Re: [Alarm API] data ... yet another DB?

On Wed, Feb 13, 2013 at 10:42 PM, Mounir Lamouri <mounir@lamouri.fr> wrote:
> On 12/02/13 14:34, Christophe Dumez - SISA wrote:
>> Hi Marcos,
>>
>> For context, you are referring about the "data" argument of this method:
>>>  AlarmRequest add(Date date, TimezoneDirective respectTimezone, optional Object data);
>>
>> I have given this some thought and I tend to agree with you. It seems to me (as well) that the client application could just as well use localStorage or IndexedDB to achieve this. I personally do not see any real advantage in using this data argument (over a database API) except maybe convenience. As you rightfully mentioned though, this convenience has a price.
>>
>> As a consequence, I would be in favour of removing this "data" argument unless someone has a use case for it that cannot be satisfied by existing Database APIs. Does anyone have any thought on this?
>
> The reason this was used in our design is that it was said to be more
> "developer friendly" in the sense that developers could simply store
> data they need to get when the alarm would go (like the event name and
> id in the app DB).

I think this is a valuable feature, and we should not remove it just
because we cannot easily find a properly formalized way to describe
it. Anyway the implementation matters.

>
> I do not know if that would be that useful to developers and we should
> probably look at how Firefox OS is using that information. Given that
> it's the last argument, I wouldn't mind having it removed and added
> later as optional.
>

Should we give it a chance still? Even if you postpone for later,
someone will have to think it through anyway. Unless the problem
solves itself, we should try to solve it now :).

Best regards,
Zoltan

Received on Wednesday, 13 February 2013 21:01:12 UTC