[HN Gopher] Push Notifications, WebXR, and better PWA support co...
       ___________________________________________________________________
        
       Push Notifications, WebXR, and better PWA support coming to iOS
        
       Author : carlycue
       Score  : 228 points
       Date   : 2022-02-01 10:18 UTC (12 hours ago)
        
 (HTM) web link (firt.dev)
 (TXT) w3m dump (firt.dev)
        
       | NicoJuicy wrote:
       | I was going to write 10 years too late, but I decided to look it
       | up.
       | 
       | I'm proud to announce i was wrong. It's actually 15 years too
       | late, since it was promised in 2007.
       | 
       | https://en.m.wikipedia.org/wiki/Progressive_web_application
       | 
       | I suppose increased regularisation worldwide has something to do
       | with it to announce it again.
       | 
       | I'm curious which features they will lag now...
        
       | rabuse wrote:
       | I've waited years for push notifications to come to iOS Safari.
       | This is wonderful news.
        
         | kecupochren wrote:
         | Same. I'm beyond excited. I thought we would never get them
         | since Apple wants their App Store cut. Today is a good day
        
       | maga wrote:
       | Sounds almost too good to be true, but here is hoping for the
       | best.
       | 
       | > Opus audio decoder (enabled)
       | 
       | Considering the track record of codec support, I bet it only
       | works when packaged as CAF, singed by QuickTime, and played
       | during the daytime to a limited audience... I joke, but the
       | amount of gotchas I encounter while working with WebAudio API is
       | mindboggling and reminiscent of good ol' days of IE6. Would you
       | guess that despite all three browsers declaring support for audio
       | recording (MediaSource API) and all three supporting opus on
       | paper, there is no way to produce the same opus file in all
       | three? Safari only supports playing opus in <audio> tag, no
       | recording; Chrome doesn't support ogg, and while both Chrome and
       | FF allow you to record opus in webm container, the resulting file
       | is not seekable... There are of course third-party solutions, but
       | it's just sad to watch in the century of the fruitbat.
        
         | jtbayly wrote:
         | > in the century of the fruitbat
         | 
         | Off topic, I know, but what in the world does this mean? I did
         | a quick search and couldn't figure it out.
        
           | maga wrote:
           | A common reference to the modern (ergo enlightened) age in
           | Terry Pratchett's Discworld series:
           | https://wiki.lspace.org/Century_of_the_Fruitbat
           | 
           | Sorry for the confusion, I'm high on caffeine.
        
           | dagw wrote:
           | It's a Terry Pratchett/Discworld reference.
        
         | istjohn wrote:
         | I just spent the weekend banging my head against the wall
         | trying to record and upload audio in Safari. Chrome and FF were
         | relatively easy. Know of any good open source examples or
         | documentation on how to do this?
         | 
         | It really is sad that something so simple is so hard. I refuse
         | to pay 10 cents a minute to a third party.
        
           | maga wrote:
           | In the end, I went with this library for now:
           | https://github.com/chris-rudmin/opus-recorder
           | 
           | There are issues reported with the latest iOS, though.
        
       | moron4hire wrote:
       | I really hope the WebXR API stuff will come with support for VR
       | headsets in Safari on desktop, and that whatever Apple's rumored
       | VR headset end up being will also support it.
       | 
       | WebXR has been great for my app at work, but we're basically
       | locked into the Quest 2 right now because Mozilla dropped the
       | ball on WebXR support. Oculus is the only one doing significant
       | VR work with WebXR in their Oculus Browser, which is a Chromium
       | build that they keep proprietary.
        
         | jimmySixDOF wrote:
         | There is also no mention yet on Apple support for glTF file
         | types which is important to keeping an open web based XR
         | ecosystem.
        
           | ngokevin wrote:
           | glTF can be rendered by the libraries, I think native glTF
           | support only matters if you want to display it isolated and
           | natively in the browser with nothing else, like an image.
           | Which is not really a use case for VR.
        
           | moron4hire wrote:
           | There's really no need for that. WebGL and WebGPU have no
           | concept of file types. Whether it's Wavefront OBJ or COLLADA
           | or glTF or some jankasaurus of JSON loaded out of a database,
           | WebGL isn't going to and shouldn't have to understand it
           | natively. It's on the app developer to translate that to
           | vertices, texture maps, etc. And that's going to change a lot
           | depending on what sort of abstraction layer you're using on
           | top of WebGL, e.g. Three.js or Babylon.js or whatever. If
           | anything, having native support for specific 3D interchange
           | formats would constrict and over complicate the available
           | options to app and 3D framework developers.
        
         | gfodor wrote:
         | Agree - WebXR is mostly off peoples' radars, but if at least
         | two vendors end up adopting it well, it will have _massive_
         | long term consequences (good ones) for the democratization of
         | VR and AR.
        
       | veesahni wrote:
       | TLDR: there's broken/non-functioning push notification support in
       | iOS 15 beta... and its disabled by default.
       | 
       | Nothing to be excited about yet.
        
         | jtbayly wrote:
         | Yes, but it is a _major_ change of direction or a _major_
         | accident for them to even put this in a beta build if they
         | still have no intention of adding it.
        
       | mwexler wrote:
       | Funny enough, apps were supposed to be just for Apple. The
       | original conception is that 3rd party tools, games, etc. for the
       | iPhone would just be webapps.
       | (https://9to5mac.com/2011/10/21/jobs-original-vision-for-the-...,
       | etc.)
       | 
       | A few adjustments to the model and here we are with lots of
       | native code and big money... but still, folks requesting updates
       | to the the browser to support more app-ish functionality. I'm
       | glad; the browser can do a lot and we kind of miss that in many
       | mobile contexts.
       | 
       | I also think the elusive "run almost anywhere" is viable in many
       | cases when the browser gets powerful enough. (Of course, I said
       | that about the JVM years ago, so what do I know?)
        
         | dbbk wrote:
         | It was also kinda weird because when they announced "the
         | solution is web apps!" they made a big deal that developers
         | could make their web apps just look exactly like native apps.
         | But they never actually released any SDK, stylesheets, or
         | anything to achieve that. In reality everyone had to mimic the
         | look through CSS on their own, and even write absurdly
         | complicated JavaScript libraries to emulate the scrolling.
        
         | tehbeard wrote:
         | I guess the main difference is JVM didn't have a workable UI vs
         | the web?
        
           | brimble wrote:
           | You had to install the JVM. It often broke or bugged you to
           | update. UI usually felt off or janky, and unresponsive. At
           | the time, Java applets were _huge_ compared to typical web
           | pages, and had crap performance. By the time hardware and
           | connections improved enough to make the last two things less
           | of a problem, JavaScript could do most of the useful things
           | the JVM could with fewer security risks (and hardware was
           | fast enough that that wasn 't _also_ so slow that no-one
           | would tolerate it, as it would have been earlier).
           | 
           | [EDIT] Oh, and it had to compete with Flash and Shockwave,
           | which had (barely) fewer issues with plugin breakage, _very_
           | nice authoring tools (still unmatched, AFAIK?), and, despite
           | having crap performance, still performed better than Java
           | applets. Plus a whole lot more fun things used them so normal
           | people had a motivation to install them.
        
           | simion314 wrote:
           | You could do a working GUI with the JVM. But most people
           | demanded Windows desktop apps, and for Windows only you could
           | use .Net and C# for a better experience. Linux community was
           | in big part against Java or Mono because of license and FUD
           | so Python was the winner for the higher level language there.
        
             | riskable wrote:
             | Firstly let me state that yes, you could make a GUI with
             | Java but it would suck because the GUI toolkits available
             | at the time were just downright _awful_ (and they still
             | are!). Want an app that 's super bloated _and_ looks awful
             | (no native controls)? Use Java!
             | 
             | Also, the Linux community wasn't against Java per se (not
             | until Oracle bought Sun anyway) it's just that writing Java
             | was a huge/complicated undertaking compared to say, Python
             | and managing JREs was a huge pain in the ass. The old joke
             | was, "By the time you have Eclipse open I will have
             | completed the program in Python."
             | 
             | Lastly, mono was _enormous_ (and still is)! Talk about
             | bloated! When mono was being most actively pushed on the
             | Linux community back in the day there was just no benefit:
             | For smaller things you had shell scripts and Python. For
             | larger programs you had C /C++, Java, and... Python :D. Why
             | would you go through the hassle (and it was a huge hassle)
             | of using mono when those other languages provided better
             | performance with much lower friction/barrier to entry? It
             | didn't have much of a "selling point" as it were.
        
           | olliej wrote:
           | why would the JVM be involved?
           | 
           | Literally the expectation was you'd be a web app - html + js.
           | 
           | Everyone said that wasn't acceptable so they added native
           | apps, and now everyone says native apps are terrible and
           | everything should be web apps. It seems like apple can't win.
        
         | feupan wrote:
         | Given how quickly the App Store was introduced (just one year
         | later) I'm betting that they never had the intention to go web-
         | app only. They probably only needed extra time to figure out
         | the experience for both developers and users (from native APIs
         | to permission model to monetization)
        
           | gjsman-1000 wrote:
           | Actually, incorrect. According to court documents released
           | from lawsuits, Apple in fact _did not_ plan for the App Store
           | initially, just like Steve Jobs said on stage. It was not
           | until a few months after the iPhone was released that some
           | internal employees convinced Steve to let it go ahead and
           | release an SDK. I can 't find the tweet containing a copy of
           | it, but the actual email where Steve gave the SDK a go is on
           | the public record.
        
             | mlajtos wrote:
             | https://www.macstories.net/linked/apple-captured-in-a-
             | single...
        
       | hestefisk wrote:
       | I really can't see many use cases for web notifications apart
       | from annoying sites trying to bug me. So I'm happy Apple haven't
       | supported it.
        
         | moron4hire wrote:
         | Do you have notifications turned off in every native app you
         | also use? The reason scummy linkspam websites are the "only"
         | ones using web notifications these days is specifically
         | _because_ it 's not a reliably useful feature. App developers
         | who want to make a good user experience around notifications
         | are instead pushed to making native apps instead.
         | 
         | Personally, I don't use the native Twitter app anymore. I use
         | the web version of Twitter, installed as an app, complete with
         | shortcuts in my launchers. I have web notifications turned on
         | and the experience is nearly identical to the native app. But
         | there are little details that I think make it a better
         | experience, like haptic feedback on mobile when clicking the
         | Like button, or being able to select subsections of text on the
         | screen. I have this on my Android phone, my Linux laptop, my
         | Windows desktop, and my Mac Mini. It's exactly the same
         | interface, everywhere, whereas the native app is ever so
         | slightly different from the web app.
         | 
         | I have a lot of sites that I've installed as apps on my
         | computers and it's great: SoundCloud, LiChess, HBO Max,
         | NetFlix, Azure Portal, TeamWork, a couple of different WebGL
         | games, my own management interface for my app at work. It's all
         | basically just a shortcut to the site, without the URL bar
         | chrome. I like it a lot more than using tabs.
        
         | kecupochren wrote:
         | It makes [0]PWAs much more viable. Not having push
         | notifications was a major blocker to make "almost native" PWAs.
         | I'm excited
         | 
         | [0] https://web.dev/what-are-pwas/
        
         | brimble wrote:
         | 100% of "use" of it I've seen on the desktop has been spammy
         | sites no sane person would ever want to receive notifications
         | from bugging me to enable them because I made the mistake of
         | clicking them in search results, and my wife accidentally
         | enabling notifications from random sites, getting spammed, and
         | having _no idea_ how they 've managed to do that or how to
         | disable it. That's why I've hoped Apple'd _never_ support them
         | on iOS, and wished they 'd never supported it on desktop.
         | 
         | As someone posted elsewhere in the thread, I hope Apple
         | restricts requests to sites you've added as an "app" already.
         | And I hope they find a way to keep sites from being able to
         | tell you've "installed" them, so actually-useful sites or ones
         | that I have _no choice_ but to use (work, kids ' school stuff)
         | can't refuse to operate until I do that (I dunno, maybe they
         | already do, I never "install" sites)
        
         | jtbayly wrote:
         | I really want it supported, but I would _also_ use it on most
         | sites just like the pop-overs, cookie permission prompts, auto-
         | play (and hover) videos, and paywalls--ie, to determine that I
         | really don 't care to use that site.
        
         | rabuse wrote:
         | My users would like to be informed when they receive
         | messages/notifications from my platform's PWA, so there's that.
        
       | asadlionpk wrote:
       | As a user, I hope Apple doesn't just enable notifications as it
       | is. I tried Android recently for 6 months and permission popups
       | on each site were really annoying!
        
         | judge2020 wrote:
         | Given Apple, it'll probably be limited to PWAs on the home
         | screen, aka. only if the user goes to the share sheet and hits
         | "add to home screen", where then it'll pop up the permission
         | for sending notifications.
        
           | asadlionpk wrote:
           | That would be awesome! But given Apple, they wouldn't want
           | people to circumvent their App Store. Push is probably the
           | biggest reason people still make apps.
        
           | olliej wrote:
           | which actually makes sense, as there is a fundamental
           | difference between "installed as an app" vs some random
           | website spamming users.
        
       | samwillis wrote:
       | It's lovely to see Safari on iOS getting some attention, but I
       | would love it if they fixed some of the long standing bugs such
       | as the incredibly annoying text selection / caret issues with
       | scrolling.
       | 
       | If you have a textarea or input inside a `overflow:scroll` and
       | scroll it out of view or have a `position:fixed` element over it,
       | the text caret or text selection hi-light remains visible
       | outside/above other elements. It's incredibly frustrating for
       | developers of PWA and apps built will frameworks touch as Ionic -
       | it result in text carets and selections being rendered over the
       | top of toolbars and menus.
       | 
       | Effectively the text caret and selection hi-light are drawn above
       | _all_ elements on the page with no detection as to if they should
       | be hidden behind an element or are outside a visible scroll area.
       | 
       | https://bugs.webkit.org/show_bug.cgi?id=138201
       | 
       | http://openradar.appspot.com/18819624
        
       | tehbeard wrote:
       | On the one hand, web push is a welcome sight and I really hope
       | they gate it a little (only allowing if you installed the pwa,
       | explicit permission et ) and had some solid spam reporting for
       | end users tied to apple dev. credentials to keep it a pleasant ux
       | for customers and not a spam channel.
       | 
       | On the other hand, not jazzed to have debug what they managed to
       | break in indexedDB this time round...
        
       | wdb wrote:
       | Yet to use/enable web notifications on iPhone. It's not something
       | I have been missing. Probably we will see the dreaded web
       | notifications popups even on iOS
        
       | lolive wrote:
       | So we can hack AR in JS, in the browser, with a reasonable
       | support in both Android and iOS ecosystems?
        
         | mlajtos wrote:
         | yup, https://github.com/pmndrs/react-xr
        
       | rgovostes wrote:
        
       | paulgb wrote:
       | I'm excited to see support for Web Push. I realize that most
       | people associate it with spammy sites asking for permission to
       | send you notifications you don't want, but I think it hasn't
       | really been given a chance because it lacks iOS support.
       | 
       | Some things that it enables:
       | 
       | - Low-friction asynchronous turn-based games (think "Words with
       | Friends" but you don't have to install an app)
       | 
       | - Single-purpose alerts (like CamelCamelCamel) or stock alerts
       | without cluttering your email inbox
       | 
       | - Easily scriptable notifications from your own code (I run
       | notify.run which provides an HTTPS-to-WebPush bridge so you can
       | send yourself a push notification from curl)
        
         | dbbk wrote:
         | If Apple do do this, which I'm not certain they will, I imagine
         | they'll be very restrictive on the permissions. Can only prompt
         | to be enabled from a direct user action (eg a button click),
         | and the latest iOS notifications prompt quite often if you want
         | to keep them or stop them.
        
         | jtbayly wrote:
         | Just to add:
         | 
         | - Getting notifications from websites I _want_ notifications
         | from
         | 
         | This should not be impossible just because there are 5 trillion
         | websites I _don 't_ want notifications from.
        
           | egeozcan wrote:
           | Exactly, why can't there be presets to set the trust level on
           | websites?
           | 
           | - Hey, it seems like you want to use our notification
           | feature, and we are delighted. For us to be able to push you
           | messages, you need to set out trust level to "App", to do
           | this...
           | 
           | I mean, nobody will open a menu and click on 3 things for a
           | spammy website, but many people will do it for their favorite
           | 3 sites.
           | 
           | - Clicking on this button will trust this website (example
           | com) as an [App] and give it the ability to store information
           | on your device, and send you notifications. Please don't do
           | this for the sites you don't trust. [Cancel] [I trust this
           | website]
           | 
           | Our grandparents can run random executable from the internet
           | and with some common-sense training and scary popups from
           | anti-virus and UAC, it is still risky but tolerable. This is
           | way more secure and easier, why cant this work?
           | 
           | - You already trust this app (example com) and it can already
           | store information on your device and send you notifications.
           | Clicking on this button will add it to your trusted apps and
           | allow it to ask for files stored on your device by other
           | apps, and also allow it to ask to access your contacts. These
           | permissions will be only available when you are actively
           | using the app. [Cancel] [Add example com to trusted apps]
           | 
           | Please some browser implement this.
        
             | mmis1000 wrote:
             | Edge exactly did that.
             | 
             | It collects feedback from users. If a site got rejected by
             | literally every user. It will be put into the `don't even
             | bother prompt the users` list.
             | 
             | And start from this point.
             | 
             | No edge user will be prompt actively by that site (There
             | will still be a icon on the url bar. If the user really
             | really want to enable it, he can click it himself).
        
               | sefrost wrote:
               | Do we know which sites are in that "no prompt" list?
        
               | miohtama wrote:
               | Sounds like a ripe for abuse on a competitor's site. How
               | does Edge identity the reporting users?
        
             | AlexandrB wrote:
             | Except make it wayyyy more fine grained so I can give the
             | app permission to store files but not send notifications or
             | to access the files of one other app but not another.
             | Something like Apple's fine-grained photos permissions
             | would be good.
        
             | philistine wrote:
             | Make it even simpler. If the website is installed as a Web
             | App on your iPhone home screen (it has an icon), it can ask
             | for and send notifications. It appears in your list of apps
             | in notifications in Settings and can be fine-tuned that
             | way.
             | 
             | You remove it from your home screen and the notification
             | permission disappears.
             | 
             | The bugs related to notifications currently seen in the
             | beta are due to this user interaction not yet being
             | implemented I'd bet.
        
           | littlecranky67 wrote:
           | The problem is, that you can always "force" the user to allow
           | you notifications or cannot proceed. And $GREEDY_MEGACORP
           | would abuse their power and send you a "verification code"
           | through the push notification, else you cannot process the
           | signup etc.
           | 
           | The good thing about the Apple Appstore, that
           | promotional/marketing push notifications are forbidden and
           | will get your App banned from the Appstore. Just allowing
           | generic WebPush on PWAs will result in companies requiring it
           | with dark patterns, just to increase your "engagement" with
           | their product (read: send you spam).
           | 
           | I have mixed feelings about webpush, I myself as a developer
           | want it for my users, but I see the history of the web and
           | how everything nice got destroyed eventually by spam.
        
             | rezonant wrote:
             | As others point out, promotional pushes are allowed on iOS.
             | 
             | A site can refuse to give you content if you don't allow
             | push notifications but in practice that doesn't happen
             | because the incentive of the site is to have more traffic.
             | Android has had this capability for an eon and I have never
             | experienced a site try to force you to accept push
             | notifications.
             | 
             | I have also never observed a site try to use push
             | notifications as a "verification" method. Do you have an
             | example of this in the wild?
             | 
             | And you can always control your allowance of push
             | notifications after the fact, either at the moment you
             | receive the offending notification or from the site
             | settings. Android specifically has a fantastic way to do
             | this without even leaving your notification area. Not sure
             | how Apple's system works in that regard.
        
             | deergomoo wrote:
             | > that promotional/marketing push notifications are
             | forbidden
             | 
             | Not anymore they're not. It wasn't enforced when it was on
             | the books; last year the rule was removed. Even Apple
             | advertise their own services via push now.
        
         | gruez wrote:
         | >- Easily scriptable notifications from your own code (I run
         | notify.run which provides an HTTPS-to-WebPush bridge so you can
         | send yourself a push notification from curl)
         | 
         | The article speculates that it might not be freely available
         | like push notifications are for firefox/chrome.
         | 
         | >Maybe we have to register the origin with the Apple Developer
         | Program, similar to the Safari Push Notification program for
         | macOS. If this is the case, the developer will have to pay the
         | annual Apple Developer Program fee to send push messages to
         | their web users.
         | 
         | In that case you're probably better off installing a
         | notification app from the appstore to handle it, eg.
         | pushover/pushbullet.
        
         | vbezhenar wrote:
         | It does not work right now and probably will require Apple
         | whitelisting your website.
        
         | olliej wrote:
         | I'd love to have a chrome person comment on the actual stats of
         | web push usage because no non-developer I've interacted with
         | has ever intentionally accepted push spam.
         | 
         | I think all developers think that their site is somehow unique
         | in the value of their notifications, despite all evidence to
         | the contrary.
        
         | mrtksn wrote:
         | So, once this is integrated clicking a link will have the
         | following experience:
         | 
         | 1) Tracking preferences overlay that will have a wall of text
         | with "Accept all" and "Customise" buttons that would not be
         | clickable the first few tries.
         | 
         | 2) Then you will have a modal that asks for your e-mail address
         | so you can receive the newsletters from the website that you
         | are trying to understand what is this all about.
         | 
         | 3) And now you will have a prompt that asks you if you like to
         | receive notifications.
         | 
         | 4) Once you are done with all this, half of the page will be
         | covered with links and menus to the other parts of the website
         | and when you scroll menus and overlays with call with action
         | will appear and disappear
         | 
         | 5) The rest of the website will be ads.
         | 
         | 6) Hopefully somewhere there will be an image or paragraph, you
         | will be able to find it if you still remember why you click on
         | the link
         | 
         | 7) if you scroll a bit too much the screen will be blacked out
         | and a call for action(sign up for a newsletter, create account,
         | download the app or buy something) will appear and will have
         | hard to identify close button that won't be working the first
         | few taps.
         | 
         | My point is, the Web is not a garbage because Safari lacks
         | notification or something like that and it is not salvageable
         | without finding a new business model.
         | 
         | Wouldn't be great to have a notification once the new Wordle of
         | the day arrives? Sure. But the business model of it appears to
         | be selling it to the people who love all the steps from 1 to 7
         | listed above. I don't blame the creator of the game, it's just
         | a testament of the state of the Web business.
         | 
         | I prefer apps for that reason. They also seem to be riddled
         | with full screen ads(especially the casual games are horrible)
         | that make your phone warm BUT you get yo see a high quality
         | screen(App Store, Google Play etc.) where you can read reviews,
         | description and see screenshots before committing into
         | installing the app(Unlike the web apps where you first
         | "install" the app and then you are on your own to figure out
         | what is this all about).
         | 
         | Besides the questionable practices of Apple/Google, there's
         | barely any reason to have browsers with PWA capabilities. I
         | would love it if the web browsers had JS off by default and
         | load JS only when the user explicitly requests it.
         | 
         | Most of the noise comes from Web technologists don't want to
         | learn anything beyond JavaScript.
         | 
         | I hope Apple doesn't cave and if they do, I hope they implement
         | it in a way that improves the experience of the USER. If I
         | wanted an Android experience, I would have used an Android
         | phone, let's keep the iPhone experience like a iPhone
         | experience.
        
           | donmcronald wrote:
           | I really want to see PWAs become a viable option for
           | distribution, but you do have a point. I sometimes end up on
           | websites using my iPad where I can't even read the content. I
           | wonder why the company would publish a site so poor that it
           | seems like they didn't even try it on an iPad and then I
           | remember they've already gotten my ad impressions, so they've
           | gotten everything they want and don't care if I get to read
           | the content.
           | 
           | I don't think those sites get penalized much either. I
           | arrived via search or a link and they don't get demoted in
           | search or have that link removed. I can't flag them as being
           | a bad site or "take back" my view. I find it very
           | frustrating.
        
           | teucris wrote:
           | > Most of the noise comes from Web technologists don't want
           | to learn anything beyond JavaScript.
           | 
           | I agree with a lot of what you wrote, but not this. PWAs are
           | the only path for allowing open publishing of apps that is
           | not gated by a large tech firm. While I completely understand
           | the value proposition of app stores, there are tons of app
           | scenarios held back by them. Two instances are small web[0]
           | and situated software[1] as app stores keep people from
           | publishing apps meant for small audiences by their very
           | definition. Both of these fall under your desire to maximize
           | the user experience.
           | 
           | Furthermore, PWAs are the closest thing we have to a
           | "universal" platform. Personally, if I could write an app
           | that worked on all mobile devices and the web in COBOL, I'd
           | be thrilled to write "PROCEDURE DIVISION" every day.
           | 
           | The web is a mess because it is open. The web can be scammy
           | because some people are scammy. Restricting PWAs doesn't
           | mitigate scamminess - it just restricts openness.
           | 
           | [0] https://ar.al/2020/08/07/what-is-the-small-web/ [1]
           | https://www.gwern.net/docs/technology/2004-03-30-shirky-
           | situ...
        
             | mrtksn wrote:
             | I wouldn't call PWAs universal platform when they are very
             | dependent on Google being nice with Chrome. Also, browsers
             | already have integrated blocking mechanism that is supposed
             | to block out phishing sites etc. but that's only the case
             | because so far the browser infrastructure
             | operators(Mozilla, Google, Apple) are acting nice. The Push
             | notification infrastructure is also possible only because
             | the browser infrastructure operators are being nice so far.
             | 
             | At any point Google or other browser makers can choose not
             | to be nice or can be compelled to follow regulatory orders.
             | 
             | The only universal thing about PWAs is the availability of
             | Google Chrome to all major platforms and that's up to the
             | courtesy of Google. They may choose not to do it at any
             | time, they may choose to discontinue their push
             | notification backend just like that.
             | 
             | Browsers are not standalone apps anymore, they are
             | dependent on the backend services of the browser company.
             | That's especially important when they need to run PWA.
        
               | tl wrote:
               | > very dependent on Google being nice with Chrome
               | 
               | Replace Apple/Safari+Google/Chrome with
               | Google/Gmail+Microsoft/Outlook and you currently describe
               | e-mail. That situation is still preferable to the tons of
               | messaging walled gardens. And the web is getting new
               | feature albeit slowly and unevenly in a way that mail is
               | not.
        
           | brimble wrote:
           | > Wouldn't be great to have a notification once the new
           | Wordle of the day arrives?
           | 
           | Nitpick, but for that specific example, not really, since
           | it's always at local midnight according to your device's time
           | setting (for certain values of "arrive", since all past and
           | future wordles are always present in the page source).
           | 
           | [EDIT] Oh and:
           | 
           | > Besides the questionable practices of Apple/Google, there's
           | barely any reason to have browsers with PWA capabilities. I
           | would love it if the web browsers had JS off by default and
           | load JS only when the user explicitly requests it.
           | 
           | I think the biting of the apple, if you will, for the web,
           | was allowing Javascript to _ever_ initiate requests on its
           | own, or modify e.g. post request content or URLs pre-flight
           | without user review and explicit OK. It 's fucking absurd
           | that clicking a link to a site means it gets to track your
           | mouse movements _and_ send them, live, to some server
           | somewhere, and that disabling that means wrecking most of the
           | web and making it a pain to use. No transmitting data without
           | explicit user action and review, I say.
        
         | HWR_14 wrote:
         | I don't get push notifications. If I don't have time to poll
         | your service, I don't have time to respond to a notification.
         | Besides, I'd rather not have my priorities determined for me.
        
           | ThatPlayer wrote:
           | That really depends on the service. If it's a messaging app,
           | then I'm not exactly going to poll a service until a new
           | message arrives. Or any other type of app that I'm not
           | checking if there's no updates.
        
         | ninkendo wrote:
         | I just wish that "web applications" were more clearly
         | distinguished from regular web sites.
         | 
         | I hate that a regular website I may tap on a search result
         | would ever want to send me notifications. But, if a site that
         | is clearly more of an "application" to me (like an online
         | spreadsheet or chat app, etc) wants to send me notifications,
         | that could be a useful feature.
         | 
         | If there were a more clear delineation between the two, I could
         | just avoid tapping anything that is labeled as an "app" in
         | search results (or just ban them from my search engine), and
         | restrict myself to "plain old webpages" only when searching.
         | And "plain old webpages" would be even further restricted than
         | they already are: No notifications prompts, no ability to see
         | my accelerometer, no offline storage available, etc etc.
         | 
         | It's nice that the web has turned into a viable app platform,
         | but it doesn't mean I want every page I visit to be allowed to
         | have "app-like" capabilities. In an ideal world 95% of the web
         | would just be plain old documents accessed over HTTP(S) and the
         | "apps" would be distinguishable as such.
        
           | kitsunesoba wrote:
           | I would find this type of differentiation useful as well.
           | "Plain" sites are generally faster to skim through for what
           | I'm looking for, (no gauntlet of spinning loading wheels and
           | dialogs or JS gumming up functions like Find in Page) so it
           | would be nice to be able to prioritize those in search
           | results.
           | 
           | One way to do this would be for the crawler to compare the
           | content of JS-enabled and JS-disabled states and if they
           | differ dramatically, the web destination gets tossed in the
           | "web app" bucket.
        
           | jtbayly wrote:
           | Interesting idea, but it sort of falls down in my mind when I
           | think about forums. I want notifications for some forums that
           | I join. And I certainly don't want to remove forums from
           | search results. They're the only place you can get decent
           | product reviews right now.
        
           | merrywhether wrote:
           | A great heuristic would just be only granting notification
           | prompt permission after you've installed the PWA to your Home
           | Screen. That shows clear interest by the user in a longer-
           | term relationship with the app, and they could grant other
           | expanded prompt permissions based on that also.
        
             | The_rationalist wrote:
             | Most users never add a shortcut to the home screen and yet
             | use plenty of web apps.
        
               | coder543 wrote:
               | Apple could certainly do more to help people understand
               | that they can install web apps to their home screen.
               | Right now, it is an option that few users are aware of,
               | buried in the menus as it is, and even fewer think about
               | it.
               | 
               | When a related app is available in the App Store, Apple
               | will show a big banner at the top of the web page
               | encouraging you to install it... but they won't do
               | anything for a PWA.
               | 
               | I think requiring "installation" of a PWA is a great way
               | to balance the concerns people have about getting spammed
               | with requests to spam you with notifications you don't
               | want... but I also wish Apple would support PWAs better
               | in general, so I can see it both ways.
        
             | rabuse wrote:
             | This will probably be the way it's handled honestly. I
             | think it's the best option.
        
           | riskable wrote:
           | Hate to break this to you but even a website that consists of
           | nothing but HTML (no CSS/JS) is still an "app". It's still
           | being executed in the runtime environment of your browser
           | when you load it. Just because the website _chose_ to limit
           | it 's use of the features to just one (HTML) doesn't mean
           | it's not an app (when executed in a web browser).
           | 
           | If you read that HTML in a text editor then yeah, that's not
           | an app but if you're reading it in a web browser you're
           | "running it".
        
       | osrec wrote:
       | Honestly, it's about time we got push notifications in Safari. It
       | was a really annoying missing feature, and meant I had to pay
       | Twilio to notify the iOS users of my web app. Hopefully that'll
       | soon come to an end.
        
       ___________________________________________________________________
       (page generated 2022-02-01 23:01 UTC)