3 ms·

That's over the top. Why? Because keeping applications sandboxed and adblocked inside the browser is better than downloading some proprietary app (doesn't apply

That's over the top. Why? Because keeping applications sandboxed and adblocked inside the browser is better than downloading some proprietary app (doesn't apply to FOSS clients). Some of those applications, you'll definitely want notifications, like the web app for work chat, etc. What you don't want to do is discourage using the browser as a platform, but you definitely should proceed with caution for enabling those notifications rather than outright disabling them.

lloydatkinson 4y ago
If you’d read the article you’d see that I specifically call out disabling browser push notifications, not web sockets or server sent events.
- toastal 4y ago
  I know exactly what the Push API is--and I resent that iOS won't enable it making PWAs not really work for a lot of use cases because developers need notifications. I currently, and purposefully, get push notifications from Element, Mattermost, Slack, and (for the time being while traffic has been low) replies on Mastodon--when I was using a web client for email, I did there as well. I think project managers would be pretty mad if I wasn't responding to messages, as much as I'd be angry if they required that I install some Electron app with tracking onto my system, where I can't make modifications via userScripts and userStyles, just to get notified (or an inbox full of "You have 100 unread message on ACMEChat from 1 hour ago"). > I have never seen them used for any legitimate or useful purpose > They take advantage of non-technical users and frighten them > Non-technical users habitually click “OK” [..] in any dialog The article is written in a way that's myopic, belittling, and alarmist--and the author, you presumably, should think about this tone when there's a good reason browser vendors agreed on a standard included the API to solve a real-world need. Communicating time-sensitive information is a valid use case--especially with PWAs where the web is the platform. This very important to Linux, BSD, KaiOS, Capyloon, etc. users where developers aren't making their application source available, but users can run a properly built web app, and it can also be good for the developer that doesn't have the time and budget to build native apps for all platforms or users that don't wish to install an "app" for everything.
  - lloydatkinson 4y ago
    No, I'm quite happy with the tone and message of the article thank you. Just because you have not experienced the many instances of compromised machines thanks to this "feature" (and feel free to find the HN thread where this article is discussed also) does not mean the feature is not dangerous. I assume you don't have a problem with popups not being heavily restricted and opt-in now. I have no idea why you seem to have a personal problem with it but, alas, it's personal. If you don't want to use them that is up to you, I don't care. Feel free to make a pull request.
    - toastal 4y ago
      It's not a personal problem--you're making that up; the purpose of Push is getting native app features in your web app which is good for consumers--especially those on less popular operating systems as almost all have some form of modern web browser (and some, like FxOS and its successor KaiOs, fully banking on the web as competent, common-denominator application platform). Opt in is perfect too because it's not hard to clearly communicate what notifications will do and making it the result of a consensual user action and not page load (if page-load-pop-up is still the case, that sucks though). Not every user needs to be coddled. Not every use of Push is malicious. Folks build real apps on the web. You didn't even bother to explain how getting real-time chat notifications isn't a real-world demand. And running through the comments are just as many folks saying PWAs are a real need to them and that this rhetoric is alarmist.
      - lloydatkinson 4y ago
        Feel free to send a pull request.
        toastal 4y ago
        When your comments are full of rebuttals to "there's no use case" and you refuse to acknowledge them, that revision is on you.

lloydatkinson 4y ago

If you’d read the article you’d see that I specifically call out disabling browser push notifications, not web sockets or server sent events.

toastal 4y ago
I know exactly what the Push API is--and I resent that iOS won't enable it making PWAs not really work for a lot of use cases because developers need notifications. I currently, and purposefully, get push notifications from Element, Mattermost, Slack, and (for the time being while traffic has been low) replies on Mastodon--when I was using a web client for email, I did there as well. I think project managers would be pretty mad if I wasn't responding to messages, as much as I'd be angry if they required that I install some Electron app with tracking onto my system, where I can't make modifications via userScripts and userStyles, just to get notified (or an inbox full of "You have 100 unread message on ACMEChat from 1 hour ago"). > I have never seen them used for any legitimate or useful purpose > They take advantage of non-technical users and frighten them > Non-technical users habitually click “OK” [..] in any dialog The article is written in a way that's myopic, belittling, and alarmist--and the author, you presumably, should think about this tone when there's a good reason browser vendors agreed on a standard included the API to solve a real-world need. Communicating time-sensitive information is a valid use case--especially with PWAs where the web is the platform. This very important to Linux, BSD, KaiOS, Capyloon, etc. users where developers aren't making their application source available, but users can run a properly built web app, and it can also be good for the developer that doesn't have the time and budget to build native apps for all platforms or users that don't wish to install an "app" for everything.
- lloydatkinson 4y ago
  No, I'm quite happy with the tone and message of the article thank you. Just because you have not experienced the many instances of compromised machines thanks to this "feature" (and feel free to find the HN thread where this article is discussed also) does not mean the feature is not dangerous. I assume you don't have a problem with popups not being heavily restricted and opt-in now. I have no idea why you seem to have a personal problem with it but, alas, it's personal. If you don't want to use them that is up to you, I don't care. Feel free to make a pull request.
  - toastal 4y ago
    It's not a personal problem--you're making that up; the purpose of Push is getting native app features in your web app which is good for consumers--especially those on less popular operating systems as almost all have some form of modern web browser (and some, like FxOS and its successor KaiOs, fully banking on the web as competent, common-denominator application platform). Opt in is perfect too because it's not hard to clearly communicate what notifications will do and making it the result of a consensual user action and not page load (if page-load-pop-up is still the case, that sucks though). Not every user needs to be coddled. Not every use of Push is malicious. Folks build real apps on the web. You didn't even bother to explain how getting real-time chat notifications isn't a real-world demand. And running through the comments are just as many folks saying PWAs are a real need to them and that this rhetoric is alarmist.
    - lloydatkinson 4y ago
      Feel free to send a pull request.
      - toastal 4y ago
        When your comments are full of rebuttals to "there's no use case" and you refuse to acknowledge them, that revision is on you.