GSoC 2019

[Timothée Giet]

Le 21/12/2018 à 12:21, Timothée Giet a écrit :
> Le 21/12/2018 à 12:11, Johnny Jazeix a écrit :
>> Hi,
>>
>> Le ven. 21 déc. 2018 à 11:44, Aman Kumar Gupta <gupta2140 at gmail.com 
>> <mailto:gupta2140 at gmail.com>> a écrit :
>>
>>     Hi,
>>
>>
>>     On Fri, Dec 21, 2018 at 3:35 PM Johnny Jazeix <jazeix at gmail.com
>>     <mailto:jazeix at gmail.com>> wrote:
>>
>>         Hi,
>>
>>         Like every year, we plan to be part of GSoC.
>>
>>         We also need to define the tasks we'll propose (KDE submits
>>         its application on Jan 15th).
>>
>>         Looking at the major activities to port the only missing one
>>         is "analog electricity" (https://phabricator.kde.org/T5954).
>>         There is still room for other activities too (adding biology
>>         ones for example) or the actual phabricator tasks
>>         (https://phabricator.kde.org/project/profile/1/).
>>
>>         The second point to work on, is to continue working on the
>>         multi dataset per activity. For now, it's still a WIP in
>>         branch multi_dataset
>>         (https://cgit.kde.org/gcompris.git/log/?h=multiple_dataset)
>>         but it will be a rework on a lot of activities to use dataset
>>         in form of json instead of hardcoding the levels in the
>>         js/qml files (the aim not being to do all of them in the GSoC
>>         but having it half time with an activity).
>>
>>
>>     Can we have an overview idea of what is the concept of multiple
>>     dataset and how are the activities supposed to adapt to it?
>>
>>
>> For now, in the branch, only the reversecount activity has been done 
>> as proof of concept.
>> In the menu, we plan to replace the stars icons with the dataset 
>> name: https://pasteboard.co/HSKeyND.png
>> On clicking on it, it will display the different datasets specific to 
>> this activity and their objectives: https://pasteboard.co/HSKeiTW.png
>> (In the code, a dataset is inherited from qml object: 
>> https://cgit.kde.org/gcompris.git/tree/src/core/Dataset.qml?h=multiple_dataset 
>> and set inside the resource folder of the activity.)
>>
>> Then, once one is chosen, we can click on the activity and play it as 
>> usual.
>>
>>     Also what is expected from this task? It can be seen that some
>>     work has already started in the branch, it would be good to know
>>     what does it aim at? How are the activities supposed to work
>>     after addition of multiple dataset?
>>
>>
>> The aim is to adapt datasets to the skill of the children who are 
>> using them. An activity may cover a few ages and the aim of it is to 
>> split the learning curve. For example, if we take the multiplication 
>> activity, a child who is just starting to learn multiplication tables 
>> will often only learn during its first year the tables until 5, but 
>> GCompris multiplication activity presents the tables from 1 to 9 from 
>> the beginning. It would be great to split in 2 different datasets 
>> (maybe more).
>
>
> Hi,
>
> I still don't think we should replace the system with stars we have 
> now, I thought we agreed on that when we discussed about multiple-dataset.
>
> The stars are still a valid indication of the overall difficulty level 
> of the activities, regardless of dataset system. It would end up more 
> confusing for the user if we replace that with some dataset names.
>
> I think it would be much better to have a bar button inside the 
> activity to select and display the dataset level.
>
>
> Another way if we really want to mix dataset concept with stars would 
> be to assign each dataset a star level, and then show on activity 
> icons in the menu the difficulty level corresponding to the selected 
> dataset.
>
>
> Timothée
>
>

To complete my previous comment, I think a simple way if we mix both 
stars level and dataset, is to also make use of the global setting for 
stars. Like, if an activity has 3 datasets corresponding to star level 
1, 2 and 3, by default activate all 3 and show the 3 stars icon on it. 
But, if on main setting user/teacher restricts to show only 1st level, 
then automatically set the activity to use only the first dataset/star 
level and then show the 1 star difficulty level on the activity. That 
would look very logical and integrate well with what we have already.

Timothée

>>     Will the end-user be able to create a dataset for himself, or it
>>     means multiple dataset like in categorization activity, except
>>     that the user will be provided with the all the dataset along
>>     with a small description for each, and they can choose one
>>     according to their suitability?
>>
>>
>> Not yet to create its own datasets. The aim afterwards, would be to 
>> have a server able to create and use datasets. But multiple_dataset 
>> is the first step towards it.
>>
>>
>>         What do you think? Do you have other ideas we can propose?
>>         Who has time (and will) to mentor this year?
>>
>>
>>         Johnny
>>
>>
>>     Thanks,
>>     Aman
>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/gcompris-devel/attachments/20181221/5aabed28/attachment-0001.html>