Regarding our language tools

[Inge Wallin]

On Tuesday, February 11, 2014 21:34:24 Andreas Cord-Landwehr wrote:
> On Tuesday 11 February 2014 14:18:29 Inge Wallin wrote:
> > I took a closer look at KWordQuiz and for that I agre with you. But for
> > Parley I do not. KWordQuiz has the simple layout and interaction that is
> > well suited for "this week's 15 words" but parley is either redundant
> > (KWordQuiz works just as well) or much too advanced for that purpose.
> > 
> > Instead I think we should target KWordQuiz to the groups you mention above
> > (Camille and possibly Gan) in the personas and Parley to Tina. There is no
> > reason to even have Parley if we don't target the more motivated and
> > advanced learner with it.
> 
> Yes, Parley is an advanced trainer and KWordquiz keeps it simple. But I do
> not see why using Parley should be harder to use than KWordquiz. What I
> learned from a workthrough with our Personas at the last sprint is that all
> of them have several difficulties to use both applications. Hence, a new
> interaction design is needed in any case, if we want to support these
> personas.

I don't think I ever said that I wanted Parley to be harder to use.  Except 
for the fact that bigger complexity always makes something a little harder, 
especially for the beginner, there is no reason why it should be harder per 
se.

Regarding the interaction design, I think the biggest problem (as I mentioned 
before) is that the applications are not focussed around the learning process.  
But I believe we can fix that without too much problems.

I spent a fair amount of time with Parley the last 2 days, both running and 
experimenting with it, and looking at the code. And one thing struck me. Often 
in the code you read about 'grades' and "correct/ / ncorrent" choices. Those 
are not terms that are used in learning. Those are terms that are used in 
tests. So I think that this is another area where we need some interaction 
design: To separate the learning process from the test situation more clearly.

The current UI gives too much negative feedback, I think. It's irrelevant if I 
haven't learned a word 17 times when I finally managed to learn it. Hence it is 
irrelevant to give the number of "wrong" answers as feedback to the user 
during the learning process. For a test it's a bit different, but the current 
implementation also doesn't work for tests. For one thing, the answer is given 
after every question so you cannot go back and correct your answer if you 
figured it out a bit later.

> For such an overhoul, I believe that it is feasable to meet the needs of all
> of them -- without loosing functionalty. And in the light of
> maintainability, I prefer to focus at one application instead of a set of
> severals (Parley + KWordquiz + Widget). Yet, if that is not possible, I
> would be comfortable with talking about providing several views to a core
> trainer engine.

Agreed.  But I never heard of Widget before. Is that an application in kde edu 
too? Also, Arikulate is part of the mix so don't forget that. Perhaps it will 
all end up in a large language laboratory type of application if we can manage 
to keep everything simple.

However, there is one area where KWordQuiz could be warranted even if Parley 
could cover all language needs: namely non-language flash card learning. There 
is an example file of US states and capitals and that works very well in 
KWordQuiz.  Maybe we could reposition KWordQuiz into the more generic 
KFlashCard (or skip the K altogether) steer the user into Parley for all 
things about language.

> So my proposal, now a little bit more precise:
> * define the usage goals of our Personas (and maybe revisit them)
> * discuss training workflows for the Personas
> * evaluate if it is possible to do this in one interface, or if several
> interfaces are required * evaluate if this can be done atop of Parley
> (which is IMO the best foundation we currently have) * draft a plan :)

Good. And maybe add a few personas who aren't strict learners but who are part 
of the learning process. Teachers come to mind, but also our volunteers that 
create your Artikulate sound files and similar people. I think we should try to 
map the whole workflow that lead up to the current personas, i.e. the learners.

> I will ping Thomas and see if he or anybody else of our interaction
> designers is interested in joining the discussion.
> 
> For the usage goals and workflows, I created an (empty) wiki page at:
> http://community.kde.org/KDEEdu/Language/VokabularyLearningWorkflows
> Currently, I am not yet sure what a good structure for this page will be,
> but I am optimistic that we will find out, once it gets more content :)

Great!

> Greetings,
> Andreas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-edu/attachments/20140212/a63d1069/attachment.html>