[Tux4kids-tuxtype-dev] tuxtype progress control

[Jordan Erickson]

I'll reply in-line to your post:

uwe wrote:
> I am writing up a little document of my thought around this idea.
> clearly I am not an expert when it comes to that, but I am a father and
> a part-time teacher and I try to imagine what I would need.

The more perspectives the better! :)


> anyway, my basic idea would be that there is a seperate module
> (web-based) that one can login to and find the answers to how good or
> not a student or class is. to compare them to averages from previous
> classes, to see progress and to find weaknesses. a place where teachers
> and students can logon to.
> then of course there needs to be a way of transmitting the results from
> tuxtype to this module. we'll have to see. would you think this is a
> good approach?

I think a web-based approach would be beneficial over others, since the
Tux4Kids apps are cross-platform. It would make it that much easier for
teachers/non-tech people to interface with.


> I think things you would want to capture are:
> - sessions a student has done
> - errors in these sessions, number of words, letters, sentences
> - indicators: number of words per minute, errors per minute or errors
> per number of words. errors per letter (weakness on certain fingers)
> - results relatively to other students of same class or age
> - trend of a student over time.
> - ready for "final exam" evaluation
> - last exercise done, number of exercise over time
> 
> and much more. but that should not be a problem once the data is there.

I think that is a great start, and like you said, the more data you have
the better. I know for a fact being able to customize reports based on
what people actually *want* to see (everyone will have a different
approach to grading students, for example on WPM, CPM, how long they
took in the lesson overall, etc.) - see my post right below yours for
some other perspectives on what school staff thinks students should
focus on.


> from a datawarehouse and business intelligence standpoint we have what
> is called "dimensions" and "facts". dimensions are the views or key
> aspects one can look at the data. e.g. per class, per time period, per
> student, per age. and so on. so class, time, student and age (a,mongst
> others) would be dimensions. then there are what is called the "facts".
> the numbers and values that we capture. it is important to capture these
> a the correct granularity. the more details one has, the more can drill
> down from e.g. class to student to month and day maybe right down to the
> hour and minute and find out the very details and answers one is looking
> for. this of course means storing more data and processing more data. on
> the other hand, if you don't use much granular data, later in the
> process there is no way of going deeper into the data to find answers.

Again, this is a great approach. I don't think that it would really
require that much more processing power/storage space to store more
'facts' than less, and again IMHO it would make for a much more
granulated and customizable reporting mechanism.


Cheers,
Jordan