[gcompris-devel] The next file format
Bruno Coudoin
bruno.coudoin at gcompris.net
Mon Aug 25 22:46:04 UTC 2014
Le 24/08/2014 14:56, Inge Wallin a écrit :
>
> On Wednesday, August 20, 2014 11:23:02 Bruno Coudoin wrote:
>
> > Hi,
>
> >
>
> > On the GCompris side we are also working on defining a new dataset
>
> > format for the new Qt Quick based version.
>
> >
>
> > While we are not specifically addressing language or grammar
>
> > application, we found the need to define a way to create, distribute,
>
> > share and play datasets for specific activities.
>
> I think it would be a good thing if we could share at least container
> format and parts of the library to access it.
>
Hi, I agree. We clearly have an opportunity here.
> > Also we have not mentioned it in this wiki page but we are already
>
> > distributing in the new GCompris voice files as Qt qrc files. They are
>
> > Qt specific but very easy to manage because you can load them
>
> > dynamically and then access their content through qrc:// url anywhere in
>
> > Qml. To us, 'qrc' is good candidate for the container of the datasets as
>
> > it is Qt native.
>
> I read up a little on qrc, and it seems that these files are
> hard-coded resources that are part of the source code. A resource
> compiler, rcc, is then used to create C source files that are later
> compiled using the normal C/C++ compiler and becomes part of the
> executable.
>
> This is a good way to collect parts of the application like icons and
> similar. But it is not what the discussion about the new file format
> is about. We are talking about external data files that can be
> downloaded or created after the program is already installed.
>
Yes, it must be external binary files. In fact there is 2 use cases for
the rcc technology, internal or external. Please read the chapter
"External Binary Resources" here:
http://qt-project.org/doc/qt-5/resources.html#external-binary-resources
Like I said, this is already the way we distribute voices in the Qt
version of GCompris. It is easy to generate rcc files with the rcc tools
that comes with Qt. It is easy to register an rcc file with
QResource::registerResource and then we can just access the resource
with a qrc: protocol.
> > Some feedback on your proposal, I am confused by the 'confidence level'.
>
> > If it is a student mark, it may not be desirable to put it in the
>
> > dataset itself because it make sense to have it on a read only storage
>
> > area (most distros will do that). On this topic at GCompris we are
>
> > interested in a teacher specific tool to help them in their daily usage,
>
> > we starting specifying it there :
>
> > http://gcompris.net/wiki/Administration_design
>
> Yes, confidence level is not the ideal term but so far we haven't
> found anything better. What it is is the level of confidence that the
> student has for a particular word. This tries to capture how strongly
> the word is put into the memory of the student, or loosely put how
> long it can be expected to be before they forget it. If you are not
> familiar with the term 'spaced repetition training', I urge you to
> look it up on Wikipedia, they have an excellent article about it.
>
> This used to be known as 'grade' in Parley but we are providing a tool
> for learning and training, not for testing so grade is not applicable.
> Besides, grades also have a negative connotation in that you are a bad
> person if you have a bad grade. Since any low confidence level is a
> necessary step to the higher confidence levels we wanted to get rid of
> the grade connotations and that was the best we could come up with. I
> guess 'mark' is vaguely similar to grade in this case.
>
This is clear now and it is what I understood. Then I don't think it is
appropriate to put this data in the container for several reasons:
- may be on a read only storage
- may be hard to centralize the student data to let a teacher review it
- an update of the file would overwrite the student data.
> Would you be interested in sharing the container format with us if we
> can agree on how we store the internal data?
>
>
Sure, we have to continue the discussion and hopefully find the path
that will satisfy all the needs.
Bruno.
More information about the Gcompris-devel
mailing list