You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cordova.apache.org by Shazron <sh...@gmail.com> on 2012/06/19 01:53:58 UTC

UUID on other platforms

I've punted this suggestion for 1.9.0 but I'll put it up for further discussion:
https://issues.apache.org/jira/browse/CB-49?focusedCommentId=13292414&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13292414

What are the implications of UUID for Android, BB, and any others
(platform policy wise)? I see a WP7 quirk in the docs already.
Do we NEED this?

In iOS 6, Apple already provides [UIDevice identifierForVendor] and
[UIDevice identifierForAdvertising]. The latter changes when the
device is erased but is the functional equivalent of the old UDID. The
former is the same for devices from the same vendor on the same device
- different in all other cases.

This problem is kinda solved for iOS 6 but not previous platforms for
iOS. My suggestion is do what Apple says (which we are doing as
suggested in our iOS quirk for device.uuid
http://docs.phonegap.com/en/1.8.1/cordova_device_device.md.html#device.uuid)
then use [UIDevice identifierForAdvertising] when they are on iOS 6.

Re: UUID on other platforms

Posted by Brian LeRoux <b...@brian.io>.
shall we open an issue for this guys? I like shaz's suggestion of
moving to a js based impl (and deprecating)

On Tue, Jun 19, 2012 at 6:40 PM, Shazron <sh...@gmail.com> wrote:
> Well considering we (may?) have consensus on a half-year time-based
> deprecation policy (not that we have agreed on deprecating this uuid
> property), I guess this will stay for now.
>
> My notice to the list is that the iOS uuid implementation as I
> proposed will change according to iOS version (but we'll cross the iOS
> 6 bridge when we get to it)
>
> We could also move this to a JS based implementation of uuid, since we
> ship this function:
> https://github.com/apache/incubator-cordova-js/blob/master/lib/common/utils.js#L62
> - we call that, and save to localStorage, and retrieve etc
>
> That will make it pretty consistent across app platforms, and for our
> docs. We should move away from privacy-breaking hardware identifiers
> because of potential platform policy reasons (read: Apple). My
> suggestion - if devs want this, they can write a plugin (it can even
> be cross-platform - basically our existing uuid core implementation
> ripped out as a stand-alone plugin).
>
> On Tue, Jun 19, 2012 at 1:04 PM, Brian LeRoux <b...@brian.io> wrote:
>> Ya for sure cookies tend to get used that way. Really they are a
>> persistence mechanism and any value in them is...well, any value you
>> want, like say, a guid! =)
>>
>> Still think this is an app developer concern and good candidate for
>> removal within the scope of the vision. Many apps dont even need this
>> capability. (At first)
>>
>>
>> On Tue, Jun 19, 2012 at 3:50 PM, Jesse <pu...@gmail.com> wrote:
>>> uuid is essentially to identify an application install on a particular
>>> device.
>>> It makes sense in an installed app context, though not in a web-app context.
>>> The web-app comparison would probably be cookies, which do not work
>>> consistently across devices.
>>>
>>> On Tue, Jun 19, 2012 at 12:36 PM, Brian LeRoux <b...@brian.io> wrote:
>>>
>>>> > ...the more things we simply offload to call 'app
>>>> > developer concerns' the less anyone will need our framework. IMHO this
>>>> > would be pursuing the 'cease to exist' mantra from the wrong direction.
>>>>
>>>> how so? should we be advocating adding a unique identifier to web
>>>> browser apis? (don't think it'll get much traction!)
>>>>
>>>
>>>
>>>
>>> --
>>> @purplecabbage
>>> risingj.com

Re: UUID on other platforms

Posted by Shazron <sh...@gmail.com>.
Well considering we (may?) have consensus on a half-year time-based
deprecation policy (not that we have agreed on deprecating this uuid
property), I guess this will stay for now.

My notice to the list is that the iOS uuid implementation as I
proposed will change according to iOS version (but we'll cross the iOS
6 bridge when we get to it)

We could also move this to a JS based implementation of uuid, since we
ship this function:
https://github.com/apache/incubator-cordova-js/blob/master/lib/common/utils.js#L62
- we call that, and save to localStorage, and retrieve etc

That will make it pretty consistent across app platforms, and for our
docs. We should move away from privacy-breaking hardware identifiers
because of potential platform policy reasons (read: Apple). My
suggestion - if devs want this, they can write a plugin (it can even
be cross-platform - basically our existing uuid core implementation
ripped out as a stand-alone plugin).

On Tue, Jun 19, 2012 at 1:04 PM, Brian LeRoux <b...@brian.io> wrote:
> Ya for sure cookies tend to get used that way. Really they are a
> persistence mechanism and any value in them is...well, any value you
> want, like say, a guid! =)
>
> Still think this is an app developer concern and good candidate for
> removal within the scope of the vision. Many apps dont even need this
> capability. (At first)
>
>
> On Tue, Jun 19, 2012 at 3:50 PM, Jesse <pu...@gmail.com> wrote:
>> uuid is essentially to identify an application install on a particular
>> device.
>> It makes sense in an installed app context, though not in a web-app context.
>> The web-app comparison would probably be cookies, which do not work
>> consistently across devices.
>>
>> On Tue, Jun 19, 2012 at 12:36 PM, Brian LeRoux <b...@brian.io> wrote:
>>
>>> > ...the more things we simply offload to call 'app
>>> > developer concerns' the less anyone will need our framework. IMHO this
>>> > would be pursuing the 'cease to exist' mantra from the wrong direction.
>>>
>>> how so? should we be advocating adding a unique identifier to web
>>> browser apis? (don't think it'll get much traction!)
>>>
>>
>>
>>
>> --
>> @purplecabbage
>> risingj.com

Re: UUID on other platforms

Posted by Brian LeRoux <b...@brian.io>.
Ya for sure cookies tend to get used that way. Really they are a
persistence mechanism and any value in them is...well, any value you
want, like say, a guid! =)

Still think this is an app developer concern and good candidate for
removal within the scope of the vision. Many apps dont even need this
capability. (At first)


On Tue, Jun 19, 2012 at 3:50 PM, Jesse <pu...@gmail.com> wrote:
> uuid is essentially to identify an application install on a particular
> device.
> It makes sense in an installed app context, though not in a web-app context.
> The web-app comparison would probably be cookies, which do not work
> consistently across devices.
>
> On Tue, Jun 19, 2012 at 12:36 PM, Brian LeRoux <b...@brian.io> wrote:
>
>> > ...the more things we simply offload to call 'app
>> > developer concerns' the less anyone will need our framework. IMHO this
>> > would be pursuing the 'cease to exist' mantra from the wrong direction.
>>
>> how so? should we be advocating adding a unique identifier to web
>> browser apis? (don't think it'll get much traction!)
>>
>
>
>
> --
> @purplecabbage
> risingj.com

Re: UUID on other platforms

Posted by Jesse <pu...@gmail.com>.
uuid is essentially to identify an application install on a particular
device.
It makes sense in an installed app context, though not in a web-app context.
The web-app comparison would probably be cookies, which do not work
consistently across devices.

On Tue, Jun 19, 2012 at 12:36 PM, Brian LeRoux <b...@brian.io> wrote:

> > ...the more things we simply offload to call 'app
> > developer concerns' the less anyone will need our framework. IMHO this
> > would be pursuing the 'cease to exist' mantra from the wrong direction.
>
> how so? should we be advocating adding a unique identifier to web
> browser apis? (don't think it'll get much traction!)
>



-- 
@purplecabbage
risingj.com

Re: UUID on other platforms

Posted by Brian LeRoux <b...@brian.io>.
> ...the more things we simply offload to call 'app
> developer concerns' the less anyone will need our framework. IMHO this
> would be pursuing the 'cease to exist' mantra from the wrong direction.

how so? should we be advocating adding a unique identifier to web
browser apis? (don't think it'll get much traction!)

Re: UUID on other platforms

Posted by Jesse <pu...@gmail.com>.
I think there is still value in having this info presented consistently
across devices by our API.  The absolute uniqueness may be questionable,
but still I think it is relevant.  guid+localStorage IS pretty easy, but so
is everything else,  and the more things we simply offload to call 'app
developer concerns' the less anyone will need our framework. IMHO this
would be pursuing the 'cease to exist' mantra from the wrong direction.

On Tue, Jun 19, 2012 at 5:54 AM, Brian LeRoux <b...@brian.io> wrote:

> > What are the implications of UUID for Android, BB, and any others
> > (platform policy wise)? I see a WP7 quirk in the docs already.
> > Do we NEED this?
>
> That is the key question. I'd say that is the app developer concern
> (generating a guid from js and putting in localStorage is pretty
> easy).
>



-- 
@purplecabbage
risingj.com

Re: UUID on other platforms

Posted by Brian LeRoux <b...@brian.io>.
> What are the implications of UUID for Android, BB, and any others
> (platform policy wise)? I see a WP7 quirk in the docs already.
> Do we NEED this?

That is the key question. I'd say that is the app developer concern
(generating a guid from js and putting in localStorage is pretty
easy).