6 ms·

This “feature” is abused terribly around the web. For every site with useful notifications like gmail, there are 10 which misuse it. Yesterday I had the misfort

This “feature” is abused terribly around the web. For every site with useful notifications like gmail, there are 10 which misuse it. Yesterday I had the misfortune of misclicking on the prompt from a website. I started getting notifications like “YOUR COMPUTER IS INFECTED WITH A VIRUS”. Turning off notifications for that website took at least 4 clicks through Chrome settings. Good to know there is a setting to turn it off completely on Firefox. If I personally actually wanted notifications from a service, I’d install the mobile app. Second, Mozilla seems to use telemetry data responsibly and well. Turning off notification prompts by default can’t possibly be done unless you have the data on acceptance rate on different types of prompts. If you’re making such decisions based on your intuition you’d likely get it wrong. I ask the folks on HN who constantly criticise the collection of such telemetry, what did I lose as a user when Firefox collected this anonymised data? More importantly, how would you have made a decision here without the data? Intuition? (I’d request that no one reply with platitudes like “with enough data nothing is anonymous” and “you’re making a nothing-to-hide argument”)

birksherty 8y ago
I don't agree with that thinking. I can say the opposite. Why did someone enabled the notifications feature globally? Was it based on intuition? Then they are wrong, they are actually. They should collect data from all people before inserting any such "features".
- nindalf 8y ago
  What? You want them to “collect data from all people” on how they use a feature right now ... before adding the feature? How would that work, exactly? > they are actually. Ok.
  - birksherty 8y ago
    I don't want that. The previous comment said that if it's not based on telemetry, then the intuition based decision is wrong. I am suggesting that if that's the case then we have to use telemetry before implementing everything. We should never have any feature without telemetry confirming it first.
anon12413 8y ago
You're making an assumption that their decision was based on data they collected. Not worth discussing data privacy with you over that-- they may have very well done simple user testing or done it on a whim of a few users that put a good case forward.
- askvictor 8y ago
  From the article: > According to our telemetry data, the notifications prompt is by far the most frequently shown permission prompt, with about 18 million prompts shown on Firefox Beta in the month from Dec 25 2018 to Jan 24 2019. Not even 3% of these prompts got accepted by users. The decision seems very clearly based on the collected data.
  - anon12413 8y ago
    No it's not, they quoted some telemetry but the feature and decision may have been decided way before then. The data may only be a justification for it.
    - zaarn 8y ago
      Well, there is little proof that they decided something and then went looking for data unless you can find some statement by mozilla that supports your theory?
    - nindalf 8y ago
      > May have been That’s a huge assumption. I’m going with a smaller assumption that Mozilla wouldn’t publicly lie for no obvious gain.
- darkpuma 8y ago
  He's also failing to acknowledge that using telemetry in a constructive way does not preclude using that telemetry in a malicious way as well.
  - nindalf 8y ago
gsich 8y ago
>I ask the folks on HN who constantly criticise the collection of such telemetry, what did I lose as a user when Firefox collected this anonymised data? More importantly, how would you have made a decision here without the data? Intuition? (I’d request that no one reply with platitudes like “with enough data nothing is anonymous” and “you’re making a nothing-to-hide argument”) You don't need telemetry to find out that notifications get abused like that.
- zaarn 8y ago
  No but you need telemetry to find out how users react to abusive notifications and how they interact with websites where they do genuinely want notifications from. You can't just guess that.
  - layoutIfNeeded 8y ago
    >you need telemetry to find out how users react to abusive notifications You need telemetry to find out how users react to abusive notifications? Wow. I know that A/B testing is in vogue, but can’t they just use some common sense? People despise abusive notification prompts. People despise abusive location prompts. People despised abusive popups to the point that every browser blocks them and websites started emulating popups via CSS to keep abusing their users (hi Medium!).
    - lifthrasiir 8y ago
      Unfortunately we are living in the world where the common sense is wrecked. People do complain if abusive notifications are blocked and websites are rendered unusable as a result, though it is completely the fault of websites! Keep in mind that browser vendors are trying to solve multivariate equations that involve users, web developers, companies, abusers and crackers...
    - michaelt 8y ago
      You need telemetry to find out how users react to abusive notifications? I run an ad blocker, so my perception of notifications is they're used for things like breaking news notifications on news websites [1]. That's not a feature I personally need, but 24 hour news channels exist so presumably someone feels they need regular news updates. So no, it's not obvious to me everyone dislikes notification requests, even if I have them all blocked myself. If you let curmudgeon developers like me dictate products' features sets based on intuition, there'd be no HTML e-mail, no third-party cookies, no WebGL, no emojis, no WebUSB.... :) [1] https://www.wired.co.uk/article/push-notifications-breaking-news https://www.wired.co.uk/article/push-notifications-breaking-...
jeanmichelx 8y ago
I like the Firefox approach. Nightly collects a lot of data (because you're a power user that want to improve the browser). Regular release collect a lot less.
- fxfan 8y ago
  This is the reason why windows doesn't let you opt out- power users would opt out. Case in point - jump lists were removed from start menu. That said- why treat regular users unfairly by not optimizing for their use cases too?
  - misterdoubt 8y ago
    I don't see how this issue has much to do with "fairness." There's no reason that any organization must treat its audiences identically.
andimm 8y ago
Can't you also block it completely in Chrome and whitelist some websites? Settings -> Privacy.. -> Content Settings.. -> Notifications (Blocked) I haven't seen any prompts since I changed that. Although I don't use this feature at all.
codedokode 8y ago
I think the bad thing is not telemetry itself but the fact that it is silently enabled by default instead of asking the user whether they agree to it. Make it voluntary and there would be no problems at all. I didn't ask for this feature. Also, one doesn't need telemetry to notice how notifications are abused on the web. You can just start a browser, visit top 1000 popular sites and count how many of them show the popup. Also, it seems like everyone tries to abuse notifications. For example, Youtube app shows a notification when the channel you are subscribed to releases a new video. Is it so urgent, that you need to distract the user? They could show this information when the user opens the app.
- naikrovek 8y ago
  Have you used Firefox? A notification about telemetry is shown on-screen the first time you launch it, including a link (or maybe button, I forget) taking you to the settings page where it can be disabled. If you're talking about telemetry collection in general, and not specifically Firefox, then nevermind. I don't find telemetry bad in any way, provided that it is non-identifying telemetry, so I don't mind that it's on by default. There should always be an option to turn it off for those that want to. Many, many more applications collect telemetry than people suspect, and for every application you know of that collects it, there are a dozen that collect it without telling you. You are right about notifications being over used. We've all gotten notification fatigue, now, because every website thinks it's own notifications are super important. I want notifications when major news events happen, when my wife texts me, and for nothing else. I do not, and will not ever, care to be notified the instant a website has new content.
- Vinnl 8y ago
  > Make it voluntary and there would be no problems at all. Unfortunately that would make the telemetry non-representative. That said, it is opt-in when possible, if you consider "using Nightly" as opt-in (it's clearly explained before and after installing Nightly). > You can just start a browser, visit top 1000 popular sites and count how many of them show the popup. That doesn't tell you anything about what behaviour led up to a permission request that got granted vs the 97% that got ignored. Furthermore, it wouldn't have told them that the notification request is denied far more often by users than the webcam/mic request. > Also, it seems like everyone tries to abuse notifications. For example, Youtube app shows a notification when the channel you are subscribed to releases a new video. Is it so urgent, that you need to distract the user? They could show this information when the user opens the app. It would be nice to have data on when users revoke permissions again as well, indeed.
SquareWheel 8y ago
>Turning off notifications for that website took at least 4 clicks through Chrome settings. It's actually just three clicks, and you can do so from the page itself. No need to go into settings. Click the Lock icon left of the omnibar. On pages that request notification permissions (eg. reddit), you'll see a toggle for it. Along with any other permissions requested.
- codedokode 8y ago
  But you have to make only one click to agree to the notifications.
  - SquareWheel 8y ago
    Yes, or one click to disagree. What's wrong with that approach?
M2Ys4U 8y ago
>This “feature” is abused terribly around the web. For every site with useful notifications like gmail, there are 10 which misuse it. I think you're off by an order of magnitude there. It's at least 100 sites that misuse it for every legitimate one.
- jacurtis 8y ago
  Yes I think that 1/100 is a more accurate ratio. I have only accepted 4 notification prompts and I know that i have encountered more than 400 that I denied.
tinus_hn 8y ago
If you like it, feel free to enable telemetry. But why force me to enable it as well?
- kowdermeister 8y ago
  Exactly for actions like this which they've done in the post. Would you prefer a more obvious switch on setup?

birksherty 8y ago

I don't agree with that thinking. I can say the opposite. Why did someone enabled the notifications feature globally? Was it based on intuition? Then they are wrong, they are actually. They should collect data from all people before inserting any such "features".

nindalf 8y ago
What? You want them to “collect data from all people” on how they use a feature right now ... before adding the feature? How would that work, exactly? > they are actually. Ok.
- birksherty 8y ago
  I don't want that. The previous comment said that if it's not based on telemetry, then the intuition based decision is wrong. I am suggesting that if that's the case then we have to use telemetry before implementing everything. We should never have any feature without telemetry confirming it first.

anon12413 8y ago

You're making an assumption that their decision was based on data they collected. Not worth discussing data privacy with you over that-- they may have very well done simple user testing or done it on a whim of a few users that put a good case forward.

askvictor 8y ago
From the article: > According to our telemetry data, the notifications prompt is by far the most frequently shown permission prompt, with about 18 million prompts shown on Firefox Beta in the month from Dec 25 2018 to Jan 24 2019. Not even 3% of these prompts got accepted by users. The decision seems very clearly based on the collected data.
- anon12413 8y ago
  No it's not, they quoted some telemetry but the feature and decision may have been decided way before then. The data may only be a justification for it.
  - zaarn 8y ago
    Well, there is little proof that they decided something and then went looking for data unless you can find some statement by mozilla that supports your theory?
  - nindalf 8y ago
    > May have been That’s a huge assumption. I’m going with a smaller assumption that Mozilla wouldn’t publicly lie for no obvious gain.
darkpuma 8y ago
He's also failing to acknowledge that using telemetry in a constructive way does not preclude using that telemetry in a malicious way as well.
- nindalf 8y ago

gsich 8y ago

>I ask the folks on HN who constantly criticise the collection of such telemetry, what did I lose as a user when Firefox collected this anonymised data? More importantly, how would you have made a decision here without the data? Intuition? (I’d request that no one reply with platitudes like “with enough data nothing is anonymous” and “you’re making a nothing-to-hide argument”) You don't need telemetry to find out that notifications get abused like that.

zaarn 8y ago
No but you need telemetry to find out how users react to abusive notifications and how they interact with websites where they do genuinely want notifications from. You can't just guess that.
- layoutIfNeeded 8y ago
  >you need telemetry to find out how users react to abusive notifications You need telemetry to find out how users react to abusive notifications? Wow. I know that A/B testing is in vogue, but can’t they just use some common sense? People despise abusive notification prompts. People despise abusive location prompts. People despised abusive popups to the point that every browser blocks them and websites started emulating popups via CSS to keep abusing their users (hi Medium!).
  - lifthrasiir 8y ago
    Unfortunately we are living in the world where the common sense is wrecked. People do complain if abusive notifications are blocked and websites are rendered unusable as a result, though it is completely the fault of websites! Keep in mind that browser vendors are trying to solve multivariate equations that involve users, web developers, companies, abusers and crackers...
  - michaelt 8y ago
    You need telemetry to find out how users react to abusive notifications? I run an ad blocker, so my perception of notifications is they're used for things like breaking news notifications on news websites [1]. That's not a feature I personally need, but 24 hour news channels exist so presumably someone feels they need regular news updates. So no, it's not obvious to me everyone dislikes notification requests, even if I have them all blocked myself. If you let curmudgeon developers like me dictate products' features sets based on intuition, there'd be no HTML e-mail, no third-party cookies, no WebGL, no emojis, no WebUSB.... :) [1] https://www.wired.co.uk/article/push-notifications-breaking-news https://www.wired.co.uk/article/push-notifications-breaking-...

jeanmichelx 8y ago

I like the Firefox approach. Nightly collects a lot of data (because you're a power user that want to improve the browser). Regular release collect a lot less.

fxfan 8y ago
This is the reason why windows doesn't let you opt out- power users would opt out. Case in point - jump lists were removed from start menu. That said- why treat regular users unfairly by not optimizing for their use cases too?
- misterdoubt 8y ago
  I don't see how this issue has much to do with "fairness." There's no reason that any organization must treat its audiences identically.

andimm 8y ago

Can't you also block it completely in Chrome and whitelist some websites? Settings -> Privacy.. -> Content Settings.. -> Notifications (Blocked) I haven't seen any prompts since I changed that. Although I don't use this feature at all.

codedokode 8y ago

I think the bad thing is not telemetry itself but the fact that it is silently enabled by default instead of asking the user whether they agree to it. Make it voluntary and there would be no problems at all. I didn't ask for this feature. Also, one doesn't need telemetry to notice how notifications are abused on the web. You can just start a browser, visit top 1000 popular sites and count how many of them show the popup. Also, it seems like everyone tries to abuse notifications. For example, Youtube app shows a notification when the channel you are subscribed to releases a new video. Is it so urgent, that you need to distract the user? They could show this information when the user opens the app.

naikrovek 8y ago
Have you used Firefox? A notification about telemetry is shown on-screen the first time you launch it, including a link (or maybe button, I forget) taking you to the settings page where it can be disabled. If you're talking about telemetry collection in general, and not specifically Firefox, then nevermind. I don't find telemetry bad in any way, provided that it is non-identifying telemetry, so I don't mind that it's on by default. There should always be an option to turn it off for those that want to. Many, many more applications collect telemetry than people suspect, and for every application you know of that collects it, there are a dozen that collect it without telling you. You are right about notifications being over used. We've all gotten notification fatigue, now, because every website thinks it's own notifications are super important. I want notifications when major news events happen, when my wife texts me, and for nothing else. I do not, and will not ever, care to be notified the instant a website has new content.
Vinnl 8y ago
> Make it voluntary and there would be no problems at all. Unfortunately that would make the telemetry non-representative. That said, it is opt-in when possible, if you consider "using Nightly" as opt-in (it's clearly explained before and after installing Nightly). > You can just start a browser, visit top 1000 popular sites and count how many of them show the popup. That doesn't tell you anything about what behaviour led up to a permission request that got granted vs the 97% that got ignored. Furthermore, it wouldn't have told them that the notification request is denied far more often by users than the webcam/mic request. > Also, it seems like everyone tries to abuse notifications. For example, Youtube app shows a notification when the channel you are subscribed to releases a new video. Is it so urgent, that you need to distract the user? They could show this information when the user opens the app. It would be nice to have data on when users revoke permissions again as well, indeed.

SquareWheel 8y ago

>Turning off notifications for that website took at least 4 clicks through Chrome settings. It's actually just three clicks, and you can do so from the page itself. No need to go into settings. Click the Lock icon left of the omnibar. On pages that request notification permissions (eg. reddit), you'll see a toggle for it. Along with any other permissions requested.

codedokode 8y ago
But you have to make only one click to agree to the notifications.
- SquareWheel 8y ago
  Yes, or one click to disagree. What's wrong with that approach?

M2Ys4U 8y ago

>This “feature” is abused terribly around the web. For every site with useful notifications like gmail, there are 10 which misuse it. I think you're off by an order of magnitude there. It's at least 100 sites that misuse it for every legitimate one.

jacurtis 8y ago
Yes I think that 1/100 is a more accurate ratio. I have only accepted 4 notification prompts and I know that i have encountered more than 400 that I denied.

tinus_hn 8y ago

If you like it, feel free to enable telemetry. But why force me to enable it as well?

kowdermeister 8y ago
Exactly for actions like this which they've done in the post. Would you prefer a more obvious switch on setup?