Push notifications are useful in browsers in all the same places they are useful in mobile apps. Note that browser push notifications also work on mobile, so you don't need to install an app to get real time alerts.
For example:
- Sites with chat can send you messages even if you don't have the tab open (or even browser in some cases)
- Online game e.g. board game or chess can push to you when it's your move.
- Home alerts, e.g. that your garage door has been opened
Sure there are a lot of spammy notifications, e.g. sales alerts, news items, etc. that most people probably don't want, but there are genuinely useful use cases.
I turn off/no to push notifications on every app that has ever asked for it, except Messaging and Calendar. Nothing else I've ever installed has warranted more than a little red notification dot on the icon.
Why cannot it be automatically "no" for all push notifications? A website should not even be allowed to ask me.
MS Teams already gives me a lot of push notifications during my work day, and if the browsers starts to do this, I will be overwhelmed with beeping nag-notifications.
Computer clutter is a distraction and lowers productivity.
IIRC by default Firefox doesn't prompt you when sites request to send you notifications. But it's possible, i just set some configuration a long time ago and forgot about it
This is the way. In addition to those, I have a silent e-mail count badge for Gmail and WSJ news alerts (because they are tangentially relevant for my job)
All the same things they’re useful for in apps (which are also prone to spam/clickbait).
Web push is a prerequisite for moving a ton of apps out of app stores and into the platform agnostic thin client model of the web: messaging clients, timers, background tasks, transaction/shipping status updates, trading tools, etc
Although apps also send junk push notifications, it seems like websites are drastically worse in terms of push notifications.
The reason is just that I visit many more websites only once and expect no ongoing relationship with them: that means that I am much less likely to want their notifications AND they are much more likely to want me to have notifications AND they have less to lose by being spammy.
Apps can assume higher organic revisits and also being too spammy can result in the downside of the me the app. A website I had no plan to revisit has no punishment for attempting not spam.
You mean not installing yet another application and having another icon on my homescreen for every freaking website I visit? And these apps almost have no need to run any native code or do anything a website can't handle in modern day. Yes. I'm all for it. Let's make browsers more capable.
Or maybe it’s because either cross platform web apps are never as good as mobile apps or that the entire web development ecosystem is a clusterfuck of dependency hell, security vulnerabilities in dependencies, complexity, and “ooh shiny” new framework of the week?
Have you thought about not installing the apps and just using the website? I have yet to see a website where I wanted to install the app to view it, and have no idea what features would make me want to do that.
Sorry, I must have miscommunication. I said "I have no idea what features would make me want to do that." And I don't. I've never wanted push notifications from any website, I'm happy that those that do are segregated to the app and that the website works fine without them. Unlike, say cookies or tracking or ads, where I have to modify the code of the website to block them.
My point was maybe they should try just not installing the apps and not getting push notifications, etc.
Have you not enountered sites that app-wall you? Sites with useful content, such as Reddit - not allowing you to view the user-generated content unless you install their app? It sucks - but it exists.
It does this under the guise of protecting you from certain content, like an nsfw post will require you to download the app, although I think you can get round that using the old. prefix rather than www
Reddit, for instance, should not be an app. It is failure on the browser's capabilities that you can't visit a reddit link without being spammed to death about installing their app. If the browser was capable of the things the reddit product team wanted to do, they wouldn't need to bug you to install their silly web wrapper app. The OS would be able to handle permissions gracefully, instead of letting the app have free reign.
> If the browser was capable of the things the reddit product team wanted to do, they wouldn't need to bug you to install their silly web wrapper app.
I don’t want the things the developers want. Reddit’s mistake is thinking that I want anything more than headlines, thumbnails, and interstitial advertising.
> If the browser was capable of the things the reddit product team wanted to do, they wouldn't need to bug you to install their silly web wrapper app.
Reddit is text with images. That is literally the core of what web is.
The only reason Reddit breaks their web version and pushes you towards the app is because Reddit wants to monetize the hell out of reddit, and doing that through the app is easier.
I disagree, Reddit does well as an app (I’m of the option that HackerNews is too , that’s why in use Octal). It’s just that Reddit has a really bad app.
Basically any situation that is some combination of in-person and online. Sure, you can build and maintain an app and then force people to install it to receive notifications. Or you can use/abuse SMS and force people to give their (highly targetable) cellphone number. But both of those alternatives cost more and are less consumer-friendly.
Personally I've found notifications to be really useful for applications which I don't want to install from external sources, but must still use for work. These include MS-teams, slack, rocket, Whapps, GMail.
Or people will get so inundated with spammy notifications on websites asking for permission to send notifications they will just disable them globally (like I will).
I work on a lot of conference and festival web sites, where there are multiple stages and events going on at the same time. We can do pretty much everything in the browser, and the users are happy with not having to download an app. The only thing that has made us look into native development is to have push notifications. Our clients wants to send push notifications if an event is delayed or if an event is at another stage, after being changed last minute, which happens pretty often.
I run a free service[1] that lets you relay messages to a WebPush notification by hitting an HTTPS endpoint. It was originally intended for tracking the progress of ML training jobs, but people have used it for notifying themselves when event tickets became available, tracking prices, getting events from Raspberry Pi home automation devices, etc. I’m glad it’s finally coming to iOS!
It would be nice if mobile browsers had push notifications in IOS because it would make it possible to have browser based mobile apps with some chance of feature parity.
- Several sites/forums with comment sections, to notify me of replies etc.
- Gitlab pipeline notifications
- Some smaller services I'm probably forgetting
Basically, anything but the scammy sites that try to force notifications upon you. Generally interactive stuff which I feel shouldn't have a 300MB app with access to file systems (and there's not a lot that _really_ needs an app these days).
To combat the blogspam from Google searches, HN/Reddit links, and whatever else tries to shove stuff down my throat, I have notifications disabled by default so I don't even see the prompts; I manually toggle them on, and I think that should be the default. I also close any website the second I see a popup for newsletters, notifications, or even cookie popups the second they block my reading, unless I can't find the information I need anywhere else.
> Every time I've seen them in action was with spam/clickbait, after someone accidentally subscribed and couldn't remove it.
I think this is the reason Apple is reluctant to add it. There are very valid reasons why a user would opt to receive web-push notifications, but its mostly niche apps/usages. The vast majority of push authorization banners are from sites you visit for the first time, with hardly any engagement or interaction nor clear intention of allowing web-push ("We think you want unsolicited notifications because it is better for our business model")
While there is no fix for dark-pattern usage of web-push, browsers and OS UX must make it super easy to withdraw notification consent anytime.
I'm not sure if it qualifies as useful, but a while back I hacked together a web push notifications site[1] for RPAN live streamers after Reddit ripped away in-site notifications. It didn't gain traction because the admins/mods banned and forbade me from speaking about it, and because I couldn't offer iOS notifications, but ~1,000 people still use it to get ~25,000 notifications every day.
For the first several months, I was getting a message or email every day asking when iOS web push was coming, and the timing couldn't be more perfect that Reddit just effectively shut down their live streaming the same day that Apple finally enabled push.
I think it's been well answered, but I feel it's a meaningful technology if web apps are to be a real consideration. There's an over-reliance on email being a notification tool, with plenty of examples such as social media sites which utterly abuse it.
The difference for me is straightforward: Email is for messages the user might want to keep, notifications are messages that the user doesn't need to keep. (To extend this further, messaging systems aren't for notifications either, but rather for interactive dialogue.)
I worked on a web app once where we used them to notify the user when a job they'd started was done. They've also been put to good use in email and calendar apps
To be fair, even smartphone apps these days will send you BS notifications.
Shopping apps will ping you throughout the day with "recommended" products, ride share apps will ask you to ride with them during the holidays, social apps will ask you to checkout the newest trend.
All sad attempts to get you to open their app. The worst is when apps don't even let you opt out, so you either get useful notifications and advertisements or nothing.
So we can have open source chat apps again. Only organizations can publish chat apps for the app store and it's resulted in a massive regression in IM app quality over the past decade.
Either they're not free (which matters when setting up group chats) or they don't have reliable push notifications (because only the group publishing the app can maintain the push infrastructure (which is required) and no one is paying them and it's a lot of work.)
It is actually completely fucked. I've looked into dealing with it myself. Apple has pretty much single-handedly destroyed internet IM. Hopefully this starts to undo the damage.
Yeah they're almost all siloed or they have the problems I've mentioned. The way Apple has decided app architecture must work forces this on everyone. I've spent a bunch of time (way more than I would have liked) looking into this and I've had this exact same discussion a dozen different times.
EDIT: I'm out of comment quota. Being open source hardly matters if the app is siloed. There's no reason we shouldn't be able to have decent XMPP and IRC clients other than Apple jerking their users around. (and before you bitch at me about leaving a socket open for push notifications using up battery, no. That's what the OS is already doing, it suspends the machine without closing the socket and reopens it if it's dead after the next wakeup.)
On top of that there's the issue of at least some of the tech that open source likes to use. Prime example is XMPP which 10 years after introduction of the iPhone still didn't have a compelling mobile story [1]
10 more years later, and yeah. I think the lack of open source messengers is a combination of many things and whatever Apple forced is just a part of it.
Every comment suggesting it brings browsers closer to feature parity with apps is a great argument against the feature. Browsers aren't operating systems. Stop trying to make them be that. It has been a terrible disaster so far, and likely will continue to be going forward, given the lingua franca of development in browsers and all the terrible development practices that have grown up around it.
So what type of app are you writing that has to accept in app purchases and are push notifications the only thing stopping you from having a much worse web app than a native app?
Do they have to be? Who knows? But since 2007, Apple, Palm, Blackberry and Microsoft have all at one point or the other have said you don’t need native mobile apps and the web was good enough. It never was.
Facebook tried the whole cross platform web app in a wrapper and decided it wasn’t good enough. Even while Google is extolling then virtues of its own cross platform tooling, it’s moving toward apps that use native frameworks.
Ahh yes, the infamous Facebook failure. Anyone can write a bad app. That's not necessarily a technology failure. But it's easier to blame the tech than to admit they did something badly.
Everyone wants web apps to help sell hardware, right up until your app store grows into a money maker. Then the incentives are different.
Regardless, today's mobile systems are much better, web browsers are much better. It is much easier to write a web app that is functionally identical to native than it used to be.
Modern reality. Conversation rates for PWAs are higher than mobile apps because there is less friction in the installation process. Load times are often lower, which can be an overall superior experience. If I'm a small shop, trying to get on someone's mobile device, these items matter a lot. Also development costs are much lower when you can share code for your mobile app and your website.
The value add of being in the mobile stores is pretty minimal. Discovery is poor, search is poor, then there is constant churn in policies, submission process, APIs, development environments/languages, fees, etc. In contrast to this, the browsers hold very very high backwards compatibility. Churn will be in the support libraries or framework you adopt and not fundamental to the platform.
2. Developers were clamoring for native apps each time. Not the mobile platforms. You act as if the modern web development environment isn’t a clusterfuck of complexity compared to modern IDEs for native apps.
3. Both Apple and Google have initiatives where you can have small instantly installable “applet” equivalents that make downloads fast and the apps are more responsive.
4. Cost is lower for the developer. But at the cost of a much worse user experience.
5. Browsers might have “higher backwards compatibility”. But the dependency hell and ever changing landscape of the front end framework of the week is real.
1) That's a totally different number... Apples and oranges. Conversation in this case is getting an icon on the users phone as opposed to sale. Mobile stores have really really low conversion rates.
2) Assumes facts not in evidence. This is not a universal opinion.
3) Still has the gate keepers, yet another platform to write for, etc etc. Lower friction might help vs PWAs, that's something at least.
4) Assumes facts not in evidence.
5) Showing your ignorance here.
Front end has been stable for a while. Old stuff still works if you prefer it. Lots and lots of sites still do.
The mobile shill is real... There are reasons to pick mobile over something else. But just shilling the tool and showing your tech bigotry isn't a good way to pick solutions. I'm out.
2. Really? Were you not around when they called Steve Jobs “pretty sweet solution” of web apps a “shit sandwich” in 2007, or when developers complained about non native SDKs on both WebOS and RIMs newer platforms? Are you really claiming that native apps are not more performant than web apps?
3. And those “gatekeepers” allow more monetization than the web ever has.
4. Are you saying that I’m not constantly getting GitHub messages about a security vulnerability somewhere deep down in a dependency chain for an open source project that I contribute to? It’s a popular company sponsored open source project.
The "browsers aren't operating systems" ship has sailed long ago, exhibit a: electron apps.
Think of it this way. When feature parity of browser sandbox closes on everything a desktop app might want, you no longer need to build with electron but can simply ship 5-15MB of js+css and use the stock browser instead.
Similar on mobile.
Which will result in:
Smaller downloads.
Lower memory footprint.
Faster launch.
The more low-level access browsers are required to expose for application development, the worse things will be in almost every dimension, but especially security and performance.
I'm curious, how does adding notifications effect security negatively exactly? Performance?
All this functionality is already available for apps including the security model. Nothing new is happening here. App security is an issue, but it's one big tech is already dumping resources into solving.
But, frankly, browsers have some of the best management features for culling analytics, telemetry, watching the traffic and requests being made. It's many times easier for me to see what a web app is doing on the network, and write something to stop it, as compared to a native app.
The threads are essentially the same because the story is the same so it ends up being an HN dupe. It was also discussed in a number of the other Safari-related WWDC posts.
This is support for web push notifications in safari it seems. (https://web.dev/push-notifications-overview/) - and it plays nice with macs. I guess it is useful if you want to support Safari for your webapp + use push notifications.
Yes – GP is claiming, as far as I understand it, that this is already the case, but their source does not support that claim, and I don't believe it's the case.
Wow. I missed this a few days ago, but it seems like huge news for the web. Am I missing some important caveat or can web developers get excited about this?
Apple’s long resistance to allowing ‘web push’ on iOS has held back PWA adoption. And they’ve obviously never admitted it, but this policy protected the hegemony of the App Store, and therefore Apple’s control over the whole ecosystem. The absence of web push on arguably the world’s most important web browser (iOS Safari) was a feature for them. It constrained PWAs to niche, toy use cases - if you wanted to make a real app with notifications (like most apps) and you wanted it to be available to iPhone owners, you had no choice but to follow Apple’s technical and content restrictions and generally play to their fiddle and be very polite to them.
So with that incentive structure in place, I didn’t hold out much hope for the situation ever changing except possibly through some anti-trust thing.
But now it’s happening. Why? What does Apple get out of this move in return for the loss of control it is likely to mean for them? Can anyone shed any light? Perhaps it’s about staying a couple of steps ahead of antitrust suits they can see coming down the line?
The only caveat I can see is it requires a $100/year developer certificate. That doesn’t sound like a major caveat to me. That cost is nothing if it means you can build one app that runs everywhere.
I guess maybe Apple could still revoke your certificate if they don’t like what you’re doing with your web app, ie hamstring your webapp by taking away your push notification ability. But I’m guessing (hoping) this would be a very rare hostile ‘intervention’ compared to the common experience today of every app update being subjected to review by Apple and possibly rejected for arcane reasons.
A few years back we had this same discussion orbiting around Service Workers. Here's a post from 2017 where the top comment argues that what's holding back the web is the lack of Service Workers in Safari: https://news.ycombinator.com/item?id=14489577
Apple shipped Service Workers in WebKit in 2021. Are web apps much better now than in 2017? Frankly it seems about the same. The best web apps are still exercises in minimalism like Wordle, which don't need Service Workers or Web Push.
I'd argue the lack of service workers was indeed holding it back, but so was the lack of push notifications. Both are vital components for webapps to be able to compete with native apps for common use cases.
> The only caveat I can see is it requires a $100/year developer certificate
Does it? I know it did for sending notifications in Safari on macOS, but I think this time it does not. See [0]: "You don’t need to join the Apple Developer Program to send web push notifications."
[0]: The only caveat I can see is it requires a $100/year developer certificate
"… today I'm happy to announce that we have added support for Web Push, and this really is Web Push! The same combination of various web standards as implemented in other browsers. We'll go over those standards more later, but… the most important takeaway is that if you've coded your application to web standards, you won't need to make any changes for it to work in Safari."
"… no Apple Developer account is required to reach Safari users."
Dumb questions about web push: would one see a gradual uptick in resources consumed by webworkers as they accumulate notifications from different sites? Would they keep the antenna/modem on a phone awake constantly?
The big advantage of the current centralized model is that it's easier to coalesce notifications and the OS only has to maintain a single connection. Would hate to see my battery take a hit as a result of a move away from that model.
Web push notifications are also delivered through APNS, the service worker only gets woken up once there is a notification, it doesn't poll, so that shouldn't be much different from a regular app.
Seems I found a bug. I can't use the spacebar to pause a video in full screen mode. I tried another video on the developer page and I face the same issue.
More than Apple (which prevented the feature from being available on iOS for 7+ years and still hasn't shipped it) you should probably thank the regulators who are investigating Apple's anti-competitive practices and the web developers who warned about them! This is the result of the pressure of the upcoming regulations around browser engines on iOS and the anticipated resulting competition for Safari.
Expectation: seamless push notifications without a dedicated app!
Reality: users blindly click "do not allow" push notification dialog the way they already do with browser notification permission dialogs, location access, "allow cookies" banners etc
Like any permission request, if you ask when the user attempts something that requires it, they'll accept. If you ask when the page loads, they'll (rightly) deny it.
If they can't figure out how to use their phone that's not my problem. I just want to be able to use decent chat apps/services again without half the people in the group complaining bout their iPhones.
Web push would be great for emails or transit updates, but most of the time there is an app. Also, I don't want to be disturbed at night or work. So I end up blocking everything except messages.
Apps suck. iOS is a treadmill and you have to maintain app specific infrastructure to send push notifications with them. It's very expensive (in terms of time) so volunteers don't do it. This pretty much guarantees any apps with push notifications are owned by some kind of largish organization. It's pretty terrible.
Hopefully this either moves all of that to the web or forces Apple to deal with the problems with their platform.
Every time I've seen them in action was with spam/clickbait, after someone accidentally subscribed and couldn't remove it.