Hey C,

I actually have been spending more time on the longer versions, as I see more 
promise in this approach than the shorter reference clips. I believe it was you 
that first suggested the longer version and this approach is certainly more 
flexible. The short reference version feels more like haiku, with its density 
of symbolism and thought. This density may better reflect the J language, but 
is not the starting point for a learner.  I do think that even with appropriate 
animation that it may take a few repetitions, if only to let the viewer learn 
the language of the animation. Once the language is 'learned', all of the 
subsequent primitives should follow the same pattern which means that 
repetition may not be as necessary. Of course, right now we are still fiddling 
with this 'language', but once we settle on it we can put together a 'watch 
this first' animation that can flatten that learning curve as well. Just a 
glimpse of my perspective of the landscape we are traversing.

You mention movement not being necessary; a halo, bounce or sparkle should 
indicate what is the target of the operation. The challenge I find with that is 
in showing that the arguments are corresponding scalar values. This can be done 
with the highlighting options, but in order to show which two scalars of the 
matrix are involved they would need to be sequentially highlighted which 
implies an order of operation we wish to avoid. One of the first animations 
used random patterns, but this was even more distracting. If we highlight all 
the scalars at once we may not know which ones correspond. The movement solves 
this by bringing the scalars into proximity which clearly shows the 
correspondence. Earlier versions place the operator between each of these 
pairs, later versions superimposed the operator (which seems to be the 
distractor you describe). Either way I'll disagree with you here and say that 
movement looks to be the best way to indicate clear argument correspondence, 
whic
 h is a key concept to convey to the learner for the matrix/vector examples. We 
can avoid movement with the scalar arguments, but I would like to use the 
scalar to introduce the movement which is key to the vector/matrix examples. 
Perhaps this consistency is not as important as I think. Let me know if you 
have other options, I haven't mentioned.

I have a few questions about the covering the numbers. Does it make a 
difference to you which argument is placed on top when the scalar arguments are 
superimposed? I am also considering dissolving in the result as the numbers 
collide (with a glow to indicate the operation). Would it be an issue if the 
arguments morphed into the result, while you were tracking them?

The animation that you mention as the best, shows to me the importance of the 
lower 'blackboard' as a way to convey concrete examples of the operation (or 
mildly abstract in the case of the grid). The animations for me have two parts. 
The scalar examples (with the blackboard coordination) and the vector/matrix 
(without blackboard). It sounds like we are pretty close with the first part, 
although we may need to make adjustments to be consistent with the 
vector/matrix solution.

Cheers, bob

On -Mar19-2010, at -Mar19-20106:46 AM, Catherine Lathwell wrote:

> It sounds like you guys have come up with some sort of consensus that is
> something like, "It's OK if the animations are too fast to grasp the first
> time because people can watch them again".
> 
> I'm the dissenter, for sure.  In my field at least, I've never seen any
> animation praised for being too fast to understand the first time around.
> 
> Even if you watch TV, and see that The Simpsons is jam packed with funny
> references, the basic story is crystal clear on your first viewing.
> 
> About that covering plus: What's bugging me is that you obscure the number
> I'm thinking about while following your operation in my head.  I'm mentally
> checking to see if I "get it" while I follow along.  The next steps clarify,
> but you've already interrupted my train of thought.
> 
> I'm not convinced you need much movement at all, to be honest.  Just some
> modest graphical indication to signal that you are applying the operation in
> the moment. A bounce movement, a halo, a sparkle all do this.  (Which is why
> Iiked the annimation no one else liked because it is successful in
> indicating "pay attention here" in an appropriate way).
> 
> The best best best animation of all your animations is the one where the
> graph animation illustrates the actual mathematical point you are making.
> Here there is agreement between picture, the movement and the meaning you
> are communicating.  Consistent, elegant & economical.
> 
> C

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to