Hacker News Reader

Comment by agust | Hacker News Reader

Please consider the relevant viewpoint before commenting. This is about the reliability and capabilities of Safari as a target for websites and web applications from a developer point of view, not the reliability from a user point of view (although, being a Safari user myself, I can also definitely say it's not "rock solid" in that regard either).

From a developer point of view, Safari is widely lagging behind other engines in terms of features, and has been rigged with hundreds of show stopping bugs for years, including relative to very basic features like scrolling. As an example, as of today, it is still impossible to reliably prevent page scrolling in Safari on iOS. Don't tell web developers "Safari is rock solid"; for us it's the bane of our existence.

View on HN

I’m a web developer, and I use Safari almost exclusively. It’s not the bane of my existence, I find it pleasant to work in, the dev tools are good. Just wanted to offer a counterpoint, not all developers hate Safari! I find it to be rock solid.

Not all developers hated IE either.

For much of the early 2000’s, IE was the only decent browser.

That is a subject statement, many people I knew preferred Netscape and then Phoenix.

I find it hard to imagine anyone preferred Netscape 4.x to IE 5 and 5.5. I think even IE6 may have existed before Netscape 6, though I definitely made the switch to Phoenix as soon as I discovered it.

I certainly did. I used a lot of features in Netscape that IE never had, and at university it was _the_ browser used on campus in the late 90s.

IE 5 was a buggy browser lacking a lot of features, but was used often by people not building things online. And forced on people by bizarre site requirements. Though this came worse with IE6.

Microsoft was embroiled in legal issues of their browser at this time, and anyone with a sense of logic (most programmers) could see being locked into one piece of software instantly degrades the value of that software.

Sept 2002 was when Phoenix was released.[0]

IE 6 came out only a year before Phoenix.[1] And you if you were involved with Slashdot back in the day you were tracking browser usage as a daily trend. Phoenix was doing well.

[0] https://en.wikipedia.org/wiki/Firefox

[1] https://en.wikipedia.org/wiki/Internet_Explorer_6

IE had over 70% marketshare through 2005. In 2002, right after IE 6 was released, it was 85%. Everything else was in the weeds. https://public.tableau.com/app/profile/neil.the.analyst/viz/...

The difference from now is that back then there was an army of dedicated tech oriented people that _knew_ MS and IE were a problem that needed to be resisted.

I knew tech oriented people, myself included, that used a Windows desktops for a few years because the other browser alternatives were unacceptably bad.

Sure. As someone who had to build sites for both, IE was far superior. Vague css support and it didn’t crash if you nested tables more than seven layers deep.

I used Netscape through 4.x, up until 2000 or so. It took years before Phoenix / Firebird / Firefox was considered "stable" (I'd say Firefox 2.0?) There was a solid 4 or 5 years there where IE was your only real choice. Apple was even shipping IE with the Mac!

the biggest problem of safari (Apple ecosystem) is the same as IE or even worse, it is integrated into the system, and old devices like ipad do not have updates ...

Old iDevices who are stuck in a non-upgradeable Apple OS are a threat to humanity.

lolwut

A lot of the developers who rail against it are in the advertising industry and those who want to push notifications to you as well as run their garbage javascript in the background.

Or perhaps people who search caniuse see it's flaring omissions

> Apple produces a vastly inferior browser and only gets to maintain market share via absolute, monopolistic control over their OS

I must have missed the implied “for developers” preface there

> Don't tell web developers "Safari is rock solid"; for us it's the bane of our existence.

i've been a web dev for 20 years. safari is rock solid but has quirks, yes.

IE was "solid" as long as you only worked inside it's constraints.

So is Crome. That statement applies to everything.

It's not the same at all, IE/Safari are/were locked to their OS or abused through monopoly.

iOS/Safari is worse than IE because it's exclusive. At least with IE/Windows you could install something else.

OK, if you really think Safari is worse than dealing with IE hacks you probably were not writing anything for the web back then. This was already a tenuous argument, but come on.

OS lock-in and slow adoption of new web features is perhaps lame but Safari doesn't have anything even approaching it's own box model, or `-ms-filter` equivalent, or quirks mode... the list approaches infinity.

Also, "install something else?" Do you know what the browser landscape looked like back then? IE had >80% dominance for years, and your other option was Netscape, which lacked huge swaths of custom features in IE.

If the one key feature in an iOS webapp _kills_ your business, then _yes_ iOS/Safari is magnitudes worse than IE ever was.

The expectations and businesses relying on browser functionality back then is minimal compared to today which amplifies the core issues at hand.

Try building an PWA with offline capabilities (trivial in modern browsers) that's your core of your business, and you will see what I see. Apple is _trying_ to strangle non-app store apps. Where IE was just a dumb lumbering problem by comparison.

> Try building an PWA with offline capabilities

I have, for one of the world's largest companies. You're missing the forest for the trees if you feel PWAs are this much of a linchpin to the broader conversation.

The only programmers that had issues with IE were when IE broke something they needed, it hardly matters what it was. Apples to apples comparisons are the relevant issue.

So if your PWA works on iOS/Safari, great for you, but it doesn't mean everyone's does.

Notice, none of these reasons why Safari is worse have anything to do with the technology whatsoever. They're all ideological.

Try using offline first features that all browsers support properly except iOS Safari. Pure tech issue.

I'm not going to argue there are no technological issues, I was just pointing out a feature of the discourse here.

If you've checked mdn or caniuse anytime in the last 2-3 years, you'd notice that Safari has been leading the gang in standard web feature adoption.

Safari offers the most refined developer experience for web dev and testing (my favourite are the automation capabilities which are incredible and private as it separates my data from test data during a run). With technology preview updates delivered directly in macOS, I'm always confident I have the latest setup to test my web apps...

https://wpt.fyi/

I really like Safari but no, they are definitely not "leading the gang" in adopting features.

Look at the front page of https://caniuse.com/

At the bottom right there is a chart and Safari is in 3rd.

I built a company based on browser testing and Safari in general was/is always behind.

They have hired some good people so maybe that situation will change but in general Safari is the currently the boat anchor of browsers. Hopefully the situation will change over time.

This just isn't true! There are so many features that have not been supported long after other browsers implemented them (webgl 2 comes to mind, though that is finally available - 4% of global web users still don't have support due to this, though!)

Exactly: Install Prompts - 7 Years. Fullscreen API - 11 Years. Push API - 7 Years. Badging - 5 years. Screen Orientation Lock - 11 years. Bluetooth - 5 years.

Nope, nope, definitely nope, not critical in the slightest, nope, and sounds like a very niche case that a native app would do just fine with.

"A native app would do just fine", why would we want to build a native app using proprietary tech, be forced to go through an app store, risk to be turn down by Apple and be taxed 30% of our revenue otherwise? All this when we can build web apps using open standards, working on all platforms and without being extorted by a giant abusive conglomerate?

You would want to because you want access to Apple's customers. Also, your example doesn't sound like it'd hold up to real world practice. If you're selling someone proprietary tech then your revenue isn't subject to 30% of Apple's fee. Only the software portion done through the app is. You obviously have a motivated buyer at this point if they bought a physical device. You could route them to a website to create an account, sign up for any subscriptions and launch the startup process that eventually has them install your native app. You want what you want, but you describe hyperbole. Plenty of apps have users manage subscriptions on a website, not on native apps.

You've got be joking right? ok so lets go through some stats:

1. Safari, performs worst on wpt.fyi by a VERY wide margin 2. In State of CSS last year had the most number of complaints in relation to bugs. 360 related to Safari, compared to 30 for chrome and 10 for firefox. 3. In the MDN developer Survey developers ranked Safari as the second biggest problem after IE. A browser that's no longer supported. 3. Doesn't really support Web Apps, no user findable method of install (install prompts), no push notifications. 4. Huge numbers of bugs (some of which haven't been fixed in over 7 years) 5. and the devtools are still far behind chromes. 6. and missing all of these: Notifications / Push API No AppStore Support App Store Support for Web Apps Fullscreen API Badging Deep Links Screen Orientation Lock Bluetooth NFC Web Share Target Screen Wakelock Keyboard Lock and Keyboard Layout APIs getInstalledRelatedApps() Background Fetch API Background Sync API WebGPU WebXR Offscreen Canvas Declarative Shadow DOM SQL SIMD WASM Threads Shared Array Buffers AV1/AVIF and VP8/VP9/WebP (Open Media Codecs) Compression Streams Scoped Custom Element Registry ElementInternals

> Bluetooth NFC

Chrome-only non-standards that both Firefox and Safari are against.

> Declarative Shadow DOM

Actually proposed by Safari, and stalled for multiple reasons. Moreover, at the start of the whole web component brouhaha Safari wanted to have declarative ways of specifying Web Components, but Chrome said "no, we want to move fast and break things"

> WebGPU

Proposed by Safari

---

Can't be bothered to check others, but off the top of my head, but most of other standards that you just dumped without understanding fall under the following categories:

- Chrome-only non-standards with multiple issues that have multiple objections from both Safari and Firefox

- Originally proposed by Safari

- Have specific technical objections from Safari, and specific reasons for not implementing

- (very small minority of specs) No idea why they haven't implemented these yet

Would you point me to Apple's alternative implementation for access bluetooth / nfc? That's right it doesn't exist. Instead they offer for you to install a native app which once granted bluetooth access can continuously scan and connect to any bluetooth device they want.

- Install Prompts 7 Years behind. Many thousands of requests. - Notifications 7 years behind. - Fullscreen API 11 years behind. - Badging 5 years behind. - Deep Links 7 years behind. - Screen Orientation Lock 10 years behind.

and what's your excuse going to be the myriad of bugs/stability issues?

Again. People who ACTUALLY build apps in Safari are the ones that run into these issues. The only people that attempt to defend it are the ones who don't even build apps for it.

> Would you point me to Apple's alternative implementation for access bluetooth / nfc?

1. This is clearly moving goalposts

2. Why would Apple or Firefox provide you with an "alternative implementation for access bluetooth / nfc"?

> Instead they offer for you to install

No. Instead both they and Mozilla consider these non-standrads harmful and have no intention of implementing them. Or are you going to pretend that Mozilla is in Apple's pocket and does this to enrich Apple?

I'm not going to react to yet another mindless list of things that you have little knowledge of.

1. If you're going to say you don't like spec because of issues, you should explicitly describe what those issues are OR offer up an alternative spec. Simply not engaging on the spec doesn't present as engineers trying to solve a problem.

2. Neither Apple or Mozilla have described in detail what they think the harms are and why the specs do not cover those harms. But Mozilla is not a gatekeeper and doesn't ban the competition from 50% of mobile devices, so IMO they should be free to decide to do whatever they want even if I disagree with their decision.

Apple on the other hand not only doesn't describe the security issues in ANY detail, or outline why the specs don't cover them, it bans the competition from bringing any solution to the problem while at the same time providing a significantly insecure solution on native which is widely used to track and surveil users and for 9 years didn't even require a permission prompt.

As for resorting to insults about mindless list of things, have you even attempted to build a web app for iOS?

In every single case of the most ardent Safari defenders the answer is no. Unsurprisingly the people that care about the issues in Safari are the ones with the most knowledge that have to work with it daily.

> If you're going to say you don't like spec because of issues, you should explicitly describe what those issues are

They did

> Neither Apple or Mozilla have described in detail what they think the harms are

They did

> As for resorting to insults about mindless list of things, have you even attempted to build a web app for iOS

Doesn't explain why you keep dumping these lists without thinking and parroting Google's talking points

I'm not joking.

You can see other replies for details lists of functionalities and capabilities often pioneered and supported by Safari long before other browsers (even Firefox).

Some loud voices from the community have very selective memory, and they momentarily forget all the evilness pumped into Chrome regularly. (remember FLoCs? Advertising API? AMP?)

All that hate for Safari can be much more productive if channeled as constructive feedback on browsers' bug trackers and governance discussions, improving the ecosystem as a whole (which is what you want if you care for the web). Such conversations can even lead to the natural evolution of standards and implementations towards more refined engine customizations on each platform.

Today, iOS allows for open browser choice on iOS, so any technical requirements for the implementation are of no meaningful consequence to iOS users. Everyone can get the look, feel, and service integration with their preferred vendor's ecosystem. Consumers and developers also get a consistent experience throughout the entire operating system - a piece of content will appear the same in any context. Simple, efficient, and reliable. All this will be lost the moment someone decides to prop a "To view this website, download browser XYZ." Imagine you are faced with this in a stressful situation when you need your phone to work the most.

iOS does not allow for any meaningful choice, it’s all basically Safari.

Says who?

Mozilla, Google, Microsoft, EFF, the uk regulator, the Japanese regulator, the EU and most importantly the group of people who actually build web software on iOS.

As for feedback. we gave up after 10 years. That’s literally how long we’d been asking for notifications.

> Some loud voices from the community have very selective memory, and they momentarily forget all the evilness pumped into Chrome regularly. (remember FLoCs? Advertising API? AMP?)

Hell, remember logging users into Chrome without permission because they logged into a Google website like Gmail and then "accidentally" synching all the user's bookmarks and browser history to Google?

from a developer point of view, i'm about to give my team a talk on container queries + subgrid, and safari is the only browser i can use to demo both.

not sure what your scrolling issue is, hidden overflow and blocked pointer events have always worked just fine for me. typically when someone gets scrolling wrong they're using 100vh instead of stretch and/or not scoping overflow correctly

Container Queries and Subgrid are only available in Safari Technology Preview, not stable, and Firefox has actually been supporting Subgrid for more than 2 years. Because CQ is not supported by either Chrome of Firefox, it will be at least 2 years before we can start actually using it.

It would have been much wiser for Safari to catch up on the dozens of features they don't support that both Chrome and Firefox do, or to focus on bug fixes for the most basic features that's been broken for years. Instead, they chose to ship shiny new ones to try and convince both regulators (from the EU, UK, US, etc) and web developers that they are leading the way in feature adoption. Unfortunately this seems to be working to some extent in the web devs community. Regulators are unlikely to fall for it though.

Here is the 7 years old scrolling bug I'm referring to, which prevents any decent implementation of modals in Safari on iOS: https://github.com/web-platform-tests/interop-2022/issues/84

Here you can see that Safari has 5 times more API failures (representative of both missing features and bugs) than Chrome, and 3 times more than Firefox: https://wpt.fyi/

Original author of the post here. I'm also a Blink API OWNER, so can happily report that Container Queries have 3 LGTMs and will ship in Chromium browsers very shortly, likely 105 (shipping to Stable in September):

https://chromestatus.com/feature/6525308435955712

Subgrid is progressing thanks to engineering from the Edge team, and I expect it to land this year.

I think the user perspective is way more important. I think users get a way better experience using Safari

> it is still impossible to reliably prevent page scrolling in Safari on iOS.

why would it?

it's a USER-Agent after all. if there is something to scroll, the user probably wants to and should be able to.

Professional web developer here: I use safari almost exclusively and I have for years. I vastly prefer it to chrome/Firefox/brave/opera/etc.

Sure, and from the point of view of thieves the locks I install are bad.

Maybe, just maybe, people should stop trying to develop webapps and instead make websites that just work without javascript. Like this one.

Nobody in their right mind would ever want to zoom & pan a map on a website, eh?

> Nobody in their right mind would ever want to zoom & pan a map on a website

1) I agree. Apps have maps, but most websites?

2) Zooming and panning can be done without WPA. At most some basic JS, although I think you can do the same via HTML+CSS.

3) The old way of zooming panning was buttons for zooming and panning. Those worked fine. In face ,they worked better than zooming and panning buggy implementations.

You might be right if one fix one mind on developer and latest greatest. But apple preposition at least for user might not be that. We n fact, I thought the whole point of browser is serving the user.

Until m1 I am not sure apple has the intel experience considering the hackintosh. But apple dominate for a while the developer market.

The question imho is whether apple provide a good enough solution for its users. And for monopoly I am not sure why os pre-installed browser is that relevant. Remember why edge exist. Just to download chrome.

Btw I use both as there are things can be easily done by one cannot by the others. I hope we have more. Not less.

Actually, didn't the latest build of Safari nightly recently over take all the other browsers in terms of number of supported features? I swear that was shared here on HN.

It 100% did not in that universe. Safari is massively lagging behind, and has been for so long it will certainly be years before it catches up with Chromium or Firefox.

it 50% did. safari and chrome are trading blows, with safari generally ahead on useful css and chrome generally ahead on useful api

https://caniuse.com/?compare=chrome+106,safari+16.0&compareC...

The caniuse data shows that the number of APIs Chrome supports that Safari doesn't is at least 3 times higher than the opposite, and among the APIs Safari supports that Chrome don't, most are either too old and useless, too new and equally useless, or very small compared to the massive APIs listed to Chrome advantage.

I really don't see how that translates to "Safari and Chrome trading blows" and Safari taking over Chrome by "50%"?

Also, here is more representative dataset, where you can see that Safari has more than 5 times the number of API failures compared to Chrome, and 3 times compared to Firefox (in the stable channel): https://wpt.fyi/results/?label=master&label=experimental&ali...

oh cool. hey quick question, how do i do a hanging quote in chrome?

How many of those Chrome APIs are privacy misfeatures? I suspect at least half of them.

It helps that Jen Simmons, a renown css evangelist and expert, works on safari.

No. Safari is miles behind. https://wpt.fyi/results/?label=experimental&label=master&ali...

Interop2022 is probably what your referring too but that’s just a small (but still important) functionality that the browsers have agreed to focus on. Interop is great, but using it as the primary metric for your browser is very dishonest. Apple tried to do this in their regulatory filing but every other vendor and developers said that the overall stats are much more representative.

Reply on news.ycombinator.com