Re: [Ayatana] Chrome desktop notifications (and APIs)

[Matthew Paul Thomas <mpt@xxxxxxxxxxxxx>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Martin Owens wrote on 21/06/10 11:47:
> 
> On Mon, 2010-06-21 at 10:46 +0100, Matthew Paul Thomas wrote:
>> 
>> If you are referring to Ubuntu's notification bubbles, Chrome's
>> notification API is fundamentally incompatible with Ubuntu
>> notification bubbles. It is also fundamentally incompatible with
>> Windows notification balloons, and with Growl notifications on Mac
>> OS X. It's very much a Chrome-specific thing. 
> 
> That's fairly damning, do you have a link where I could read up on the
> technical explanation?

I replied when Google first proposed this in March 2009:
<http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-March/018769.html>

They're working on a draft that they may eventually submit to the W3C:
<http://dev.w3.org/2006/webapi/WebNotifications/publish/>

(Skimming through that, I see that I may have been mistaken about it
being incompatible with Windows notification balloons. I thought it
allowed notifications of indefinite duration, contra Windows Vista and
later, but it doesn't seem to.)

>                        I'd like to understand how text could be
> incompatible.
>...

It's not the text, but the interaction model. The draft says "while the
Notification object offers an 'onclick' event, applications may enhance
their functionality by listening for that event, but *must not depend on
receiving it*, in case the underlying notification platform does not
provide that capability."

This is, really, as much as we could reasonably hope for from a
specification of this nature (it even name-checks Notify OSD). But
Ubuntu has negligible market share amongst Web designers. So
pessimistically, if Chrome on other platforms allows onclick(), authors
will expect it to be possible on every platform -- it won't even occur
to them to test with Notify OSD.

The current draft also has this use case: "A calendar application alerts
the user for an upcoming meeting, and allows the user to easily specify
a 'snooze' delay of several possible time periods." I don't see how the
draft API would cater for that, so they may be planning to extend it
further.

- -- 
Matthew Paul Thomas
http://mpt.net.nz/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkwfS8AACgkQ6PUxNfU6ecoEuQCcDxXpNeA4JkKAN/6UPmm6iGo8
viQAoKkvS8gGeDBxMHPw6wvi6alGJH0m
=N6WY
-----END PGP SIGNATURE-----