Safari (especially on iOS) appears to have become the new Internet Explorer, holding back progress for everyone. The table in the article certainly suggests it. At least on Windows years ago you could install another browser. You can install an alternative on MacOS, but not really on iOS (the "alternatives" on iOS can't replace the important parts, so all other iOS browsers are actually Safari). I don't see any likelihood of getting that fixed. Is there a plausible solution in the future?
The only solution I see is the FTC stepping in. I don't understand why they're sitting on their hands. What Microsoft did in the 90s was pale by comparison—they just bundled some extra software. By contrast Apple today is locking down their platform and explicitly decreeing that nobody is permitted to compete with them. It's amazing how brazen it is.
What Microsoft did in the 90s was pale by comparison—they just bundled some extra software.
Microsoft threatened to cancel the Windows license of several PC manufacturers including HP for making Netscape Navigator their default web browser instead of Internet Explorer. This would essentially put those companies out of business, so they had to comply.
Not to mention, for corporate and home users, Netscape Navigator was commercial product that customers had to purchase while Microsoft bundled IE with Windows for free. This was back in the day when most people didn’t have broadband and stores sold shrink-wrapped boxes of software containing floppy disks or CDs.
I’ve written this many times here on HN: a monopoly isn’t illegal; what is illegal is using your monopoly in one market (PC operating systems) to disadvantage competitors in another market—web browsers.
Apple doesn’t have a monopoly in either PCs or smartphones; there are plenty of alternatives to iOS and macOS. There are lots of examples of platforms where the platform owner decides what can run on that platform, starting with game consoles.
There are other requirements on effective monopolies. But the US executive branch has decided not to bother enforcing most of them. Used to be, using monopoly power to drive other companies out of business or out of a market was an abuse of that power. Now all the USDoJ will enforce is charging excessive prices.
Reiterating, being a monopoly in the US is not illegal, but it having chosen to have a monopoly makes you subject to extra fairness laws. These laws should be enforced as written, but no recent administration has been willing to step up, and Americans are the poorer for it.
If Google forced alternative browser engines off the play store, the EU would jump all over them.
Apple gets to break all the rules Microsoft and Google get dinged for, because they are vertically integrated. It feels like a bug in the law that corporations are incentivized to take everything in house. Somehow being a bigger, more powerful company makes you less vulnerable to anti-trust laws.
No, everyone who’s a registered developer knew they couldn’t use a different browser engine. That’s been the policy from day 1.
It would be different if Apple allowed 3rd party rendering engines and then forced them off, but that’s not the case.
Among other things, 3rd party engines would be incompatible with Apple’s Lockdown feature. Because browsers are required to use WebKit on iOS, they automatically work with lockdown mode; it would be much more difficult if every browser on iOS had it’s own rendering and JavaScript engine:
On a more technical level, a number of web technologies are turned off,
including just-in-time (JIT) JavaScript compilation (where code is run
and compiled at the same time.) In addition, configuration profiles
(for work or school for example) can't be installed.
You are right about Microsoft, and what you're describing is also exactly what Google is doing (threatening to cancel Play Store licensing, while Play Store is the dominant licensable app store).
Google is already sued for that, but the sailant point to me is that it took so much time, and it's far from a clear cut case.
We are way past the point where Apple, Google and Amazon abused their position in provable ways, but they got left alone for so long by the regulators it's disgusting.
PS: the definition of a monopoly comes down to the definition of the market. If we were playing games, Microsoft wasn't a monopoy when looking a computing platforms in general, for instance, it was "only" on the PC home market. Apple has a monopoly under many framings, now it comes down to wether a judge agrees or just kicks away Apple opponents.
Apple has a monopoly under many framings, now it comes down to wether a judge agrees or just kicks away Apple opponents.
How so? Google has 80-90% of the search engine market and Microsoft had 95% of the desktop operating system market and used that disadvantage competitors in the nascent browser market.
The bottom line with Microsoft was: should Microsoft be allowed to use its monopoly position in operating systems to essentially force customers to take your web browser, whether they wanted it or not? It certainly wasn’t okay for Microsoft to use the threat of canceling HP’s ability to be a Windows’s OEM (and therefore destroying their PC business) if they made Netscape Navigator their default browser.
People forget how dominant Word Perfect and Lotus 1-2-3 were back in the MS-DOS days; Microsoft used Windows to make Word and Excel the dominant word processor and spreadsheet.
Apple’s worldwide market share is barely 30%; it’s closer to 50% in the U.S. It’s also not illegal to disallow 3rd party developers to “break” your platform, especially when they entered into a legal agreement with you to not break your platform.
There’s nothing illegal about controlling a successful platform and deciding who and what can be present on that platform. Game consoles have long operated the same way but that seems to be okay.
Apple doesn’t have a monopoly position in either phones or computers; if a potential customer doesn’t like how Apple operates, they are free to buy something else. That wasn’t the case with Microsoft back in the day and it’s pretty much not the case today with web search, since it’s realistically not an option for a business to not be available on Google search or to ignore its advertising platform that reaches billions of users.
A lot of percentage are thrown around in your reply, but all of these are relative to a specific market you define. Wether it’s the relevant market definition is up for grabs.
You should have a look at the Apple vs Epic fillings, where Epic came with a pretty different perspective on which definition should be taken into account. They of course lost as the US judge sided with Apple’s definition (while Korea came with a different verdict…), but that’s at least an indication of how important that definition is. “30% of device share” or “80% of in-app purchases” are completely different perspectives.
> if a potential customer doesn’t like how Apple operates, they are free to buy something else
Customers are mostly irrelevant. To get back to your Microsoft example, no developper was stopped from developping for linux, nor were customers stopped from buying barebone machines, or Apple, or beOS, or mainframes. The issue was Microsoft making backroom deals with the vendors to make Windows alternatives pricier and non competitive. This makes it worst for customers as a result, but that’s the result, not the core issue (I’m referring here to the EU rulings). Definition of a anti-trust issue should be about how a player distorts the market by pressuring other business, not wether the customer has theoretical choices or not.
>I’ve written this many times here on HN: a monopoly isn’t illegal; what is illegal is using your monopoly in one market (PC operating systems) to disadvantage competitors in another market—web browsers.
Ah yes, the free market will come up with a solution to this.
> By contrast Apple today is locking down their platform and explicitly decreeing that nobody is permitted to compete with them.
Microsoft got in trouble for anti-competitive practices with IE, because they were tightly integrating their OS with their browser in ways that meant that their browser was uniquely enabled by the OS to do things that third-party browsers couldn't do (not having access to the same APIs), thus using their position in the OS market to steal browser market-share away from third-party browser companies like Netscape, in ways those browser companies could do nothing to respond to.
In contrast, on iOS, there's nothing Safari can do that third-party browsers can't. Third-party browsers do exactly the same things Safari does, rendering-engine-wise — no more, no less. And meanwhile, they all do more than Safari does UI-wise. There's nothing Apple is doing to bias users of iOS in favor of Mobile Safari, such that it steals market-share away from competing browsers. Instead, their OS is just ruining the browser experience for everybody, equally. That's not anti-competitive — it's just dumb. :)
Apple blocks allocating executable memory which is necessary in order to create jit compilers, which are necessary for JavaScript engines to run at acceptable speeds. This restriction does not apply to safari’s js engine though.
I’ve implemented a WKWebView for a purely personal app with very specific but very minimal requirements and… “just” is doing a lot of heavy lifting here. You can “just” use the API and you get a fully functional web browser without any UI, but you certainly aren’t going to attach any meaningful custom UI to it that way without digging deep not just into internals at the binding level but at the WebKit level. I’m not saying it’s a bad design either, but it’s definitely not a trivial investment.
Don't want to be snide, but you're talking about the high barrier-to-entry for implementing a nice, customized WKWebView in a project that isn't a web browser. In such a case, the WKWebView brings with it a bunch of complexity that wasn't already there otherwise.
If your project is itself a web browser, then all that complexity is already inherent to the project, whatever you choose to build it with. It's just that the "digging deep" would be in the docs for how to glue Apple's low-level DOM/JS/etc library frameworks together in mostly-undocumented ways; rather than in the docs for how to feed WKWebView the right delegate logic in just the right place to get it to do something that'd be easy if you were directly doing it yourself.
Didn’t take it as snide but even just the delegated logic isn’t so simple. Most of my personal project was just a specialized browser, and really basic functionality was much more complex than I’d hoped. I got it good enough to use and haven’t maintained it since because it was just for my own use, so maybe the APIs have become less complex since.
Microsoft got in trouble for anti-competitive practices with IE, because they were tightly integrating their OS with their browser in ways that meant that their browser was uniquely enabled by the OS to do things that third-party
It was actually worse than this--when Microsoft was ordered to separate IE from Windows, they lied and said IE was integral to Windows and that it could not operate without it.
Turned out Windows worked just fine without IE and Microsoft ended up settling with the government.
But Safari is the default. And if all the options are equal, then there’s no reason to switch from the default. It’s hard enough to convince people to switch browsers in the best of cases, let alone when the list of potential advantages or differentiators you can offer is severely cut down.
iOS supports changing the default browser to something that’s not Safari.
It’s not only about the rendering engines; Brave on iOS has other features and a UI/UX that I like that’s not present in Safari.
3rd party browsers on iOS have other ways of differentiating themselves, especially if you use those browsers on other platforms, like bookmark syncing, etc.
90% of end-users couldn’t care less what’s being used to render their websites on iOS.
The free and open web, to the extent we have it, only exists today because Microsoft wasn’t allowed to force IE, Active X, Silverlight and the rest of their technologies down everyone’s throat. Oh, and JScript [1], their proprietary version of JavaScript would have been the industry standard.
We wouldn’t have the W3C or WHATWG creating open standards for the web had Microsoft been allowed to essentially turn the web into another Microsoft runtime environment.
> The free and open web, to the extent we have it, only exists today because Microsoft wasn’t allowed to force IE, Active X, Silverlight and the rest of their technologies down everyone’s throat. Oh, and JScript [1], their proprietary version of JavaScript would have been the industry standard
The justice department at no point stopped Microsoft from bundling IE with Windows.
Besides that, up until 2008-2010, the “free and open web” was littered with proprietary Flash apps. It wasn’t until Apple refused to support Flash on iPhones that anyone started worrying about standards.
> Microsoft been allowed to essentially turn the web into another Microsoft runtime environment.
Nothing that the FTC did prevented that. No part of the consent decree caused MS to behave any differently with respect to IE.
And if you don’t remember, the “modern web” as it exists today with Ajax was actually a Microsoft addition that everyone else added.
JScript was no more or less proprietary than Netscape’s “LiveScript” before Netscape jumped on the Java bandwagon and called it Javascript. It’s not like JS came from a standards body.
JScript was no more or less proprietary than Netscape’s “LiveScript” before Netscape jumped on the Java bandwagon and called it Javascript. It’s not like JS came from a standards body.
Popular programming/markup languages don’t originate from standards bodies; it’s usually one person or a small group to start. Ruby, Python, Lisp, Pascal, Perl, PHP were created by a single person or a small group and (sometimes) would be submitted to a standards body.
Javascript was created for the Netscape Navigator browser in September 1995 and was submitted to Ecma International November 1996 [1] to start the standardization process. That was probably in response to Microsoft reverse engineering Javascript to create JScript in… 1996. Funny that.
So… yes, turns out JScript was significantly more proprietary than JavaScript and was only created to further Microsoft’s dominance of the web--“Best viewed in Internet Explorer”. JScript, like most of Microsoft’s tech at that time, probably relied on features that only existed on Windows and no where else.
> Javascript was created for the Netscape Navigator browser in September 1995 and was submitted to Ecma International November 1996 [1] to start the standardization process. That was probably in response to Microsoft reverse engineering Javascript to create JScript in… 1996. Funny that.
So it wasn’t “proprietary” when Netscape created it. But it became “proprietary” when Microsoft copied it?
> So… yes, turns out JScript was significantly more proprietary than JavaScript and was only created to further Microsoft’s dominance of the web--“Best viewed in Internet Explorer”
Were you around then when before IE, there were plenty of sites that had “Best viewed in Netscape Navigator”?
And don’t pretend that Navigator was a great product. In its heyday in the mid 90s, it was a point of nerd pride how well your operating system handled a Netscape crash.
IE3 was much better than Navigator on the Mac and Windows. In fact, the Max version of IE was one of the most standards compliant at the time.
The justice department at no point stopped Microsoft from bundling IE with Windows.
Nobody suggested otherwise; but that actually wasn’t the point. Microsoft made a deal with the government to stop most of its anticompetitive behavior. It was less about the bundling of IE with Windows than giving OEMs the choice to ship other browsers in addition to IE.
Besides that, up until 2008-2010, the “free and open web” was littered with proprietary Flash apps.
But that was a choice made by the market to adopt Flash instead of having of having all of the Microsoft stuff forced on them. Nobody forced developers to adopt Flash and there were other choices for audio/video, even if they weren’t as popular as Flash.
And if you don’t remember, the “modern web” as it exists today with Ajax was actually a Microsoft addition that everyone else added.
Well, to be accurate, Microsoft shipped XMLHttpRequest in IE 5.0 in 1999, as a proprietary Active X component for their products; it was never intended to be a web standard. But it was a cool idea so the other browser makers implemented it.
The term AJAX was coined by a Google employee in 2005 and was standardized by the W3C in 2006 [1].
It’s no different than Apple developing the canvas [2] element, which became a web standard, except they were more intentional than Microsoft was about XMLHttpRequest which would be renamed as AJAX.
Things would have been very different if the FTC and European regulators (who did make Microsoft ship a version of Windows without IE [3] hadn’t gotten involved.
> But that was a choice made by the market to adopt Flash instead of having of having all of the Microsoft stuff forced on them. Nobody forced developers to adopt Flash and there were other choices for audio/video, even if they weren’t as popular as Flash.
The comment I replied to gave credit to the FTC. Which is not true.
> Microsoft made a deal with the government to stop most of its anticompetitive behavior. It was less about the bundling of IE with Windows than giving OEMs the choice to ship other browsers in addition to IE
Microsoft didn’t just promise “we will be good little boys”. The parent poster gave the credit to the FTC. Firefox didn’t become popular for awhile because it was bundled by OEMs. MS basically just stopped development on IE. Then Chrome took off because it was heavily advertised on the Google home page and bundled with third party downloads.
As far as IE not being bundled with windows 7 because of the EU. This is an example of government ineffectiveness. Windows 7 came out in 2009. Web browsing on the desktop was starting Gibbs yesterdays news by then. The iPhone had been out for two years and Chrome was already ascendant.
And that’s about as useless as the EU caring to protect users privacy and writing a 99 section 11 chapter law and the only thing users got were cookie pop ups.
Depends how you see the relevant market. You could define the market as browsers for iPhones, instead of browsers for all phones. After all, it’s not like users switch readily between different phones for different apps. There’s a whole body of case law on what the relevant market is. The gist is that it turns on substitutability, but it’s more complex and flexible. Easy knob to turn to account for tech platform dynamics.
If the friction is high enough, antitrust law has always seen that as bearing on the definition of the relevant market. You can’t lock people in and then start charging monopoly prices with impunity. One issue is what happens when the stickiness is mostly psychological friction instead of something that an economist analyzing rational action would see as a barrier. These are laws for people, after all, not for hypothetical rational beings.
it is anti-competative by denying 3rd parties from provinding more.
Imagine a truck maker who could somehow prevent other truck makers from making a truck that caries more or runs faster or is more fuel efficient. Of course that would be anti-competative.
The issue raised is not about video codecs, but about browsers. Apple is using their market position to artificially prevent Safari from being competed against by other vendors—that is classic antitrust behavior.
For comparison, imagine Microsoft banned all browser engines other than EdgeHTML from running on Windows. Would you find that acceptable?
You may not be aware, but chrome does exist on iOS, but even so, it is actually running the safari rendering engine because apple prevents other browser engines from running. This effectively prevents chrome or Firefox from offering a browser with av1 support (as far as I am aware) so the browser engine is relevant to the codec policy discussion.
Given that Chrome abuses its market position to ram through and continuously release hundreds of Chrome-only non-standards I view this as a net positive.
Which is the reason why I choose iOS: I trust Apple to implement the integrations in the WebView for best battery life. To get new OS features in the browser right away. Would Google Chrome implement picture-in-picture on the iPhone if they were running their own engine?
I use Firefox on iOS for the password and tab sync integration, and am supper happy that it’s using the WebKit framework. I care about the iOS experience more than some video codec support.
If the other browsers could implement their engines, you would still have the option not to install them.
So it’s okay for user who (in theory) installed a 3rd party browser on iOS to have a worse experience and potentially make themselves vulnerable to malware? That’s exactly what Apple doesn’t want.
Your argument is a straw man, he's saying that apple should be force to allow competing browser engines on iOS so other vendors can support what they don't want to...
I mean it may save them bandwidth on Apple TV+ or iTunes TV/movie purchases. But it seems they’ve chosen not to bother.
Maybe all the re-encoding would negate the benefits for them?
Maybe they didn’t think it was enough of an improvement over h264. Maybe they bet on h265 and it worked out well enough (for them) or it didn’t but they figure by the time they got everything set for VP1 something better will be ready.
I believe their main reasoning for not supporting AV1 would be the lack of hardware decoding on their chips.
A minor improvement in network bandwidth and storage costs wouldn't be as high a priority as "up to 16 hours video playback (streamed)" on the iPhone 14.
Sure they could optionally support it in Safari without hardware decoding, but a service like YouTube will want to negotiate the lower bandwidth protocol over one that will make the device's battery last longer.
Maybe the licensing cost of h265 works out better for them than rushing out AV1 decode support on their chips or lowering the average battery life of their devices.
That would certainly have been the argument at first. Battery life is EVERYTHING.
But Netflix announced they would be using it in 2016. YouTube started using in 2018, the year the standard hit 1.0.
Apple had plenty of warning and time to get the necessary hardware acceleration blocks available by now. Chips have multi-year timelines but it’s been 4+ years.
Now on the other side, Netflix actually started using it in early 2020/late 2021, depending on device. YouTube rolled it out wide in 2020. Twitch is working on it.
So Apple is at most 2 years behind the content. It’s not like it’s been 5+. And if they decided to wait to see how those early rollouts went due to hardware timelines it might make sense support only appears to be coming now (based on someone pointing out AV1 shows up now in the AVPlayer framework documentation).
I believe their main reasoning for not supporting AV1 would be the lack of hardware decoding on their chips.
I doubt it.
Apple Silicon is more than fast enough to decode AV1 in software. Sure, the battery life won’t be as good but there are other trade-offs that can be made.
Their latest chips were developed before AV1 had gained enough traction to be concerned about. I suspect that future M2 and M3 and A-series SoCs will have hardware decoding for AV1 built-in.
In the meanwhile, a future version of macOS Ventura and iOS 16 will have AV1 support; it’s just going to take a little longer. They probably want to roll it out simultaneously on iOS, macOS, tvOS, and iPadOS.
When AMD/Nvidia/Intel and Samsung/Mediatek/Rockchip/etc were able to put AV1 blocks into their chips, I'm sure Apple would be capable of doing the same -- if they wanted. Now it looks that they do not want.
Apple might have made a deal with one or more major h.265 patent holders that discourages them from fully supporting av-1, similar to how Google pays apple nine digits a year to keep Google search default in the Apple ecosystem.
Court action could compel any deal like this to be made public.
> What Microsoft did in the 90s was pale by comparison
You really don't know what you are talking about. The issue with Microsoft was not bundling a piece of software and not fixing the box model.
Microsoft (of the 90s) deliberately worked to undermine the web using their dominant market position, adding their own extensions to ensure lock in. Their behaviour in the web standards process was appalling, and directed from above (sometimes by phone outside meeting).
"Embrace, extend, and extinguish" was not hyperbole.
> I don't understand why they're sitting on their hands. What Microsoft did in the 90s was pale by comparison
Windows had, IIRC, something like 97% of the market. It appears[0] that iPhones have a bit over half the phone market today, but they have nothing like the monopoly that Microsoft held.
The difference is a) the FTC was way more lax in enforcing anti-monopoly laws than back in the day, b) Microsoft was a monopolist, c) downloading software from the internet hasn't been as much of a thing back then, so obtaining internet explorer replacements was more complicated, so preinstalling something was going to influence pepole's descisions way more than now where installing an app is just one visit to the app store app away. Only years later with improved internet speeds for everyone we got the ability to install chrome with a simple download. Lastly, d), you also have to consider that Chrome and Chromium based engines basically have a monopoly right now. So I'm not sure what the government stepping in will bring to users: Right now it's more that Safari is the only non-Chrome browser that most web devs care about.
Re b: Apple has never been a monopolist so they never got into that situation. This allowed them to have such opinionated choices that would have never worked for Microsoft. However, this might be changing. Recent statistics show that they have extreme market shares in the teenager demographic. If those kids grow up and keep buying iPhones, Apple will become one, in the US market at least, and this means they will have to change.
It's possible the EU has sorted this out for us already. The Digital Markets Act comes into effect 1st November and really begins to apply in May next year.
One of the thing that it specifies is that Gatekeepers (Apple in this case) can't prefer their own services and apps over other people's - they must allow third party interoperability. If I understand it correctly:
* You must be able to uninstall preloaded first party apps
* You must be able to install your preferred third party equivalents (e.g. Firefox)
* I believe also - you can install alternative app stores
We shall see what actually happens once Apple's lawyers dig out loopholes, and we'll see if the Gatekeepers of this world have an alternate 'EU vs rest of world strategy'. Also, if that does happen, which side the UK falls in...
I may be misremembering things, and I'd appreciate the correction if someone can show me my error here.
When the iPhone was first released, telcos were extremely paranoid, because there were legit scams by which an abuse of the telco network could result in monetary transfers. If you called the wrong number, whether a US (and Apple was a US company first and foremost) 900 number or an international number, you might owe a great deal on your next telco bill. The original iPhone had no native code applications, only web apps, because of the paranoia resulting from scams that had been perpetrated in the 90's. The telcos required a level of security that, in the early 00's and its embrace of Adobe Flash and other high-vulnerability-surface technologies, seemed prudent.
A short while later, Apple proved its security model enough to allow native code applications. This happened because, despite the initial fears of a "smart phone" being a lure for scammers, Apple's security model, which included mandating WebKit so that its security model would be imposed on anything touching the web, kept those web things separate from telco charges.
Fast-forward to 2022: now we wonder why Apple keeps locking things down so that no app can use a security model that Apple has not vetted. In this era, rogue calls to Nigeria don't drain thousands of dollars from your account. Now, we wonder why Apple seems so paranoid.
For me, whose memory reminds me that there was a time that telephony involved the risk of connecting to the wrong number and losing money, there seems to have been a time that "smart phones" ran the risk of presenting an attack surface that could incur dire financial consequences for individuals. I also suspect that those times have passed, but I'm not sure. I have to wonder what the intersectional implications of a more open environment for a device capable of both web browsing (with its security issues) and telephony (with its billing issues) might be. When I wonder about things like that, I'm happier with walled gardens that I would not accept on my laptop, which is not connected to a phone number and billing account.
A computer running Windows and using IE is one thing, namely a computer on an IP network; a phone running iOS is two things, a computer and a telephone, and there are (or were) mechanisms by which a telephone could charge money that a computer cannot (or could not).
Why do people act like the results of the FTC case is lost to the annals of history? Absolutely nothing happen with the case with respect to bundling. There was never a browser choice initiative nor was there ever a time since IR was bundled with Windows was it not bundled.
This “the new IE” meme is getting sillier by the WebKit Technology Preview release cycle and has been silly for two whole Safari release cycles now. The WebKit team has significantly expanded and isn’t only bringing compatibility up to date at a rapid clip but shipping some highly anticipated APIs first.
I objectively would prefer other browser engines were available on iOS, but at the moment that would just mean further entrenchment of Chrome/Blink, which I don’t think would be better. I’m happy to see WebKit continue to improve and embrace the open standards that benefit the web. It’s gonna take some time but given the options I think patience is a better option than ushering in Google’s total dominance of browser technology.
2. The product stagnated and became behind on all of the web technologies.
3. Eventually they caved and started catching up.
vs
1. Apple was relying on Google's developers which forked and stopped developing WebKit.
2. The product stagnated and became behind on all of the web technologies.
3. Eventually they caved and started catching up.
Even before IE was discontinued I had a much better experience with them than Safari (about 8y ago now). Not only did it support more recent APIs but I could download a VM with IE and Edge installed to test. Now almost a decade after the Blink fork they are committing to adding new features that people have been asking for for years. We'll see if they actually manage to catch up like Microsoft did. (Even before they switched to Chromium they were much more compatible than Safari).
1. Apple was relying on Google's developers which forked and stopped developing WebKit.
2. The product stagnated and became behind on all of the web technologies.
3. Eventually they caved and started catching up.
1. Google was at some point the leading committer on WebKit, but there were other companies and organizations who contributed code as well [3]: Nokia, RIM, Igalia. They weren’t dependent on Google; but in hindsight, they knew at some point they were going to fork WebKit to create Blink, so they had incentive to put in a lot of effort.
2. Apple had different priorities than Google; many of the early CSS and other web features came from Apple/WebKit. Google had no problem just implementing things whether other browser vendors agreed or not.
If you look at the Interop 2022 dashboard [1], you’ll see that WebKit is either at parity or ahead on the focus areas.
Apple shipped :has(), the most anticipated CSS feature in a long long time before anyone else. They’re shipping in iOS 16 all of the cutting edge features; I documented them earlier in this thread [2].
Internet Explorer was extending the standard to the point of incompatibility. That was a goal and we know because of the trial:
"kill HTML by extending it"
(As recalled by Steven McGeady Intel VP told by Paul Maritz Microsoft VP)
So Internet Explorer was quite the opposite of today’s Safari: Microsoft was trying to be faster than Netscape, and to put forward new non-standard specifications faster than Netscape could or would add them. AJAX for example started as an ActiveX component, a technology that tied most components to Windows.
If we draw parallels, it looks more like what Google is (accused frequently of) doing than what Apple is doing (releasing yearly, adding standards after others)
Of course when Microsoft drove Netscape to bankruptcy, then they slowed improving their monopolistic browser, as only Windows was important to protect for them.
As a user, I appreciate my user agent restricting its codec support to those that can be played back efficiently on my hardware. I see this in contrast to Google not caring about my battery life or device thermals and forcing CPU decode on me if it saves them a fraction of a cent on royalties or bandwidth.
Google isn't in control of the hardware Chrome is running on. Many of those devices do support Av1 decode at a hardware level so it is efficient. Apple has complete control over all their hardware, and they don't support av1 decode. They're absolutely holding progress back.
It’s fine to disagree, it’s not unreasonable to have the take that making users pay a cost in terms of extra cpu usage is worth it for the benefits of AV1. But it would be better if you could make that argument without the hostility.
Anyhow, I will also point out that there is plenty of non Apple hardware in use that also doesn’t have dedicated AV1 decoders. I realize that’s changing, but it’s still got to be a low % of devices at the moment with dedicated AV1 support. I think a more user friendly posture from Google would be to only report AV1 capability in WebRTC negotiation or video tag source selection where it can be done energy efficiently. Happy to be shown to be wrong on that, but my understanding is they don’t do that today.
Hostility wasn't my intention. Sorry, it came across that way. Please forgive me. I can see my mistake--I'll try to improve going forward. Thank you for your correction and for keeping a cool head. Last thing I want to do is make the internet a more trollish place. It has plenty of that already.
> Safari (especially on iOS) appears to have become the new Internet Explorer, holding back progress for everyone. The table in the article certainly suggests it.
Ah yes. Tables. Let's look at the actual text and text in some linked articles [1]
=== start quote ===
Last year, Google began requiring AV1 support for new Android TV and Google TV devices
...
[1]
Google is requiring makers of Android TV devices to support AV1 starting this month. Additionally, Google also seems to push makers of smart TVs and streaming devices not based on Android TV to use AV1 for YouTube.
Google has long forced device makers to use the free VP9 codec for 4K YouTube streams.
=== end quote ===
So, Google is abusing its market position to ram through its own standards. But sure, it's Apple and Safari who are the bad guys.
AV1 is the best royalty-free video codec, and it's pushed by the AOM industry group (which includes Apple!), not just Google. Apple doesn't support it; Apple is trying to get people to use the patent-encumbered HEVC codec. Both are leveraging their market positions, one to push a royalty-free video codec, one to push a proprietary one. The fact that you describe what Google is doing as "abuse", but let Apple off the hook, makes no sense.
They are switching their product, youtube, to av1. If youtube couldn't use av1 then it would be device manufacturers who are abusing their power. Can you explain the abuse here? AV1 is an open and royalty free standard.
They are not "switching youtube to AV1". They are "using their dominant market power to force everyone else to switch first to VP9 and then to AV1, or else"
Device manufactures couldn't care less about AV1. And yet, Google made AV1 support mandatory for Android, Android TV etc.
Or else... what? As you say, hardware manufacturers are holding back codec progress. It's google's product, they can force manufactures to comply, but it's not abuse because it's a better and open standard. All consumers benefit from this move.
>The mere fact of it being better or open doesn't make what Google is doing any less of an abuse of their position.
Except it literaly does. Abuse in the market it not defined by the loss of another company but by the loss of the consumer. Anti-competition laws are ultimately meant to protect the consumer. The consumer. "Abusing" a company in itself is perfectly fine: it's a free market after all.
Abuse of dominance occurs when a dominant business (or group of businesses) engages in activity that stops or substantially reduces competition in a market. These anti-competitive activities may be:
predatory (incurring short-term losses to eliminate a competitor and gain future market power);
exclusionary (trying to prevent a business from operating in a market);
disciplinary (trying to punish a business); or
intended to adversely affect competition (e.g., by making other companies want to compete less and denying consumers the benefit of competition)
I agree. Google should be able to make whatever they want, declare it a standard by fiat since they own YouTube, shove it down everyone’s throat, and then get some partners so the standard claim isn’t as ridiculous.
If YouTube hadn’t pushed it, would it be popular? That’s a BIG question in my mind.
I understand it is the most efficient (unless h265, which is patented, is better).
My point about YouTube is it gives Google power. They can say “we’re doing this with billions of hours of video” and people must follow. They can demand chips that encode faster. Or decode faster. Or say “No YouTube app on your custom gizmo device without VP1 support.”
They have the power to force changes in the market the way a startup or even Microsoft don’t. Those two would have to convince people the change is worth it to everyone. Google just does it, and others follow because it’s now there.
I’m basically arguing Google abused their position as opposed to Apple.
Using open standards can almost never be abuse because nobody controls it. The reason Google wants it benefits everyone who streams video. MV3 is a better example. It's technically "open" but it is removing functionality. I don't think anyone here disagrees with you that Google has power. But someone always has power; in the tech world it's mostly uses for profit. In this case it happens to benefit everyone and we should be happy about it.
Isn't HEVC comparable? It may lack some sheer bandwidth efficiency, but it seems to make up for that with less compute cost. And the entire M1 line has hardware and software HEVC support.
edit: Oh, you said royalty-free. Fair I guess? I guess I'd be pretty annoyed if I decided to encode some personal project I ended up making money from with my M1's HEVC encoder and got dinged by royalty police.
I mean maybe? But does that matter? It's by far the highest efficiency (in terms of bandwidth) open codec out there. It takes a lot of money to develop these sorts of things, especially given the patent risk, so I'm not sure how else you'd expect something like this to be created.
Netflix and Facebook were also early adopters of this. Between Google (Youtube), Facebook (various) and Netflix you have three of the biggest content providers. Other big ones like iQiyi and Vimeo jumped on it pretty quickly too.
Is there any evidence for this beyond kCMVideoCodecType_AV1 in the 16.0 headers? I'm not sure that alone implies any particular version. Could be iOS 17 or later for all we know.
Do you happen to have a link? I actually do follow them and their blog pretty closely and this is news to me. I realize we are all reading tea leaves here, but any public info at all from an Apple source would be interesting to me.
Sure it's nowhere near as bad relatively speaking as IE was at it's worst, but Safari is still consistently slower to make new features available than other major browsers.
This still isn't really a problem (if it gets bad enough, devs will just stop supporting it), except for on iOS where it's the only browser available, so not supporting them is not an option.
but Safari is still consistently slower to make new features available than other major browsers.
Safari was the first browser to ship the most anticipated web feature of the last 3-4 years by web developers: the parent selector :has() [1] back in March.
When you check the Interop 2022 dashboard [2], Safari Technology Preview is ahead of both Firefox nightly and Chrome dev for the shipping the latest web features. Safari Technology Preview is passing 97% of the interop tests.
In case nobody noticed, the WebKit team kicked ass by shipping a ton of new features this year:
* dialog element
* lazy loading
* inert
* :has() pseudo-class
* new viewport units
* Cascade Layers
* focus visible
* accent color
* appearance
* font palettes for color fonts
* BroadcastChannel
* Web Locks API
* File System Access API
* enhancements to WebAssembly
* support for Display-P3 in canvas
* additions to COOP and COEP
* container queries
* subgrid
* web push
* shared workers
* CSS Offset Path
* AVIF
* Passkeys
Plus it’s faster and has better battery life on macOS than either Chrome or Firefox. What’s not to like?
Another point here: you can't update safari without updating ios itself. That means people who don't update or have old devices won't get new fetures. Also, safari STILL doesn't support a lot of new wasm extensions. So not really the picture GP is painting.
Another point here: you can't update safari without updating ios itself.
True. But each release of iOS gets to around 90% of the installed base by the time the next version of iOS gets rolled out; it’s non-issue for the vast majority of iOS users.
iOS 15 was updated 12 times [1] between September 2021 and July 2022; iOS users either got bug fixes or new features (usually both) in Safari each time, including users of the iPhone 6s, which shipped September 2015, more than 7 years ago.
Also, safari STILL doesn't support a lot of new wasm extensions.
According to caniuse [2], Safari’s WASM support seems to be just as current as Chrome and Firefox.
[1]:
Version Build Release Date
15.0 19A341, 19A346 September 20, 2021
15.0.1 19A348 October 1, 2021
15.0.2 19A404 October 11, 2021
15.1 19B74 October 25, 2021
15.1.1 19B81 November 17, 2021
15.2 19C56, 19C57 December 13, 2021
15.2.1 19C63 January 12, 2022
15.3 19D50 January 26, 2022
15.3.1 19D52 February 10, 2022
15.4 19E241 March 14, 2022
15.4.1 19E258 March 31, 2022
15.5 19F77 May 16, 2022
15.6 19G71 July 20, 2022
> …but Safari is still consistently slower to make new features available than other major browsers
See I’m not so sure that’s a problem. Many of the things they haven’t added I’m glad they didn’t. I like that they aren’t just rushing to implement everything someone proposes or a competitor ships.
Google is effectively shipping a cross-platform “metaplatform”, for which the browser is the trojan horse. Apple is shipping a web browser that only runs on its OS and hardware.
So Webkit is slow or unwilling to implement features that may hinder battery life, user privacy or that simply have better native alternatives. Webkit is actually on the forefront when it comes to CSS features, like the recent has() selector, or JS memory usage and overall page speed.
What I think is a fair criticism is that Safari is generally tied to the OS and gets outdated when the user is unable to upgrade the OS (or oblivious). Evergreen browsers are a revolutionaty concept, not without its drawbacks, but all in all, were very positive to the web.
I use Safari but I would prefer it wasn’t shipped on the same schedule as the OS. I understand why (I think) but I’d rather have more frequent updates, even if only every 3 months or so.
That’s a problem with most Apple apps. Bug in Mail? Better hope it gets fixed in the next iOS or MacOS point release. Because the fix isn’t coming out earlier.
Safari Developer Preview is stable and available out of sync with the OS.
It does make sense that it is shipped with the OS since Safari is just a wrapper around the OS WebKit library which is used across many apps and services.
A browser is so fundamental to modern computing that if you can't run the latest version, you might as well have a paper weight. You can run mail, office, games, all in the browser if the native alternatives fail you.
Chrome still supporting High Sierra is what makes old Macs usable.
How many of those "actual HTML standards" in the past 10 years are actual standards and not "whatever Google scribbled on a napkin and pretended it was a standard"?
> (The alternatives can’t replace the important parts)
I’m genuinely curious what you mean by this. I don’t think I’ve opened Safari in iOS in 5+ years. I am also not as tech savvy as most of you so the question is genuine lol
Whatever you use (I’m guessing Chrome?) is actually Safari under the hood on iOS, not Chromium.
Apple has stated reasons besides “control everything”. Many people dislike/disbelieve the reasons and have objections. It’s not worth litigating here (see: every thread anywhere mentioning iOS or Safari).
The point is, from a technical point of view there is only Safari on iOS. Chrome can’t add new APIs or rendering features that aren’t built into Safari.
And since Safari on iOS is very common and adds things when/if Apple decides, it’s often compared to IE6, rightly or wrongly.
(Not trying to debate the issues, just explaining the comment)
I mean you can “remove” it. It won’t show on your home screen anymore.
But under the covers the code that does all the heavy lifting will be the Safari code. Firefox on iOS will get new features when we Safari does, because that’s how (it’s forced) to work.
—-
Don’t worry about the last part of my comment. I’m happy to explain this, I’m just tired of the debates of “policy X is an antitrust issue.” Or “They’re doing it to kill PWAs.” Or “The JIT rule is just an excuse.”
The debate has existed since iOS 2.0 in 2008. So 14 years of roughly the same arguments. And I’m just tired that it’s impossible to talk about Safari or iOS or Apple without the exact same argument happening. Like elsewhere in these comments.
That’s the discussion I don’t want to have. Your question is totally fair.
Apple makes other browsers on iOS use the safari engine under the surface. So the core functionality you're seeing of, e.g. rendering the webpage and video codec support is just safari. Chrome/FFox are just sugar around the safari engine, they won't allow other engines
Every browser on iOS is essentially safari but reskinned. Apple forces browser apps to be WebKit based. That means supporting other codecs, and different types of web technologies, etc, only happens on iOS when WebKit supports it, because all browsers have to use the WebKit that’s provided by apple.
On iOS, yes. My understanding is that Apple forbids installing full Safari alternatives on iOS. On iOS the only permitted browsers are Safari reskins that use the Safari webkit. Source: https://en.wikipedia.org/wiki/Firefox_for_iOS
If I'm misinformed please let me know.
This greatly impacts ALL web development. Many people use iOS, and Apple prevents the installation of real alternative browsers. As a result, it's typically financial suicide for any web developer to use a client-side feature not currently supported by iOS Safari.
There is one partial exception: Puffin. It uses Chrome on Linux and a custom lightweight remote desktop-like client to let the server render the pages and the phone view them.
It's an impressive work around, though it adds it's own problems, not the least of which putting a lot of trust in a third party.
When Apple deployed support for HEVC in 2017, all devices that supported iOS/tvOS 11 or macOS High Sierra gained the ability to play HEVC video (some via hardware, some via software). Something very similar will happen when Apple deploys AV1 support.
The old Apple TV got 4k 60fps VP9 decode in Youtube from A12 onward, and the AV1 software decoder is roughly similar in load, so 4K should be doable.
Wouldn't be surprised if the old one was already using AV1 for Youtube.
edit: there are 3rd party apps for the Apple TV that play back VPR and AV1 in software, this thread suggests that even on the old hardware, AV1 decoded better:
> 4k HDR @ 60 fps encoded with AV1 - Plays much better than previously, there are still some frame drops in more complicated scenes (for example when they show palm tree on a beach in Costa Rica video). Unfortunately it does not trigger HDR mode on my TV (all other HDR and Dolby Vision content does it correctly).
> 4k HDR @ 60 fps encoded with VP9 - It works, but performance is not as good as AV1. Does not trigger HDR mode on my TV. More dropped frames, video can freeze (audio plays, but picture is frozen).
Worth noting that Youtube often has lower relative bitrates/quality levels than other uses.
Right I think this is a point many miss. Most Apple HW can do AV1 decode in SW plenty fast, it's just much more energy than with a hardware decoder (big deal for phones in particular). What's really missing is the SW support, but I'm delighted to hear that it's _finally_ going to happen.
To be fair the A series chips are very powerful (compared to something tiny like a Roku stick) and don’t have the thermal/power limitations of an iPhone. They may be able decode VP1 at full speed without VP1 specific hardware.
I don’t think YouTube will abandon VP9 anytime soon either though. They want AV1 for improving quality and lowering bandwidth - but VP9 and MPEG before it will stay a long time yet.
ffmpeg project doesn't ship binary; patents do not apply for study and research. Now shipping ffmpeg binary in some jurisdiction would need licensing. If you compile and use it yourself, the licensing in these jurisdictions is on you.
Another factor for AVC against HEVC is, that at some resolution and bitrates, AVC still provides better results.
Anyone can set up a patent pool claiming to be for anything, and hope fools will buy it. This scam has been done to previous VPx codecs too.
And anyone can sue anybody for anything. Until there's a judgement it doesn't mean much. But there is an alliance around AV1 patents, and it has strong backers to fight off the trolls: https://aomedia.org/membership/members/
> AOM is being sued by the EU for not offering FRAND terms
Can you expand on this? If I'm reading correctly, the issue is 'mandatory royalty-free cross licensing' which is supposed to bring down innovation? If yes, this sounds like some grade A bullshit.
Fairly certain that VP9 support in Apple hardware existed in shipped devices before they 'turned it on' with a software update. I'm assuming that if someone had done a deep dive on the hardware and found this then we'd know about it, but not sure whether lack of that proves it doesn't exist.
They certainly had intel chips with VP9 decode that was blocked/unused by the OS, but I was specifically referring to Apple in-house designed/made chips which apparently also had VP9 hardware support that was unnannounced.
It always had Bluetooth support, it was just locked to the Nike integration they had a partnership with. The update gave it support for other devices though.
Nvidia's 4000-series GPUs include pretty good hardware AV1 encoders, marketed towards streamers as a way to have higher quality streams on Twitch for the same bitrate.
