random and fixed features in mutipledataset

[Johnny Jazeix]

Hi,
let's not forget there are 2 main use cases of GCompris: at school
(under a teacher, meaning they want to use GCompris to teach a
specific feature) or at home (where we should be able to provide
multiple different levels and not always the same ones).

Depending on the activities we port to multiple dataset, it may make
sense to have a multiple dataset supporting both conditions to please
both worlds.
Maybe the activity can be reworked to keep the current behaviour and
on some keywords, making a static dataset.

The multiple dataset only provides data, it's up to the activity to
handle it how it wants :).

Johnny


Le ven. 2 août 2019 à 15:16, Timothée Giet <animtim at gmail.com> a écrit :
>
> Hi,
>
> I'm sorry I don't agree with that idea:
>
> -having a fixed list like makes the activity predictable for the user,
> so it's better to keep the randomness for the user to always have to
> really look and think to provide the result accordingly.
>
> -you talk about "if I'm working on a new number...", but that is not the
> way it works, as each dataset provides a range of number, not a specific
> new number.
>
> Timo.
>
>
> Le 02/08/2019 à 15:02, Emmanuel Charruau a écrit :
> > Hi Akshay, Hi mentors,
> >
> > I see a major lack on the pedagocical use of multipledataset.
> > It is only working in random mode at the moment, which causes me problems.
> >
> > Let me give you an example.
> >
> > If I am working on a new number in falling dice, for example 7 I want
> > to be able to recall the other numbers but 7 more often than the other
> > numbers.
> > At the moment we will only give to the config file the following configuration :
> >
> > "level" : "3",
> >          "objective" : qsTr("Select the number on dominoes up to 7"),
> >          "words" : [
> >             "1",
> >             "2",
> >             "3",
> >             "4",
> >             "5",
> >             "6",
> >             "7"
> >          ]
> >
> > and apply random choices within these numbers.
> >
> > We should have the choice to be able to provide exactly the numbers we
> > want to use through a list which will not be  random.
> >
> > Example if I want to teach 7:
> >
> >        "words" : [
> >             "7",
> >             "2",
> >             "3",
> >             "7",
> >             "5",
> >             "6",
> >             "6",
> >             "1",
> >             "2",
> >             "7",
> >             "4",
> >             "5",
> >             "6",
> >             "7",
> >             "1",
> >             "7",
> >             "3",
> >             "4",
> >             "5",
> >             "7",
> >             "7",
> >             "1",
> >             "7",
> >             "6",
> >             "4",
> >             "7",
> >             "6",
> >             "7",
> >             "7",
> >             "2",
> >             "3",
> >             "4",
> >             "5",
> >             "6",
> >             "7",
> >
> >          ]
> >
> > Note how often 7 is present in the list (also 6 since theoritically it
> > is the less well known numbers being just taught).
> >
> >
> > I would also suggest to rename "words" into StaticListDataList and
> > RandomDataList or whatever like this, as the keywords needs to be more
> > generic than words and have modes (random or fixed list)
> >
> > In the activity I am writting at the moment, it will look like this:
> >    "words" : [
> >             "12345,2",
> >             "1245,5",
> >             "42,125",
> >             "456,85",
> >
> > you can see that these data are not words.
> > There are also other activities where we present images and not words.
> >
> > Finally if we use new keywords we can choose to keep the actual word
> > "words" and not break the compatibility.
> >
> > Mentors, Akshay feedback is very welcomed :)
> >
> > Regards,
> >
> > Emmanuel
>
>
>