W3C home > Mailing lists > Public > public-web-notification@w3.org > August 2013

Re: Re[2]: Moving to Last Call

From: Justin DeWitt <dewittj@google.com>
Date: Wed, 7 Aug 2013 09:35:45 -0700
Message-ID: <CAE54qEb9-K4h70QorzHhCtY9bYoA5kZ3wJ4+D3Deukid_KsmKQ@mail.gmail.com>
To: sokoldv@ukr.net
Cc: Andrew Wilson <atwilson@google.com>, Jon Lee <jonlee@apple.com>, WG <public-web-notification@w3.org>
I have made a Chrome extension for calendar work with our newer
chrome.notifications API, and I found that it's useful information to know
that a user has dismissed a notification.  In this specific case, I
implemented event reminders using notifications created 10 minutes before
the event and then again at two minutes before.  However, if the user had
explicitly closed the first notification, I took that as confirmation that
the user had acknowledged their upcoming event, and didn't need further

Without guarantees that the user caused the close event, I would no longer
be able to assume that they had dismissed my earlier notification and would
need to either always re-notify the user at the 2 minute mark - potentially
annoying them - or not show another notification, and risk failing to alert
the user if they hadn't in fact seen the initial notification.

Of course, this use case only makes sense in a world where notifications
are aggressively hidden from view, which is where Chrome is headed.


On Tue, Aug 6, 2013 at 2:35 AM, <sokoldv@ukr.net> wrote:

>  Hello everyone,
> We have no use knowing that notification was closed by user or platform
> cause in second case we still didn't know whether user have seen it or not
> (platform closes notification before user actually see it or user just
> ignore it).
> Also we can close it manually before platform does it (actually I do so).
> --- Ориг╕нальне пов╕домлення ---
> В╕д кого: "Andrew Wilson" <atwilson@google.com>
> Дата: 6 серпня 2013, 11:05:37
> On Mon, Aug 5, 2013 at 11:22 PM, Jon Lee <jonlee@apple.com> wrote:
> Is there a specific situation where understanding the reason is crucial to
> the page's functionality? On iOS and Mac, for example, notifications are
> expected to act like taps on the shoulder, and not necessarily return vital
> information that users cannot get from the web page or app. Given that the
> spec declare the event model is best-effort, the web page will not get an
> accurate picture of which notifications a user has seen and acknowledged,
> even with this additional flag.
> Jon
> In the app that i'm familiar with (gmail) I can't think of any reason why
> I'd care about system-close vs user-close. I would like a robust way to
> tell if a notification has been closed, but I'm not sure I care about the
> source.
> That said, web apps can present a much better user experience if it has
> some idea about whether the platform closes notifications automatically.
> I'd rather have some way to know if this is the case (perhaps just by
> requiring this behavior in the spec), per my previous email.
Received on Wednesday, 7 August 2013 16:36:13 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:53:14 UTC