Re: CR: Web Notifications

On Thu, Jun 4, 2015 at 10:37 PM, timeless <timeless@gmail.com> wrote:
> http://www.w3.org/TR/2015/CR-notifications-20150519/

I reviewed against https://notifications.spec.whatwg.org/ since it
seems some was still applicable. Here are the changes I made:
https://github.com/whatwg/notifications/commit/26c3f5dc79cf5c5ffdc46b11a245cc411fadd9ef


>>    Return notification, but continue running these steps asynchronouusly [sic].

This was not.

>> When this specification refers to displaying notifications on the "desktop",
>> it generally refers to some static display area outside the web page,
>> but may take several forms, including:

This was not.

>> It does not specify exactly how a user agent should display these notifications [...]
>
>> This specification does not define exactly how the user agent should display the notification, [....]

This was not.

>> In particular, notifications as specified here only can contain text and icon content.

This was not.

>> in case the underlying notification platform does not provide that capability.

This was not.

> I'd suggest: occurrence -> that something happened

Fixed.

> While that's a good thing to encourage, it's probably worth explaining
> *why*, (insert link to an A11y / WAI document).

Pointer appreciated.

>> behaviour

HTML is in en-GB. Forgot to change this after copying this text.
(Ideally we abstract all this out at some point I suppose, but no
clear plan for that as of yet.)

>> When *a* notification notification is activated by the user,
> ...
>> for each Notification object representing notification,
>
> Why "for each"? there's only one ("a") at the beginning of the sentence.

This was not. (Originally a single notification could be represented
by multiple objects, even within the same global object.)

> What happens when the display space shrinks?

This is now left up to implementations.

>> Once fetching has finished and the image format is supported,

This was not.

>> set notification's icon image to the decoded resource.
>> (Otherwise notification has no icon image.)
>
> This parenthetical should be some sort of else...

I don't think so.

> Presumably platform -> platform api

Fixed.

>> If the notification platform supports icons, new's icon URL is set
>> and has not yet been fetched, fetch it and wait for the resource to be fully downloaded.
>
> "fetched" doesn't seem to have a note about "staleness" (I only
> quickly read the If I fetch a resource and it expires in 1 minute, and
> then I queue a notification a minute later, has it been "fetched" for
> the purposes here?

For now that level of detail seems unwarranted.

>> If options's lang is a valid BCP 47 language tag, or the empty string,
>> set notification's language to options's lang, or set it to the empty string otherwise. [LANG]
>
> .... or ... otherwise -> .... otherwise ...

This was not.

>> If the notification platform supports icons, the user agent may start fetching notification's icon URL at this point, if icon URL is set.

This was not.

> you use both ".... or ... otherwise" and ".... and ... otherwise", I'd
> encourage you to avoid both. (There's probably a pattern for your use
> of both, but English doesn't support it.)

This was not.

> http://www.w3.org/TR/2015/CR-notifications-20150519/#tags-example
>
> Say
> Instance 1
> // Instance notices there is new mail.
> x= new Notification("New mail from John Doe",
>                  { tag: 'message1' });
>
>
> Instance 2:
> // Slightly later, this instance notices
> // there is new mail.
> y = new Notification("New mail from John Doe",
> { tag: 'message1' });
>
>
> Instance 1:
> // Slightly later, because it feels like it
> x.close()
>
> What happens to y?

Nothing.

> Are any events triggered?

Nope.


-- 
https://annevankesteren.nl/

Received on Friday, 3 July 2015 08:43:27 UTC