On Thu, Jun 9, 2016 at 10:23 AM, Andrei Chis <[email protected]>
wrote:

>
>
> On Wed, Jun 8, 2016 at 9:28 AM, stepharo <[email protected]> wrote:
>
>>
>>
>> Le 7/6/16 à 12:00, Andrei Chis a écrit :
>>
>> Hi Peter,
>>
>> On Tue, Jun 7, 2016 at 12:05 AM, Peter Uhnak <[email protected]> wrote:
>>
>>> Hi,
>>>
>>> Privacy>>sendDiagnosticsAndUsageData should be ternary, not binary.
>>>
>>> Because right now if I refuse sending the data I will be asked again
>>> every single time.
>>>
>>
>> Indeed right now, at least in Spotter, the notification is shown every
>> time it is opened in
>> a new image, even if the setting was explicitly set. It was done like
>> this as a reminder,
>> especially if you have the setting set to true.
>>
>>
>> Andrei for the shortcut reporter we put a number of show. Like that the
>> user will be reminder a number
>> of times and the system will get calm. May be you could do the same
>>
>
> In a new image the message to send usage data is shown only once.
> Indeed we could add a counter but that would have to maintain state
> between images.
>

This should also include an API for third-party apps to correctly ask the
question.

For example Roassal just checks the value and if it's not true it will ask
again…

RTRoassalExample>>open
| pager |
self checkSendingData.
...

RTRoassalExample>>checkSendingData | result | Privacy
sendDiagnosticsAndUsageData ifTrue: [ ^ self ]. result := UIManager default
questionWithoutCancel: 'Please, do you want to let us know how you use the
Roassal example browser? ' title: 'Send Us Usage Data'. result ifNil: [ ^
self ]. result ifTrue: [ Privacy sendDiagnosticsAndUsageData: true ]

With an approach like this a counter wouldn't help.

So maybe a new method is warrented… something like


RTRoassalExample>>checkSendingData | result | Privacy
askForDiagnosticsAndUsageData: 'Custom Message' for: #Roassal.
Privacy sendDiagnosticsAndUsageData ifTrue: [ self enableWhatever ]

And `askForDiagnosticsAndUsageData:` would open the popup if the counter
hasn't been hit and do nothing otherwise.

It's also a question whether it shouldn't be per-app, because maybe you
want logging enabled for an app X, but not Y; but that's extra complexity.

Peter


>
>
>>
>>> So the proper behavior (imho) should be:
>>>
>>> ask Privacy for the setting… if the setting is not defined, then show a
>>> popup.
>>> If the setting is defined then respect it and do not show another popup.
>>>
>>
>> The current behaviour with showing the notification in Spotter and
>> Nautilus should be a temporary one.
>> Ideally we just need a single uniform way of showing this notification
>> plus the option to control
>> and see the data that each tool wants to record at a fined-grained level.
>>
>>
>>>
>>> Also it would be nice to know what happens with the data.
>>>
>>
>> We store it on a server and use it for various analyses.
>>
>> what can you tell us on processor use in Spotter.
>>
>
> Not right at this moment. I'm way to busy with other urgent things.
> Some older results were discussed here:
> http://scg.unibe.ch/scgbib?query=Kube15a&display=abstract
> Just for these result to carry more value we still need more people to
> send data.
> Right now few do.
>
> Cheers,
> Andrei
>
>
>>
>> It is also publicly available: GTEventTool default download.
>>
>>
>>> I mean my projects are open source so "sendSourceCode" shouldn't be an
>>> issue… but what you can possibly learn from it?
>>> Why not just analyze the content of SmalltalkHub/GitHub?
>>>
>>
>> Because if you have the setting enabled it does not always work to map
>> different actions that with the code that you are working on.
>>
>> Cheers,
>> Andrei
>>
>>
>>>
>>> Peter
>>>
>>>
>>
>>
>

Reply via email to