current retail price.


Regarding the hardware, are you deliberately ignoring Nintendo? They have low cost, cheap hardware, good margins, and are able to get a huge part of the market.

I agree on the need of a shift in product development, but the hardware isn’t what is lacking here.


This certainly has not always been the case. Nintendo lost money on consoles same as SEGA and everyone else.


Not sure how that’s relevant to what is discussed here. Would you only consider game console companies that never lost money?


Given the size of the iPhone gaming industry, I think Apple is making Apple-money from gaming hardware.


> No one is making money off gaming hardware.

Where does this nonesense come from? There's a whole market segment just catering to PC non-console games with companies manufacturing RGB lighted RAM, specialized motherboards and plenty of gaming periphelias. Even gaming laptop market is booming.

And that's even ignoring "tiny" companies like nVidia.


Jobs’ view always seemed short sighted to me. Yes, they were a hardware company back when he made these decisions, but software sells the hardware in the market he was in.


> Jobs’ view always seemed short sighted to me

Halo was originally intended as a Mac release, and Steve was livid[1] when Microsoft bought Bungie and made it an Xbox exclusive. I think his apathy toward the games business probably grew from experiences like that, if not that one specifically.

1. https://www.mcvuk.com/business-news/publishing/steve-jobs-ra...


I could see that. Can't remember if Isaacs' biography covered that debacle.


This is the type of bullshit comment that makes me login and bother replying. Anyone in the industry that has needed to work with Apple knows their stance on games is inane. This is in spite of the fact that games make the lion’s share of revenue on their app store and they know it. This in spite of the fact that the majority of games on their iOS platform suck. Games is really not in their dna and boy does it show.


>No one is making money off gaming hardware.

Some gaming keyboards full of fancy LEDs and software and stuff and military-grade whatever retail for the price of a small car.


> everyone except Linux and Android have solid back compat stories

Linux is very backwards compatible. I've run an OpenGL application I wrote 10+ years ago without any trouble and have run non 3D applications that were over 20 years old without a need to recompile. This is because the Linux kernel has a very stable ABI.

Unfortunatly, this isn't the case if you ship with dependencies on dynamic libraries but that's pretty easy to avoid. Static link everything and only dynamically depend on SDL or something.

You could also go the windows route and ship all the DLLs. This is what steam does under Linux, they ship their own Ubuntu 14.04


> Notice how of the above list everyone except Linux and Android have solid back compat stories.

Out of curiosity, what exactly were your struggles with Linux/Android backwards compatibility? Considering how many apps I run still on my phone still use Gingerbread-era UIs, and how Linux executables from as far back as the mid-90's are supposed to be able to run on even the absolute latest kernel (Linus Torvalds having publicly chewed people out on the LKML for breaking said backwards-compatibility), it's surprising that these would somehow be classified as not having solid backwards compatibility.

The main thing that comes to mind with non-Android Linux is library versions, since those do occasionally break backwards compatibility, but this shouldn't matter too much if you're compiling static executables, or using some isolated runtime like the one Steam (IIRC) provides.

As a point of comparison here, World of Goo's Linux version came out in 2009, and it still works on my system (Slackware64-current) which - while not at the absolute most bleeding edge - is still using library and driver and kernel versions substantially more recent than most "LTS" distro releases.


* 3. Small market. Mac users make up a share of sales similar to Linux. "What are you talking about! Linux is 1% while Mac is a whole 4%!", nope! Those numbers are the same insignificance per a video game business plan*

As a Mac user, I buy other hardware to play games on! Probably the only real Apple game platform is iOS. And even then I’m a bit more careful of what games I buy.

The only Mac gaming I do is with emulators like DOS Box. The market must be really tiny.


On the flip side, I have and maintain a Windows computer for the sole purpose of gaming. I hate it to the point where I can’t understand why anyone would voluntarily subject themselves to Windows other than sheer ignorance.

You can’t swap sound outputs while an application is running. Fullscreen Direct3D games can freeze in a way that obscures all other windows, including the Task Manager, preventing you from ever killing them. USB device power management is abysmal, and devices often just won’t wake up after sleep. The OS itself is a complete clusterfuck of half-new, half-old, half-ancient things like the Control Panel. Everything prompts constantly—usually with modal windows—for absolutely pointless decisions (and no I’m not talking about UAC).

This is before I even get to what awful citizens of the platform their developers are. Everyone wants to hijack the right-click menu. For all the complaints about the Mac App Store, the Windows Store is—as best I can tell—bordering 100% crapware with no way to find anything of reasonable quality. Everything’s UI seems to have been built with entirely custom components designed by a color-blind child who’s seen too many sci-fi movies.

To top it all off, I am continually just floored whenever I read Raymond Chen’s blog. He often has posts explaining why $thing works some unexpected way or has some sharp edge, and the answer is always because thirty years ago a team at Microsoft built something insane and convoluted (admittedly through the lens of someone watching today), it’s now baked in at the lowest levels of Windows, and it can’t be touched, changed, or removed because of their commitment to backward compatibility.

I struggle to find any redeeming qualities of Windows as a user.

I get that the Mac is a poor platform for gaming development. I really do. But goddamn would I kill for that to change, and for me to be able to run games on my Mac some day.


> You can’t swap sound outputs while an application is running.

Then the application in question wasn't properly written and assumes that the output sound device never changes. About the only application running on my Windows gaming desktop that doesn't follow the sound adapter I select (headphones, integrated, USB DAC, etc) is the Java version of Minecraft.

I use MacOS, Linux and Windows and every single one of them has their own different pile of warts to deal with.


I’d be a little more circumspect before accusing others of ignorance and then going on a tirade of declarative statements about an OS you’re obviously not invested in enough to solve the problems you’re experiencing.


I’m not using “ignorant” as a pejorative here. I literally mean this in the sense of not realizing that other systems exist and are dramatically less awful. I too used to be ignorant here, until I started using Linux extensively before eventually switching to a Mac.

And yes, I understand that there are people who are forced to use Windows professionally or for gaming (like myself), there are people who use Windows because they don’t like (or want to pay for) Mac hardware, and I understand that there are even people who aren’t ignorant about other operating systems who prefer Windows. On the whole though I conclude that the overwhelming majority of users who could use a Mac and don’t simply do so out of a combination of unawareness and/or inertia.

To your second point, literally none of my complaints are something I could have any possible level of control over, no matter my level of personal investment. Windows’ nested layers of awfulness driven by slavish adherence to backward compatibility are quite literally something I could never visibly impact even if I dedicated my life to doing so. The fact that Mac apps, by and large, try to be “good citizens” of the platform while Windows is a free-for-all of developers doing anything and everything however they see fit is in a similar boat.


>> Probably the only real Apple game platform is iOS

Actually, I’d say the ‘platform’ is Apple Arcade. I imagine that’ll grow more.


The backwards compatibility for games on Linux is being improved.

We do have Flatpak[0] which could be used for indy, DRM-free games.

Valve is also working on a Steam Linux Runtime[1] which uses Linux namespaces to make older titles work on newer distributions.

[0] https://flatpak.org/ [1] https://steamcommunity.com/app/221410/discussions/0/16386755...


If Apple ever changes their mind, they could own this market though.

I have a 6 year old imac 5K. I don't game a lot but I always had a weak spot for x-plane. X-plane 11 came out shortly after I got this machine and I've been running it throughout. It has improved a lot over the years and they just released a beta with metal support. It's quite amazing that I can get decent performance with this version in 2020 that is better than what I got in 2015 with the same hardware. I run it with number of objects maxed out at a low resolution and mostly can keep the fps above 25-30, which is the minimum for an acceptable experience in a flightsim. The GPU is a still pretty decent AMD R9 M295X with 4GB ram.

However, my Steam library looks pretty depressing. As of Catalina, only a handful of games actually still work. I wasn't really playing most of them but I did spend money on this stuff at some point. I've actually been considering switching to Linux recently specifically because of Steam and having access to a bit wider selection of games. Apparently a lot of stuff works pretty well these days; certainly much more than what works on macs.

IMHO if Apple ever does a move on the gaming front, they have a good shot at domininating this market. If they can manage to mass market a good enough product, the numbers start adding up. That's why the IOS store is such a money maker. I've always been puzzled why they never made more of an effort to position the Apple TV as a gaming console. They tried, a little, but it never added up to much.

Apple making their now widely rumored move to their own CPU/GPU could unlock this. IMHO any VR/AR plans they may have are probably dependent on that. I doubt they'll make life easy for the likes of Steam to run in such an environment though. Apple has never liked middlemen.


> 1. No cross platform builds allowed. We can compile to Windows, Linux, Android, Switch, PS4, Xbox One all from one windows build machine.

I build for OSX from Linux using clang all the time. You just need to grab the sdk from xcode and get your toolchain all setup.

https://github.com/tpoechtrager/osxcross


License terms explicitly forbid using any of the SDK on non-Apple hardware, so many companies will not be comfortable with ignoring that.


What if linux is on apple hardware?


Then why not run macOS?


> Then why not run macOS?

You value controlling as much of your computing environment as possible? Want source to all the software you use?

Do we really need to explain the merits of libre software still in 2020?


I think you’re responding to a strawman rather than my comment, which was in a thread about developing for Macs being difficult on non-Apple hardware because they can’t run macOS and the “solution” being buying Apple hardware and not running macOS on it.


The ideal case isn't running OS X, it's for (Apple to allow) Windows (or Linux) to output OS X builds. Switching to OS X doesn't solve that. A quick read of Apple's license would imply that's prohibited, thus it's a clever bit of rules lawyering to suss out that Apple's license doesn't require that their software be running on bare-metal on OS X, just Apple hardware. This means it's possible to output OS X builds while not running OS X, in limited cases with the proper configuration.


On the other hand, Apple is a top dog in mobile gaming. Maybe it should not be generalized as "Apple doesn't like games".


That's more a factor of having a large install base than actively courting or targeting "gamers"


Apple actually has a gaming platform called Apple Arcade and gaming is a separate and prominent category in the App Store. They do "game of the day" and editorials all the time, there's a platform called Game Center that can be used as an account between devices and a place to keep scores.

Their graphics performance on iPhones has been top of the line for many years now, they always boost about the graphics performance of their devices they have game demos and game introductions incorporated into their product announcements and finally they introduced game controllers support in the latest OS.

I don't know what else they need to do to target gamers. RGB LEDs maybe?


Get the graphics support on their desktop platforms/OS to a usable level again? I barely play on mobile because most of the games I find there are either shallow or don't work as good with touch screens


While Apple Arcade was a nice idea, its clear now that is has been out a while they they are still not taking gaming seriously. New game releases are slow and the games are very shallow. There are very few games in the Arcade that you would play for more than a day or two.


Do you not suspect that the marketplace for "shallow" games is a lot bigger and more valuable than whatever section of the market you think they're missing?


We don't need to guess the market. Lets use App Annie to tell what games are going well in the App Store.

Top Grossing seems to me to be about 50/50 casual / core games.


You are right but it doesn't have anything to do with the abilities of the hardware or software they provide.

The medium is more approachable, therefore attracts more casual gamers and many of these casual games have stunning graphics, physics and gameplay.

There are also games that require more serious skills, games like Pubg Mobile and to be honest, I like the mobile version more than the desktop version.


So then why are you in this thread at all? Apple goes its own way, people who like interesting, mechanically deep games with cool tech go their own.


Sure we can go out own ways! But I was challenging the Apple "not taking gaming seriously" accusation.

Mobile gaming is gaming now, if you count where gamers spend their money. Apple sell the easiest, most powerful mobile platforms for developers, and theres are loads of iOS exclusives as a result.

Big budget, cutting edge graphics, fiddly controls, single player big screen adventures... Those games are a shrinking niche for developers. And in that niche, MacOS can only be a 3rd place behind consoles and PCs.


FYI, most people don't let capitalism define what something "is" (in italics, even).


tbf even Apple realizes the problem, which is probably why they worked to make iOS and MacOS more cross compatible with each other development wise


I will say that the newer MBPs have decent video cards. They’re not great, but they’re ok-ish.


Unfortunately due to the mobile architecture you can’t sync frames. This is true on the windows side too but there you can get a “gamer” laptop which has a desktop cpu and video architecture. We could only demo castar for example On a gamer laptop or a desktop. Even a cheap desktop worked great.

I’m a Mac user and develop on the Mac so you might think I’d bend over backwards to make an excuse, but...not possible in this area. Those gpus do render beautiful pictures and can do some cool gpgpu crunching, but aren’t appropriate for gaming.


Genuinely surprising. I’ve switched over from gaming on my older PC (very old CPU but a 1050ti) to my MBP for cable logistical reasons, and the MBP performs admirably. Admittedly I’m no hardcore gamer, but I’m satisfied with what my MBP can do.


>can’t sync frames

What does this mean?



Right, but display buffering is a solution to the vertical sync issue that has existed for as long as raster graphics. You can’t buy a GPU without it.

Now that I’ve had more time since I read the post, I think they may have been referring to adaptive refresh rate.


I don't think adaptive refresh rate is what that post was referring to. Instead, I think it was referring to the difficulty in doing frame synchronization when the game is running on the discrete GPU but the display is connected to the integrated GPU. But I don't know how Apple's hardware is set up on that score.


Yeah, the problem was making sure we didn't have screen tearing. T he only path to video was through the mobile integrated graphics interface even when using the external graphics chip. On a desktop CPU you can DMA into graphics memory directly and get the frame sync.

I think you have to do this if you want to be able to switch between integrated and external. Which we didn't care about but perhaps most people do?


> I think you have to do this if you want to be able to switch between integrated and external. Which we didn't care about but perhaps most people do?

Most people need to be able to power down the discrete GPU when they're done gaming in order to have reasonable battery life, and the most user-friendly way to make that possible is for every application but the video game to use the integrated GPU.


It’s not necessarily about the hardware, but (for me at least) more to do with the OpenGL hacks I have to do to get reasonable cross platform support rolling.

Took about 2 days to take a win project to linux, that same win project took about 3 weeks to port to mac.

That was the first time I’d done it of course (to either mac or linux), but I just found the mac documentation to be severely lacking.

Also all macs are now stuck on OpenGL 4.1 forever, it may not matter much now, but it will show in future.

And yes I know Vulkan is a thing but I think that’ll take a while to fully be adopted by the games community.


> Also all macs are now stuck on OpenGL 4.1 forever, it may not matter much now, but it will show in future.

Not forever: OpenGL is deprecated on macOS and will be removed in the future. I will be very surprised if OpenGL survives the ARM transition.


Imprecise language on my part. I meant that’s the final version that they’re ever going to support (if you can call the current state “support”), but yeah you’re 100% correct.


There is no Vulkan support on Mac as far as I know, except with MoltenVK...


There isn’t, you’re right. I was more future proofing my answer :)


But "decent" isn't really good enough for VR, where high refresh rates matter quite a bit.

Another factor in play might be that, if rumors are to be believed, future Macbooks will be shipping with ARM chips, with yet-to-be-determined CPU and Graphics performance. That might be a bridge too far for crossplatform projects who are already frustrated about having to maintain a Metal backend just for Mac.


Totally fair. VR has much higher requirements than regular gaming.


And yet it works and sells well on a 3 year old mobile chipset in the Oculus Quest.


Indeed, new machines can have decent GPUs. Except the install base is where your game will be played.

Thus the average matters.


The install base of your buyer matters, not the install base of Apple. I doubt that the average MacBook Air buyer is the average Mac video gamer, and I bet that MBPs are over represented in Mac video game uses.


I would bet otherwise. Most gamers are casual and casual games usually run on integrated GPUs.

Only top-end MBPs are half decent for gaming anyway.


I have a gaming PC but also used my MBP for gaming on non GPU intensive games.

I was happy with this setup until Catalina killed 32 bit support. Most of the games I played on the MBP was 32 bit.

Killed Mac Gaming for me.


Those GPUs are horribly underpowered for the screen they're driving. They would be fine on a 1920x1080 display, but on a Mac display the GPU barely squeezes stable 30fps on modern games... even in lowest settings.


The GPUs are not underpowered. A Radeon Pro is simply not intended for video games. There are less powerful standard Radeons that have better game performance.


How is the proportion of MBPs in the macOS users?


Probably high as long as you include work machines, and otherwise low.


> Video games are in development for 2-3 years, and stay on the market for 5+ years with no planned recompiles.

Do video game makers plan for updates? Am I getting this wrong, or are patches and such deployed without a game recompile?


Updates are planned, and more and more games get longer support periods. Still majority of games stop getting updates within a year.

Every game is different, but for every Subnautica with 5+ years of active development post initial sale you'll get a bunch of successful games and only marginal updates for 2 months.

For my next game, Railgrade, I am planning a couple years of updates, but even that will be dwarfed by the long tail years of continued sales after updates are done. I expect Railgrade to be on the market for 5+ years. For consoles it should be purchase-able for at least a decade until Nintendo shuts down the eShop. Then it should continue being playable for decades more.

At some point emulators will take over and my effort can keep bringing players joy for eons more. I am not interested in Apple's "expiring" model of art.


Railgrade looks very cool, look forward to checking it out.


Games eventually stop getting updates. They're proprietaty software so they're always in danger of becoming incompatible with newer systems. This is true even on Windows: older games regularly fail to work on Windows 10 in my experience and many didn't recover after hunting down the missing DirectX stuff. It's gonna be ironic if Wine becomes the best way to play these games one day.

In a way, this is the whole point of the emulation community. Proprietary software was developed for really old systems and they don't work on modern computers. The old systems had to be emulated as virtual machines on top of which the old software could run.


Maybe for the first year or two, but many many games are pretty much untouched after that, even though there is a long tail of players still buying and playing them.

And yes, many expansions and additional content are usually data-driven, so it might not require a recompile.


> 1. No cross platform builds allowed.

Mozilla cross compiles Mac (and soon Windows, I believe) Firefox builds on Linux build machines.


Someone had mentioned this previously and I was curious if you were linking against the SDK or TBD files.


As one of the few who tried to game on macOS using an eGPU; I don’t blame you. Also, Tiny Metal rules :)


IMO #3 is the only significant issue. Developer will overcome any struggles if payback worth it.


I suppose they do that so you know what using a Mac is like when you develop for it. Macs are, in my experience, different than other machines.


Games are different, though. They don’t play as part of the ecosystem; they run full-screen and do their own input processing and use no system widgets or controls. They essentially take over the lowest level of hardware they can get access to, and in many ways once launched would be indistinguishable between the various platforms (quirks and rendering idiosyncrasies aside).


How are they different?


Awesome trackpads, and cohesive hardware from a single maker vs a massive ecosystem of different arrangements of components. Different ethos imo.


What things have have Macs had in terms on non-backwards compatibility? Genuinely curious here - don’t say 32 bit, they released the last desktop Mac OS that ran on 32bit hardware more than a decade ago.

(I’m not saying that I disagree with some of the other issues)


32 bits was dropped this year, not a decade ago. That killed hundreds of macOS games.

Another major issue is broken OpenGL: bad drivers, very old version and nowadays deprecated.

In 2010 macOS gaming was in a good spot. Nowadays, Apple is making sure you don't use Macs for games.


No, 32bit hardware was dropped a decade ago. There has not been a Mac sold that isn’t capable of running 64bit software for even longer.

32bit was dropped in Catalina, the prior release a year earlier started making it clear it was going away.

Neither intel nor amd sell 32bit only desktop processors, and possibly not even laptops.

Game developers complain about performance all the time, but given simply recompiling you 64bit gives (IIRC) a 15+% perf win it continues to amaze me that devs insist on producing 32 bit only software.


32-bit hardware has never been dropped. All your Macs are capable of running 32-bit. In fact, they can even run 16-bit operating systems.

It is software support that was dropped in Catalina.

"Simply recompiling" for games is impossible for technical, legal, financial and other reasons. Some companies do not even exist.


I started developing an iOS client for an API we use at work recently. Downloaded Xcode again, built a small POC Mac app and ran it. It wouldn't run because I am still on Mojave and the default settings are to build for 10.15. I was a little shocked/not surprised to get a "This application is not compatible with your software" type dialog from my own brand new code!

I really like a lot of the ideas that Apple implements, but the upgrade treadmill is notorious and not one of them.


You can write applications that run all the way back to Mac OS X 10.6 using the latest Xcode. However, our of the box a new project is going to be configured to build for the latest macOS supported by the version of Xcode you’re using. It’s very easy to set your deployment target to an older operating system though.


Yes, you're absolutely right and this is not generally a problem (although obviously APIs get deprecated, etc). All my projects have a cycle, though, where I open them back up and have to turn several knobs in the build settings to get them to build. It's not often an issue with the code aging out, but the Xcode configuration. I got a pretty good chuckle when that happened to a brand new project right off the bat.


Metal?


How is using something like unreal engine not able to compile to all operating systems. I'm not a game developer so pardon my ignorance but can you not just run macOS on a VM from Windows to compile?

Hardware isn't horrible from Apple. It's just more expensive for macOS and the complete build quality that has been slipping over the years. All Apple users know we're paying more for the hardware than machines preinstalled with windows.

Did Steve Jobs really dislike games? I assume he might have viewed them naively as a waste of time. Most people consider that assumption true and unless you consider them some type of therapy to unwind from the world but even then it can be argued there are healthier options for the majority of people that play games.

edit: ah yes, I get downvoted for writing an honest comment because that's HN crowd.


You cannot run macOS virtualized on anything but Apple hardware. It's against the Apple Terms of Service. Is it technically possible to do? Yes. However most companies will not do so because that's ground for getting sued, or having your developer licenses revoked.

For the same reason as macOS games must be build on Apple machines, iPhone and iPad applications must also be build on Apple hardware. There is an entire market of companies that rent out "server farms" of MacMini's in datacenters for this exact reason. It massively raises the cost and complexity of a modern development/integration system to have "special snowflakes".

As it comes to hardware, the problem is you can't get something "reasonable". You can build a $500 Windows compatible PC that's not half bad. You can stick a low in Nvidia RTX card in that (for a few hundred more) and get some really reasonable gaming performance out of it. However to get an RTX level gaming performance out of a mac... you're out however much it cost to buy the new MacPro... so let's call that $10k. Most normal people are not going to pay that.


Sure, you're right about running macOS on non apple hardware but I know a lot of windows user do that and if you're a serious developer why not just buy an apple computer. I really doubt the cost of the apple computer to then compile isn't going to be paid off from macOS users if the game is any good.

Renting out a mac mini isn't raising the cost or complexity to something unreasonable either. It's like people are forcing their ideology upon a company that decided whats best for itself.

The hardware comment your wrote in regard to mine is cherry picking out the part where apple users know they're overpaying for the hardware they get.


There are quite a number of costs here relating to the "why not buy an apple computer" point. I know because I paid them in both previous jobs and as a sole developer.

Firstly, the hardware cost. It's expensive and slow, but for a business it's a justifiable expense. The real cost is in the ongoing hosting and maintenance. Keeping Mac minis or whatever in a datacentre is a costly waste of space. There's zero remote management, and any failure means a trip to visit the system along with a monitor to hook it up to. Then there's the ongoing system administration cost, keeping it updated, keeping the development tools and build dependencies current. This is an ongoing expense.

Typically, if you need to keep clean environments you need to virtualise, and this means using VMware Fusion or similar. It makes a slow Mac mini system even slower. But if you need to support multiple MacOS versions, or multiple environments, what other choice do you have? It's not like there's a containerisation system for MacOS.

Every other platform can be run virtualised on big metal, like VMware and OpenStack clusters. Linux, BSD, Windows, everything we need to care about. MacOS is a special snowflake exception. This nonconformity adds costs.

As for renting. It's bloody expensive for very little in return. That entire market segment exists solely due to how terrible the hardware and software licensing options are. And the hardware and software procurement and management pain is reflected in the uncompetitive pricing. If Apple offered a developers-only generic VM licence for MacOS this market would vanish overnight.


As a single developer, that's an incremental cost that would be hard to justify for a small potential market.

As a company, adding an additional different testing infrastructure (rent a macmini farm) with the additional pipeline costs (setup, build pipeline etc) as well as the additional opex, again, for a small potential market, is not economical.


> Did Steve Jobs really dislike games?

Dislike may be a bit strong, but John Carmack described him as indifferent:

> Over the years I've been through a number of initiatives where Apple wants to get serious about games, and we've done things with them. The idea way back with Quake 3 on there, that was my deal with Steve Jobs: if Apple adopts OpenGL rather than going off and doing QuickTime3D or something else of their own which was going to be a bad idea, then I'll personally port the Quake 3 stuff rather than working with a partner company on that. And we went through all that. All of our Apple ports have been successful - they've all made money - but it's marginal money, and we have worked with Aspyr usually on all the other ones after that, but I do think it kind of comes from the top.

> The truth is Steve Jobs doesn't care about games. This is going to be one of those things that I say something in an interview and it gets fed back to him and I'm on his shithead list for a while on that, until he needs me to do something else there. But I think that that's my general opinion. He's not a gamer. It's difficult to ask somebody to get behind something they don't really believe in. I mean obviously he believes in the music and the iTunes and that whole side of things, and the media side of things, and he gets it and he pushes it and they do wonderful things with that, but he's not a gamer. That's just the bottom line about it.

> There are people at Apple who want to support all this - and there's no roadblocks for us right now, we're going to support the Mac on Rage, we hope to get a version of Quake Live going up on the Mac there - but it's just that's not what the Mac platform's about, and I don't really expect that to change because it's a tough equation now that you've got everybody dual-booting their Macs and everything: why would you want to go to the extra trouble of [developing games for Mac]?

https://www.eurogamer.net/articles/id-softwares-john-carmack...


This always struck me as highly shortsighted of Jobs and other Apple leadership. Games helped to drive and define the Apple II and Macintosh successes, and the virtuous cycle of gamers growing up on those platforms and then wanting to try their hands at developing their own games for them is how many, many software developers got their start, including some who later found themselves at Apple. What's especially strange is that that list of former videogame developers includes Jobs and Wozniak.

Apple later lucked into a gaming market with iPhone, but they're still awkward at it (just see the games they choose to showcase onstage) and there is very little cross-pollination that considers the Mac as a gaming platform.


When the Mac was originally launched, there was an industry perception that GUIs were toys not suitable for business, so Apple was very worried that being a games platform would reinforce that image[0]. That initial worry seems to have left scar tissue in their corporate culture that would haunt Apple for decades.

[0]e.g. see the concluding paragraph on https://www.folklore.org/StoryView.py?project=Macintosh&stor...


Games helped to drive and define the Apple II and Macintosh successes

Apple II, but Macintosh? I'm not sure I agree with that.

I think a lot of Jobs' opinions about personal computers gelled in the 80's. Back then, with the various Sinclair machines, the Ataris, the C64, and the Amiga, there were a lot of people who said that games actually harmed those platforms. In North America at least, having a large library of games made businesses see those platforms as toys and game consoles. Why would you buy an office full of game consoles? You're a serious business man wearing a serious business suit, running a serious business, and you need computers from a serious business computer company that can run serious business software! You need an IBM!


> You're a serious business man wearing a serious business suit, running a serious business, and you need computers from a serious business computer company that can run serious business software! You need an IBM!

To be fair, that mindset predates the use of personal computers to do serious graphics business. The Mac is largely responsible for making that mindset obsolete with things like desktop publishing and Photoshop. The same attributes that made the Mac a good platform for those use cases also made it much more suited to gaming than DOS-era PCs.


> much more suited to gaming than DOS-era PCs.

Is that really true? I was alive during DOS-era gaming and still remember kids from Mac families standing out on the sidewalk in the rain crying like Oliver Twist weenies because they couldn’t get their game on. Meanwhile PC - read _intel_ - games we’re writing their own low-level memory managers to eek everything out of the platform possible. I was inside playing Ultima VII or playing Falcon 3.0 with a remote friend _over modem_ drinking hot chocolate. I think any review of the games catalogs would show that the market clearly judged the Mac as less well-suited for gaming.


In Europe most serious graphics business during DOS-era PCs, was done in Amiga and Atari STs.


> just run macOS on a VM

As far as I know it is against the terms of service to run Mac OS on non-Apple hardware, so you can't just run a macOS VM.

> Hardware isn't horrible from Apple. It's just more expensive for macOS.

Is that not the same thing, effectively? For a given buying power, a player that chooses mac will have objectively poorer hardware.

Yes, there are a few capable rigs for Apple systems, but they are ludicrously expensive and thus provide a minuscule install base for your game, so it's basically not worth the effort for a triple-A studio. Less demanding games might be more viable, but it's still a lot of costs in porting, build infrastructure and publishing for very little return.


When games are targeted to the Macs potential users have, they have always sold reasonably well. Back when the Mac market was far smaller, and it was much more difficult to develop a game, developers did OK.

If you start with the premise that you should just barely achieve 60fps with all effects enabled on this year’s top-of-the-line PC video card you’re artificially constraining your market. Make your game fun and make it run on as broad a set of users’ systems as possible and you have much better chances of doing well in any market.


I understand and agree with "make the game fun despite computer power", but you got to understand that at least in the current industry climate it means less games are even viable for porting at all.

Basically every triple-A studio banks on having The Latest Technology for their games, so that is a whole nonviable segment.

Sole developers or small indies often can't bother keeping up with the product lineup (XCode often requires the latest macOS version and that means buying a new machine way more often than for other platforms) just to be able to keep building their game. Remember, these are possibly multi-thousand dollar purchases, not to count the actual porting and testing costs. So, another segment made largely nonviable, at least for initial/planned releases.

This leaves us with what's informally called the "double-A" part of the industry: Studios large enough to have a bunch of cash to spare and one or two people who can work on sussing out the multitude of platform requirements and compatibility/performance issues that arise from the Mac platform.

Not only that, but if you are a "double-A" studio that works with a in-house engine, the costs might be increased still due to the required proprietary Metal graphics API, that must be integrated or at least integrating something like MoltenVK (assuming the engine is new enough to have a Vulkan based renderer already).

The only remaining viable projects are on an exception basis, and that leaves very few candidates for porting.


> When games are targeted to the Macs potential users have, they have always sold reasonably well.

The mac market has always been small, and the supply of games even smaller. So that supply has sold pretty well and for example Aspyr and Bungie made pretty OK money off of that.

If the mac supply of games was more competitive I'd wonder if those economics work anymore.


What you ask is basically impossible: most Macs sold only have an integrated GPU, while all major platforms (gaming PCs, Xbox, PS...) do have dedicated GPUs. This means the performance is not even comparable.

Targeting macOS is only feasible for simple/casual games and some eSports.


MacOS Vm's on any other platform other than Macs are disallowed by the EULA. And while you might or might not believe in EULA's, Apple's legal department certainly does and will prosecute commercial entities (for example, if you are trying to build and sell a game)

Re Steve Jobs on games, you can believe in your own skepticism, or you can hear it directly from John Carmack, the creator of DOOM, who tried to work with Steve Jobs to get a proper foundation for gaming on the Mac. He did not succeed.

https://www.digitaltrends.com/gaming/john-carmack-talks-stev...


> Did Steve Jobs really dislike games?

'[Steve] claimed to have never read a comic book in his life (“I hate them more than I hate video games,” he told me)...' — https://www.vanityfair.com/news/2019/09/bob-iger-remembers-s...


Everyone's talking about the slower hardware. IMO the biggest hurdle to gaming is Mac OS's baked in mouse acceleration and whatever else makes 100% of mice feel absolutely terrible on a Mac. Even the magic mouse feels janky.

I followed a bunch of online guides to do everything I can to disable the acceleration and everything else, but for some reason when I try to play a game it still feels terrible. Even when the game has a high framerate the mouse never feels smooth.

Tried multiple MBPs. Multiple gaming mice from multiple manufacturers, etc.

Has anyone ever had any luck making a mouse feel as smooth as it does on Windows? If so, what did you do?


Removing macOS mouse acceleration is mandatory for me to not go crazy. I'm using this script now: https://github.com/docwhat/killmouseaccel

Downside is it's a one-shot sensitivity which can't be adjusted, owing to the hacky nature of it, which was semi-found-by-chance to boot. After lots of searching on threads, seems this is one of the only hacks that works, so doesn't seem to be much support from macOS on this stuff.

Used to have an app format of this so I didn't have to run the console command every time at login, but need to find an updated solution for Catalina.


That's disappointing - I've never used a Mac but the first thing I do on new Windows PCs is turn off "enhance pointer precision" (mouse acceleration) in control panel. It's awful for anything where you have to aim, I can't imagine gaming on a system with mouse acceleration that can't be disabled.


FPS games typically use raw input, which bypasses any acceleration settings in the OS. This works on both Windows and macOS, assuming the game is properly written. This is why I insist on keeping mouse acceleration enabled since it helps with everyday computing tasks and has no effect on at least the games I normally play.

Regarding the comment above, it is definitely possible there are subtle configuration differences that cause the described behavior (different acceleration curves, polling rate/DPI settings, display refresh rates, etc). FWIW, my “gaming” mouse works fine with my MBP, but it takes a moment to readjust to the different sensitivity and acceleration behavior.


Getting rid of Mac OS's acceleration isn't simply a matter of clicking "raw input" in a game's options.

Even if you use the command line to actually disable it for some reason the mouse still feels off.

If you google, for example "counterstrike mac os raw input" or something similar, you'll see lots of posts about people having the same issue. The raw input setting doesn't work properly. Granted the posts are from a couple years ago. The last time I tried to play a game on a Mac was probably around 2018. I tried every workaround I could find on the internet and it was still terrible.

What I described has nothing to do with a difference of sensitivity or polling rate/dpi. Although if a person is unfamiliar with those things they could have an issue with them.


I don’t think you understand what raw input means. It doesn’t mean telling the OS you don’t want it to mess with your input, it means bypassing the userland OS input framework entirely and grabbing the input at a lower level preventing the default OS stack from even processing those input events.

A well-designed game would do that, anyway. I wouldn’t be surprised if the counter strike port for macOS doesn’t qualify as such.


I understand what it means. There's an option in CS and it doesn't seem to actually do anything on Mac OS. Its one of the most popular games ever created and unlike many AAA devs Valve is still actually really good at what they do. If they can't get it right on a Mac it doesn't inspire much confidence that its an acceptable gaming platform.

Can you name an online FPS with a decent sized community with proper raw input on Mac OS? I'd be willing to give it a shot and see if its different. There are plenty of games that aren't CS that people have complained about over the years.

Also, raw input or not I was talking about multiple issues. Every mouse I have ever used on the 2 MBP's I've had felt terrible and nowhere near as smooth as the same mice on a PC even for general desktop use. Even Apple's mouse feels pretty bad.

I don't really know why, their touchpads are amazingly smooth and responsive.


GeForce NOW seemed to handle raw input correctly when I tested it on macOS. Though that's more of a game streaming thing.

Regarding Apple's mice, the Magic Mouse has a rather unusual polling rate of 90 Hz (likely to save energy), which probably explains why it feels awful to use.


Magic mice are also dicey ergonomically, imho. Best to avoid in general.


Games should use raw mouse input, but there's enough that don't support it (or that don't by default) that it's easier to just leave it turned off. I also find it annoying even on the desktop - I'd rather have mouse speed relate to hand speed linearly, but I do use higher DPI than the 800ish standard of regular mice.


Weird, I use a gaming mouse on macOS and never noticed any problem or jankiness, running at 1000 Hz polling rate. Sure, the acceleration curve is different that Windows' and I only game on Boot Camp, but it's never felt "wrong" to me for day to day usage.


What's sad is that once upon a time (back when macOS was called OS X, and when iMacs used PowerPC CPUs) the mouse acceleration was great, and the cursor buttery smooth, at least so I thought.


I really like a lot about MacOS but unfortunately I didn't get a chance to ever use it until 2014 so I wasn't able to experience that.

I'd get rid of my Windows computer in an instant if it wasn't for the poor support Mac has for gaming. Windows is fine but I enjoy the Mac workflow more.


This is definitely subjective. While you find mouse acceleration on Macs miserable, I find mouse acceleration on anything _but_ Macs miserable (looking at you especially, Linux/Mutter).


I'm mostly commenting on acceleration in gaming. The acceleration alone doesn't bother me for productivity. However, for gaming it doesn't matter what platform you are on acceleration sucks. Technically a subjective opinion but its pretty close to a consensus for anyone that plays 3d action games where you control the camera and/or your crosshairs with a mouse.


I remain skeptical of the long term viability of VR; I’m old enough to remember when 3D TVs and monitors were going to change the face of both TV and video games, and yet here we are.

I priced out VR recently and I was unimpressed with the price per value available, with most systems adding $500-$1,000 to my gaming system’s cost, and some requiring that I modify my room to work properly. And that’s not even accounting for the high cost of a VR capable gaming rig, if you don’t own one. This is a pretty steep financial and logistical hill to climb given the current set of titles available.


Yeah but have you played beat saber?

I'm dead serious though the experience you get in a good immersive VR rig is completely unrivaled. Well worth the hassle and overhead. This will get better as costs come down and developers continue to innovate.


I have tried Beat Saber. I rented an Oculus Quest for a couple of weeks over Christmas. Several things, Beat Saber included, were really impressive at first, but the thrill quickly wore off. The kids quickly went back to playing their PS4 and Switch games, and never even mentioned the Quest after it was gone.

My basic take is that the technology is really impressive; I agree with proponents that the Quest finally nails it technically. However, I don't think it's enough better than other options to make it worth the significant drawbacks. Something being immersive is ultimately a function of the player experience, not the hardware. And as is demonstrated to me whenever I try to get the kids to turn off almost any game, non-VR hardware is definitely immersive.

If anything, facehugger VR seems less immersive, because the stuff is way less comfortable than other gaming hardware. The kid who loved rhythm games would play Beat Saber less than non-VR games because of neck discomfort and face irritation. Battery life was also lower than the Switch, and of course infinitely lower than the Playstation. Which also cut into immersion.

I did really appreciate the new 6-axis controllers; they were very cool, and gave a great sense of spatial freedom. But unfortunately, that freedom is a lie; my sense of immersion quickly ended every time I hit a virtual boundary. To successfully play Superhot, I found I had to continuously maintain two spatial frames of reference: VR and reality. So it ultimately felt less immersive to me than a console game, where I could just plop down on the couch and forget my body entirely.


> but the thrill quickly wore off. The kids quickly went back to playing their PS4 and Switch games, and never even mentioned the Quest after it was gone.

Same experience (Vive Pro) with niece, nephew and their friends. They tried Beat Saber and all other age-appropriate VR games. After those few times, not once did they ask to try again. They would much rather play Fortnite and The Sims, with a mouse and keyboard.


That’s fascinating. I would totally be willing to accept that I’m on the other side of a generational divide over VR, and that I’m now too old to “get” new stuff like that. But with kids these days (heh) finding low replay value in VR, my thesis might be closer to the money than I thought.


Yeah, I had similar thoughts. One really interesting thing to me was that after a while, the kids would sit down to play Beat Saber. Instead of moving arms boldly as one would with real swords, it degenerated into hand/wrist twitches. Part of the fun for me of VR was getting a real workout, but a) I want that, and b) I mostly play video games in short bursts. It makes perfect sense that people who play longer want it to be nice and easy. And if you're already on the couch, the big TV and the regular controllers are just fine.


I haven't played Beat Saber, but I remember Avatar. As beautiful as it was, it remained unique. All new films for the past decade have been released as 3D, but I cannot think of a single one other than Avatar that I would prefer watching in 3D over 2D. Is there reason to believe Beat Saber would not become VR's Avatar?


Avatar did not wow me in 3d. Beat Saber and the current VR games like Alyx and Pistol Whip made me a believer in VR.


A believr. (sorry)


You're comparing a fixed medium, movies, with a dynamic medium, games. I'm not sure VR will be a long term success, but it will have no relation to 3D TVs failure.


What is fascinating that both of these technologies require glasses - something burdensome and often uncomfortable.

I often said ‘when 3D TV’s don’t require glasses, I’ll buy one.’ I guess that was the general opinion, and when that didn’t happen quickly enough, it died.


Since everyone's really going on about Beat Saber, I have to say out of all the games I've played on my Vive, the one that I had the most fun with and kept coming back to was SUPERHOT.

The entire concept of the game (time only progresses while you're moving) fits _so well_ with VR. It makes you feel like a complete badass as you're able to perform ninja-like feats of dexterity and dodge bullets all the while being a just challenging-enough puzzle and providing a decent workout.

Everything else I've played was either basically a tech demo or felt like our old games built within the limitations of VR (what really differentiates Beat Saber from DDR, Guitar Hero, etc?). SUPERHOT was one of the few games I played that really felt like a proper VR game.


I will proselytize Beat Saber until the day I die. I have played hundreds of hours of video games in 2020, which reflects the time spent since I was 8 playing Donkey Kong on the Super Nintendo. After writing this, I'm probably going to jump into some FF7R. In the morning, I'm going to log into Animal Crossing and do some chores around the island.

Beat Saber is the most fun game I have ever played. No qualifiers. No shifty language. Nothing else is even in the same order of magnitude as Beat Saber. I dare to even call it a video game, because very few other games even possess the language to compete with it. The combat in FF7R is visceral, but you're not actually swinging Cloud's sword. The loop in Animal Crossing is entrancing, but it doesn't become your entire world like when you put a headset on. Even other VR games, which have that language they can use to tell their gameplay, often feel like the VR experience is bolted onto something which could, and maybe should, be non-VR. You often fight with the game; the game transplants you into a world, not just on a screen, but every degree surrounding you, yet you're limited in how you move, you're limited in how the game allows you to interact, nothing feels natural, and you soon adapt your actions with what the game wants.

I am solidly in the camp that, honestly, VR will not become a household item like a game console is. Even with headsets priced about at the level of a game console, like the Rift, the platform has too many downfalls.

Startlingly few games work well in VR. Even an experience like Half Life Alyx, the pinnacle of VR technology, feels like a shell of a game like Half Life 2. Its different. It can't tell a strong story, because players don't generally spend more than 30-90 minutes in VR at a time. It can't really innovate in puzzle design, because movement and viewport control is so heavy that it becomes tiring if you ask the player to do too much. The mechanics feel incredible. No one enters HL:A for the first time, loads that pistol, and doesn't have this HUGE GRIN on their face. But, its also a novelty; it wears off far too quickly once you begin realizing how many fundamentally core things Valve had to sacrifice to make the game work at all.

Nearly every game in VR is, bluntly, a tech demo. The platform is still figuring out if it is viable. Of the top selling VR games on Steam, #7 is a sim game which models a hyper-realistic anime girl, allowing you to mess with her clothes in a predictable anime way, and #10 is Skyrim. Of course, HL:A and Beat Saber are #1 and #2, but roughly this state is where VR was six years ago. It hasn't really evolved, and I'm not convinced HL:A was the catalyst I think Valve is hoping it is.

But, maybe I'm wrong. I will say this: There is no price you can pay for a VR headset that does not make Beat Saber worth it (within your means, of course). It is that good. I have an Index ($1000); beyond spending an hour or two in as many games as I can, 100% of my time is spent in Beat Saber, and I do not regret the purchase. Everything else VR has to offer is just icing. But this, a platform, does not make.


Games that do a good job of keeping the arcade format are a lot more immersive. Fact is, VR today can only put you in a single room. So the entire game needs to be build around that constraint to maintain immersion.

Space Pirate Trainer is another game that did a really good job of this. Before Beat Saber, that was easily my favorite, though wow Beat Saber is _so good_. Blarp is another game that did a great job of immersing you into a single room and making the most of the mechanics.

More "full" games don't work as well, because they have to figure out some way to transport you around the world without breaking immersion or causing nausea. And I don't think anyone has pulled it off well.

I am confident we will have more games that stand up to Beat Saber. It's a platform that requires discarding a lot of the strengths that video games have in order to work well, and developers have to re-learn a lot more from scratch than I think anyone realized when the platforms started.


> So the entire game needs to be build around that constraint to maintain immersion

I haven't tried the new Half Life, so I'm wondering if you're right.


HL:A definitely does more with something resembling an adventure formula akin to the traditional HL games, versus most other VR titles. Per my original comment; I think I have to consider it as experimenting, and it's only partially successful.

(low-level spoilers)

There's one moment early on where you're asked to solve a vorgon puzzle in a drainage ditch. Every person I've asked who has been through this area disliked it, and I haven't found anyone who solved it legitimately. Its confusing. It has parallax/perspective elements which ask you to move around a large area, position yourself perfectly, and possibly even crane your neck in the right way to get an angle on what is the intended solution.

But, movement sucks in any non-room-style VR game. You can't use the joystick to just move your character, like you would in a non-VR game, because it causes unreal motion sickness in, like, 100% of users (this is available as option. its an experience that i think everyone should try; I had no idea I could get motion sickness until I did, but I was literally falling over). Instead, you press the joystick in the direction you want, a targeting beam pops up, and you "teleport" to the desired location. This sucks. Its imprecise, its slow, and it makes you feel, to some degree, like you're actually watching a powerpoint presentation of the game.

And, head yaw can also suck in VR. This may be surprising; its the entire point of VR. But, depending on whether your headset has a tether, and how big your play area is, many VR players are stuck stationary and with, maybe, 180 degrees of freedom in front of them. In one of these setups, you cant naturally look behind you; you have to rotate in-game, which calls back to all of the negatives that movement itself has. But even the in-game rotation is in-precise; in HL:A, it jumps by roughly 60 degree increments. Imagine there's something 45 degrees to your left you want to center; you can either turn your head 45 degrees left to center it (uncomfortable), rotate your body 45 degrees left (scary), or rotate in-game 60 degrees left, followed by rotating your head 15 degrees right or body 15 degrees right. In small play areas, the body rotations are a slippery slope that eventually ends in you being turned around, contacting furniture, or getting tangled in the headset tether. This is, at least for me, always front of mind; I can't escape the reality that my play area isn't large. So, you end up with these nearly (only nearly) subconscious calculations involving how many times you need to rotate in-game to make neck rotations less uncomfortable.

This is the problem space that any non-room VR game has, and will have, to solve. It is unsolved. I believe it is unsolvable. Even if you can get the player/object interactions down to perfection (this is insanely difficult, and different for every game; HL:A is the first "complex" VR game I've seen get close), even if you can craft a story and world that is both compelling, and works well with the intrinsic movement limitations VR has (no one has done this yet; again, HL:A gets close), even if you can remove slideshow-teleportation movement without motion sickness (I believe this is impossible without something like an omni-directional treadmill); you still run up against most players just not having the space for room-scale VR & N-degree freedom of movement.

VR is filled with technical challenges, and every technical challenge developers have to solve limits their ability to be artists. This is why arcade-style games are the showcase piece for VR. And, in my opinion, will remain that way for a good, long while.


> And, head yaw can also suck in VR.

Aren't tethered headsets a thing of the past now? I have a Quest and never had that issue.


Absolutely not. The most advanced headsets on the market require a tether. And, to make matters worse, outside-in lighthouse-style tracking. Valve Index, Vive Cosmos Elite, even the Rift S. The Quest is literally the only headset capable of operating fully untethered; its possible to set up wireless tethers on most headsets, its just more expensive, and it still requires a PC.

The Quest does two things that are interesting, and rather unique from a full market perspective; local rendering, and inside-out tracking.

Local rendering is probably the future for VR platforms. Unfortunately, there's really only one company on the planet with a mobile SoC design capable of driving the performance necessary for truly current-generation VR experiences, and that company is not interested in sharing their technology (Apple). The Quest, with its android smartphone-class specs, is not, in my mind, a current-gen VR headset from a performance perspective. It has other (huge!) advantages which make it an awesome purchase (generally, the best purchase for most people), but performance is not one of them.

I think the jury is still out on inside-out tracking. It may approach a level where it is good enough for most use-cases, while still being worse than outside-in tracking is today. Arguably, its there today. Fundamentally, if your environment is dark, or lacks the many visual intelligence landmarks necessary for the AI to operate, its going to struggle, not to even mention the challenges it has in tracking hand positions behind your back (which is very common in FPS-type games, where you often reach into your backpack to grab more ammo).

Plus, the local rendering, the inside-out tracking, and internal battery, all add weight to the headset, which already cheaps out on the circumcranial support straps to properly distribute weight (many people swap their Quest straps with, say, those from a Rift S or even an Index/Vive, because they're so much better). The Quest is both one of the most and least advanced headsets on the market.


> The most advanced headsets on the market require a tether

I've heard streaming to the Oculus Quest is basically like owning a Rift. I haven't tried it myself though.


Obviously I have not tried beat saber. If it’s that amazing, I wouldn’t be posting cynical takes on VR.

I think one of the problems of VR is one of cost and credibility. Even if I knew you personally and trusted you, there is no way in hell I’m going to float thousands of dollars getting a new gaming PC, VR headset, and customizing my room for the full setup just because someone said it’s amazing. The risk/reward ratio for that is just insane, and not in a good way.

Heck, even $500 for an entry level setup is a pretty dicey proposition, assuming my PC can lift it. $500 can buy me a lot of non-VR gaming satisfaction with effectively 0 risk.

If it was $200, I’d consider it.


> If it was $200, I’d consider it.

Wait 10 years and then ask.

It's brand new, genuinely computationally intensive, tech - it's going to be expensive. The Original PlayStation cost more in today's money than I bought my Vive for.

Also, if you can get a used Vive or similar there is almost no risk. If you genuinely don't like VR then maybe its not for you but I've played a lot different games and I can't think of anything that remotely compares with VR. Super Hot with a mouse is brilliant, Super Hot in VR makes me genuinely excited that I'm going to be around to see the near future. Beat Saber is very good fun, too (quite hard to describe but even in a vive the graphics are very crisp and smooth).

> If it was $200, I’d consider it.

Apparently half life alyx is enjoyable in a windows mixed reality headset. There is also the Oculus Go at that price point although I've never touched one because I have a proper VR headset.


Not 10 years.. 2 years at most. Budget headsets right now are as good as the Vive. If we have to wait 10 years it'll be dead.


I'll say for sims, (both flight and racing) the immersion and ability to move your view hands-free make VR well worth it. At this point I can't imagine going back to regular monitors for DCS. I do agree that there is significant pain getting into them though, as there's no real <$100 option to get your feet wet. I was lucky enough to have a local friend who had already purchased a headset, allowing me to get a full demo.


It's funny that you're making a really strong assumption about something that you've never even tried. You can get $2xx.xx Microsoft backed WMR VR headsets new. Facebook's all inclusive Oculus Quest is $400 new.

ymmv but I've been able to buy used VR systems like the Vive Pro for $375 on ebay. Just to use them for either driving or flying simulators, VR is already worth it.

That said, there's still plenty to complain about for VR but that doesn't mean it doesn't have mass market potential.


You’ve got it backwards. One must make a strong argument for spending hundreds on something with uncertain value, not make an argument against spending the money. Basic prudence says that items of uncertain value should be weighted less strongly, and receive fewer resources.


I spent like $2000 on a VR setup so I could play Half Life Alyx without ever trying VR before. Remember that the gaming PC is very usable without the VR headset! I built a backpack portable mini-itx build with a used NVIDIA 1080Ti. Then 6 months later I got a Valve Index. Didn’t change my room at all other than moving my coffee table when I want to play. My current game is Audica which is such a blast.

But yeah. It was like $3000 in total or something ‘~’


Wow, that’s quite a large amount of money to spend for one game.

I find $60 for AAA titles a marginal proposition.


When the Oculus Rift and its development kits first released, early adopters would post invitations to interested people in their area to try them out. Many shopping malls and arcades now have VR demos or games on display.

There are options available to try high-end VR before you buy, and ~$200 low-end versions if you prefer going in blind.


I tried an oculus. It was ok, but I walked away utterly unconvinced of the value proposition.


My house is open, any friend who wants to give beat saber (or half life, or space pirate trainer, etc.) a spin is welcome. No need to spend $1000 without getting a proper demo.

I know not everyone has a friend with a high spec rig, but that is improving over time as well.


You can also just rent. I rented one here: https://www.lensrentals.com/rent/oculus-quest-all-in-one-vr-...

For a try-before-you-by thing, I'd recommend doing 2 weeks. For the first few days, it was, "OMG, the future!" Then for a few days it was, "Not overwhelming, but still cool!" By the end of the 2 weeks, though, we were happy enough to return it.


I literally don’t know a single person who owns a low-end VR rig, let alone a high spec one. That probably has something to do with my skepticism.


The Oculus quest is 400$ and it's pretty amazing.


I never heard anyone really passionately speak about the 3D experience. It was always marketing speech with movie-goers sometimes buying into the expectation. But after experiencing it, no one ever was really passionate about it. At most they would slightly prefer, and even so I think it was because the confounding factor that 3D sessions are also played in the largest screens and the best sound system.

But for VR? It is a niche, for sure, but I know a few people really enjoy it.

It can be a niche forever (like high end sound equipment) or it can grow into mainstream (like TV), I have no idea, but I doubt it will fade away.


Unfortunately, I don't think it can be a niche forever, assuming it's a gaming platform. As we see from game developers in this thread, a platform with a too-small audience is uneconomical. That's true even when the platforms are as similar as Windows and Mac. It'll be way more true given that VR is a very different experience.

Apparently Quest sales figures aren't public, but I see estimates that they sold a million units last year. Nintendo sold that many Switches last month. And the month before that. And the month before that. (And in December it was something like 5 million/month.)

However much niche users enjoy it, I have a hard time seeing how they'll sustain enough of a developer ecosystem to keep them entertained. High end sound equipment can draft off of every music player in the world for content. But good VR is very much VR-specific.


Dude, Half Life just came out on VR. It's not a niche anymore.


That doesn't prove it's not a niche, dude.

Valve, like a lot of other companies, hopes that VR will become a thing, so they're making a headset. Because they'd like to sell headsets, they took a major property and made a VR version. If they lose money on their VR version of Half Life, they're fine with that.

Some game makers will surely try VR versions as well. This could be a big wave, and they want to be ready, so it's ok to lose money. But if unit sales aren't there, they're not going to keep doing it. In which case, things could well end up just like every other attempt at VR going back to Nintendo's 1995 effort. Or like the various waves of 3D movies and TV, where everybody's really excited for a hot minute and then people stop caring.


You can try to reason all you want, the fact is that today there is a Half Life game that is only playable on VR. It's a significant fact.


I am not denying the fact. I'm saying it doesn't prove that VR has gone mainstream. It's a valiant attempt on Valve's part to make that happen, but it's just an attempt.


We’ll see in a year.


That would be great, but I doubt it. I've been having this discussion since 1997 or so. VR will always have true believers who believe that this is the year. A year later, they just have a different reason. It's the same deal with Bitcoin advocates: https://twitter.com/williampietri/status/1071833726294749184


> I've been having this discussion since 1997

I tried VR back then, and I tried VR now. It sounds like you haven't tried it now? I've seen people crying because they couldn't believe how crazy it was. Just give it a try!


Ah yes, the assumption of ignorance and the belief that if I only try it I'll become a convert. Two more tiresome features of quasi-religious discussions.

I have tried it a number of times, including the current generation. https://news.ycombinator.com/item?id=23048814


what games have you tried? I guess trying a shitty experience does not help


‘Doom just came out for the Atari Jaguar. It’s gonna be huge!’


From the perspective of everyone not in the niche, the difference between it fading away and being a niche is minimal.


Let me say in no uncertain terms that there will be a time soon when desk monitors and tvs will be present but not the main display, and their sales will drop at that point. To me, it's more "when" than "if." To reach that tipping point two things need to first occur: there needs to be an ~300 USD headset with specs comparable to the current Valve Index, there need to be improvements to hand & finger tracking so existing workflows begin translating directly into the virtual environment. When those two things transpire, give it like 3-5 years tops, and I'll think you'll see a state change.


Have you tried the Quest? It's a gamechanger in my opinion for ease-of-use. Hardware isn't top of the line, but it's good enough for a great experience, and they are doing all sorts of fancy things with the existing headset via software that is wringing more mileage out of it (Oculus Link, Virtual Desktop, Hand Tracking, pass-through, etc).

Quite bullish that the next generation will be absolutely incredible. I bought an upgraded dev machine to try out VR dev because of the Quest -- it seems like it will be the next mass-computing interface.


I've been owning a VR headset (a htc vive) for more than a year, now, and I concur with other comments : it's something else, and I think the only reason it's not taking over is the high price tag for a setup.

That being said, I use my VR headset probably even less than my 3d printer. As mindblowing as it is when I put the headset on (and truly, it never gets old), I realized it's just not for me.

I usually play games on my sofa, with a steam controller and the screen on my wall through a projector. I always play turn by turn or low intensity games rather than fast-paced action games. I have my laptop on front of me on my coffee table, and constantly check my mailbox, rss feeds and many sources of news, on the laptop between playing two turns in the games or message with friends on my phone.

This is a funny thing, because I didn't realized that before I bought a VR headset : basically, I'm always multitasking during my leisure time, and you just can't do that in VR. Each time I think "hey, let's chill in No Man's Sky", my immediate reaction is "oh wait, but if I do that, I can do _only_ that", and I play a classic game instead.

Now, this is a very personal experience, maybe that won't be a problem for anyone else. But as far as I'm concerned, VR fullfills all its promises, and yet it's not for me.

EDIT : fun fact, the time when VR is really useful for me is during week-end during the day when I don't feel like working on a side project or learning a new thing. I can't use the projector with daylight, so in those cases I'll just put the VR headset on rather than closing shutters.


I think VR is for everyone, like any medium, it's just that you haven't found a game/app that makes you come back.

For me it's the multiplayer games that I like, more of my friends are getting VR headsets right now and I suspect that as more of them get into it the more we'll end up playing VR together.


Oh no, don't get me wrong, I'm totally mind blown by what I've seen. No Man's Sky and Skyrim VR most notably were just game changing compared to what I experienced in flat screens.

But yeah, VR requires my full and undivided attention, and this will just not happen :)


Have you tried multiplayer games in VR though?

It's the same for me and gaming in general, I have very little patience except if I'm playing warzone or rocketleague with friends. It's just not the same :)

I wrote about my mind-blowing experience playing the Catan board game in VR: https://p1x3l.com/story/239/social-virtual-reality-and-the-o...


Not every medium is for everyone; human preferences don’t work that way.


If you see VR as another medium like a laptop or a smart phone, then I believe they are for everyone.

People underestimate the amount of applications for VR. But Mark my word, in a few years your comment will have poorly aged as every household will have a VR headset the same way every household had a printer or a desktop PC 20 years ago.


The peak of 3D TVs was like 5-7 years ago, so you hardly need to be "old enough" to have experienced that. Pricing is the easiest hurdle for a new technology to get over. As long as there is a market fit and steady demand, VR will get affordable.


Come back after you played HL: Alyx, Beat Saber, and No Man's Sky VR. Any one of these experiences is proof that VR is a huge deal.


I’m not floating that much money to try a game, and I’m unaware of anyone in my social circles who owns any level of headset.

“Go spend a few hundred to prove the value” is a hard sell.


So you haven't tried VR?


I tried it years ago, and I thought it was a cute tech demo with zero replay value.


That was my experience a few years ago as well. Things have changed a lot since then.


> I’m old enough to remember when 3D TVs and monitors were going to change the face of both TV and video games

but 3D TV and movies were pretty bad... It was always a gimmick. The experience you get in VR is nothing like that.


The Oculus Quest is 400$, an amazing technology, and constantly out of stock. I'm pretty optimistic for VR.


Apple feels like it's about as blind to where graphics and gaming are going as Microsoft was to mobile. in the end it's probably for the best though


There's a small market, little developer support, and they don't support NVidia graphics. Not worth doing. Good for Valve.


I don't understand why a Mac is such a bad platform for gaming, what is the technical challenge here? Can someone enlighten me?


As the article alludes, lack of a build your own, flexible yet affordable desktop option seems to be a pretty deadly aspect.

Pretty much Mac owners are on laptops. Even in the PC world few laptops can power VR. Those that do tend to be bulky for cooling and have nVidia GPUs; both things which MacBooks seem unwilling to have.

Having a extensible desktop at 1k that supports nVidia would probably go a long way.


lack of a build your own, flexible yet affordable desktop option

From Apple's perspective, that's a machine with razor-thin margins. Why would they want to sell that when all of their other products have enviable margins?

Moreover, such a machine would cannibalize their high-end Mac Pro sales. After all the money they spent to develop it, having your pro customers scoop up the thin margin gaming machines seems like a poor strategy.


I'm unconvinced that it would cannibalise Mac Pro sales. The Mac Pro is so unnecessarily expensive that it has to exclude all but the richest companies and individuals. It's for people who have money to blow on £600 novelty chrome wheels as an afterthought. Apple have clearly noticed that there's a market for people with more money than sense, but it's a small niche.

The market for mid- to high-end desktops is indeed lower margin, but it's a much bigger market overall. The old Mac Pro was at the high end of this market. Expensive, but justifiable if you needed it. If Apple had the will, they could be profitable in this space. They have chosen to exclude themselves from it.

Back in the mid '00s, I saw Mac Pros in use in scientific research settings, because they allowed use of lots of RAM compared to equivalent PC mainboards of the time, could take lots of storage, and had lots of horsepower to throw at bioinformatics and modelling problems. We had compute clusters as well, but for some problems these were a better fit. Today, Macs are banned. The academic funding bodies don't consider them good value for money. And they are not wrong in that assessment.


That doesn't convince me, I use to run windows with bootcamp on a macbook pro and play there. It worked.


As a gamer with a macbook pro. I just use shadow to play games over the cloud. Otherwise I would invest in a eGPU. So it's not like apple gamers are missing out if they want to game with good graphics.


> Otherwise I would invest in a eGPU. So it's not like apple gamers are missing out if they want to game with good graphics.

You're not going to get very far without drivers an you'll be stuck on OpenGL 4.1 and Metal. That and you'll have the problem that there are less games built for MacOS than Windows an Linux + Proton.

Investing in a GPU for MacOS for gaming would be a huge waste of money IMO.


People can bootcamp if they want to run windows with a good gpu on their laptop. You're also missing the cloud gaming part.


I'm not sold on cloud gaming yet, I gave pubg on stadia a whirl today (on my google fiber) and it was somewhat laggy and blurry compared to even my not-highend rx580 which only cost me about what a year or so of stadia would (not including the games)


What? You'll run an unsupported card over 4 slow PCIe lanes over thunderbolt in some third party box? That's a terrible experience.


See other comment and it hasn't been a bad experience for me. Have you personally tried it? I used to have a windows desktop just for gaming and before I decided to limit what I own.


Having to purchase overpriced box on top of already costly GPU and wiring all up including one power adapter each for laptop and eGPU is already a bad experience.

Already a terrible experience.


At least in theory (going off marketing materials¹ for one such device, though I personally lack first-hand experience here, and said device doesn't advertise itself as Mac-compatible), the eGPU should function as the power source for the laptop, so you'd only need one power adapter (for the eGPU).

Really, this experience actually sounds pretty pleasant.

----

¹: https://www.asus.com/Graphics-Cards-Accessories/ROG-XG-STATI...


The main reason is that Apple's pushing their own graphics API (Metal) rather than supporting the cross-platform Vulkan API. Apple also doesn't make any machines targeted for gaming use, and they haven't used NVidia graphics chips for quite a while and don't ship drivers for recent NVidia chips, so even Hackintosh users have trouble building a great Mac gaming system.


TBF, Apple’s Metal was developed before Bulkan while tackling the same problem, and for the NVidia chips, the reason on the bad relationship of NVidia/Apple is mostly on NVidia, right?


It doesn't matter if Metal was first, or better.

The problem is that it's unique and different. The problem is that I (as an open source developer for a popular gaming-related project with an OSX port) can't realistically develop and support a metal backend without having to buy an expensive mac with a good modern GPU.

And with OpenGL being deprecated on Mac, with old and (sometimes) broken drivers that are missing 10 year old features, along with apple breaking things and adding new signing/notarising requirements, it's hard to justify continuing to support mac.

Same kind of thing goes for Nvidia/Apple. It doesn't matter who's fault it is for starting that feud, it's still hurting the users.


I’ve mentioned it in a sibling comment, but I’ll just copy-paste my text:

> The problem space Vulkan and Metal solves is not high level cross platform graphics (which OpenGL tries to solve), it’s low level control over the GPU. That’s the problem they’re trying to solve, it’s not about cross platform. You might ask, ‘Then why is OpenGL deprecated?’. The answer is that Apple is just deprecating it’s own support since there’s a more low level API, OpenGL should be implemented as a library on top of Metal, which (for example) libraries like MoltenGL is.


Vulkan is the vendor-independent derivative of AMD's Mantle, which pre-dates Apple's Metal. There were several proposed new 3D APIs around that time, but ISVs quickly converged around Vulkan. Apple's reticence to adopt or even permit it is worse than Microsoft's continued efforts to push Direct3D.


The team that developed Mantle got ignored by AMD management, who then left and subsequently got hired by Apple to develop Metal. Vulkan's release was held up for over a year by AMD management and red tape, which was by then already a year later than Metal's release.


None of that matters one iota.

I'm not going to spend valuable time and resources on a non-standard graphics API for a platform with anaemic graphics hardware. Instead, I'm going to target Windows and Linux with Vulkan or OpenGL. If I do support MacOS, it will be via MoltenVK, and it will be a second tier platform.

From my point of view, by dropping standard OpenGL they have turned their platform into a second-class citizen.


Then you will only target Linux.

OpenGL and Vulkan only happen to work on Windows via the ICD driver model introduced in OpenGL 1.1, which Microsoft hasn't yet bothered to remove.

Instead they aren't supporting it on UWP and Win32 sandbox model, which only allow DirectX.

Microsoft's way forward is also to build OpenGL wrappers on top of DirectX like ANGLE.

https://github.com/microsoft/angle/wiki

https://devblogs.microsoft.com/directx/in-the-works-opencl-a...

And then there are the game consoles.

Khronos has just messed up providing SDKs that are at the same productivity level as commercial APIs.


So long as the API works, I really don't care about the implementation details.

I also don't care about Microsoft's UWP and sandboxing stuff.

They might have a desire to take the platform in that direction. But I have no obligation to follow, and if they want to retain application developers they won't be able to simply cut us off. We've always got the option of abandoning Windows entirely if they make it too inconvenient and troublesome to develop for. It's not like we don't have a precedent for abandoning MacOS X as a top tier platform for exactly those reasons.


Yes they will, Windows 10X only allows for sandboxed software.

"How Windows 10X runs UWP and Win32 apps"

https://www.youtube.com/watch?v=ztrmrIlgbIc&list=PLWZJrkeLOr...

MSIX

https://docs.microsoft.com/en-us/windows/msix/overview

> Apps that are packaged using MSIX run in a lightweight app container.

Yes, Win32 can still choose between both models in current versions, depending on Windows 10X reception, I bet it might eventually change.

Sandboxing resistant developers are always free to keep their Windows XP or 7 copy, I guess.


It's way to early for 10X to be influencing anyone's plans. It's still in the middle of transitioning from a pile of Microsoft's pipe dreams to a product that may or may not bear much resemblance to those original ideas.


Maybe, lets see.


Metal's first release was 2014, one year after Mantle in 2013. Vulkan's first release wasn't until 2016, and Mantle was proprietary.


Does it really solve the same problem? Can I run metal API on my windows or linux machines?

Also, to answer your question: no. Nvidia is in the GPU selling business. They wouldn’t mind having their high priced GPUs in high priced laptops. Apple keeps control of what hardware they sell, which dictates what software gets developed. Why would Nvidia or Apple spend resources to make software for a computer that no one is selling? It’s Apple who refuses to buy rather than Nvidia being unwilling to sell.


The problem space Vulkan and Metal solves is not high level cross platform graphics (which OpenGL tries to solve), it’s low level control over the GPU.

That’s the problem they’re trying to solve, it’s not about cross platform.

You might ask, ‘Then why is OpenGL deprecated?’. The answer is that Apple is just deprecating it’s own support since there’s a more low level API, OpenGL should be implemented on top of Metal, which (for example) MoltenGL is.


MoltenVK is an exciting prospect, but until I see games running it I will be skeptical that it exists.


Can you run direct3d on your Linux and Mac systems?


I can run OpenGL and Vulkan on my Windows system. MacOS is the only platform that has officially deprecated OpenGL support, and there is no plan for any future Vulkan support.


Windows only has support on legacy Win32 thanks to the ICD driver model. Vulkan and OpenGL aren't supported on Win32 and UWP sandboxes.

Game consoles never had OpenGL support, other than a timid attempt from Sony with GL ES 1.0 using Cg as shading language.



We’re talking about Vulkan. Why are you asking about D3D?


There are Vulkan and OpenGL -> metal libraries. The original post gave windows (where the better performing native api is direct3d) As an example.


>TBF, Apple’s Metal was developed before Bulkan while tackling the same problem


The point is that Windows has an alternative to their proprietary APIs, but mac doesn't


Metal is one hurdle, Apple has a weird habit of putting many small ones up in front of you for whatever reason it gets old after awhile and the sales aren't there to justify jumping over them most mac users who want to game don't seem to have issues just buying a console either so it's not like a large group of customers isn't being reached

iOS is an exception due to it's position in the market but I've been seeing less and less interest in mobile overtime too


Since after High Sierra (Mojave I think?), Apple eliminated the ability to use non-metal drivers on macOS wholesale. Nvidia’s drivers weren’t developed in Metal and there’s zero reason for them to do so (only way to use Nvidia is intentionally either as an addon to a Mac Pro or as an eGPU, which they don’t see a large market for it) and Apple has a long history of killing off support of something on their platforms, users be damned.


> only way to use Nvidia is intentionally either as an addon to a Mac Pro or as an eGPU, which they don’t see a large market for it

Not quite. From Nvidia: "Apple fully control drivers for Mac OS. But if Apple allows, our engineers are ready and eager to help Apple deliver great drivers for Mac OS 10.14 (Mojave)...Unfortunately, Nvidia currently cannot release a driver unless it is approved by Apple." — https://en.wikipedia.org/wiki/Nvidia#Apple/Nvidia_web_driver...

Apple is preventing Nvidia from releasing Mac Pro and eGPU driver support.


Which means if Apple allows them to use non-metal drivers. Historically, there’s little reason to think Apple would waiver (wish they would) and the value for Nvidia to write their drivers using Metal is minimal.

I’m kind of curious though: what value do their web drivers have to them now that they couldn’t just release them as OSS and allow a determined group to rewrite them as Metal.


No, you've misunderstood again.

Nvidia's eGPU web drivers already supported Metal in High Sierra [0]. Apple refuses to sign their drivers for Mojave and up. Without Apple's signature, GPU drivers will not run on macOS, OSS or not.

[0] https://www.insanelymac.com/forum/uploads/monthly_2020_03/84...


Ah, that's interesting.

I know there are some older Nvidia cards that do work on Mac, but most of the recent GTX/RTX/etc. ones don't. Are they similarly Metal (and just missing a signature)?

Because whenever I was actually keeping up with this, the stalemate was over Nvidia not leveraging Metal.


I just upgraded to Catalina and now almost none of my steam games work. Turns out they were compiled for 32 bit support, and 32bit support was removed in Catalina, basically destroying my gaming rig.


Game developers who insist upon working a certain way and insist that they have to target a certain kind of high baseline hardware say this.

Game developers who want to reach a broad market and are willing to meet their deployment platforms halfway do not say this. They just port and ship.


Mac tends to fail meeting minimum graphics requirements. There’s no practical way around it.


Sure there is: Design your game for the systems the majority of the users who will want to play it—after you’ve marketed it to them to them—will have.

Don’t start with system requirements first and make a game targeting that because you want to show everyone how elite a hacker you are.

If Sony didn’t require games achieve 60FPS at 1080p on PlayStation 4 before they’d allow them on the platform, most of the developers would be targeting PS 4 Pro because they want to show off. I expect that’s mostly the case anyway, and these developers have to be dragged by their corporate masters kicking and screaming back to baseline PS 4 support.


Yeah well but then your game looks like garbage for most of the people who will proceed to buy some competing product that's more up to their standards. If you want to make a game and sell it only to the low end mac audience go for it, just don't expect to get your investment back. I don't expect game devs to target my 2011 thinkpad in the 2020 gaming market, that'd be ridiculous.


Apparently plenty of game studios seem to happy to target the lame iGPUs from Intel on PC side.


Danieru explained pretty well in a sibling comment: https://news.ycombinator.com/item?id=23048633


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: