Thanks Skip,
I'll work on delaying addition. It is amazing how much difference a tenth of a
second makes to some of these processes.
I am tempted to add the length error example, but the higher dimensions will
probably wait for either other primitives that are more germane i.e. Rank (")
or a tutorial animation down the road. The J language has so many facets that I
feel the information could get pretty muddy if we try to show too many concepts
in one animation.
Cheers, bob
On -Mar10-2010, at -Mar10-20106:27 AM, Skip Cave wrote:
> Bob,
>
> Very clean.. I think you've got it. Your latest video shows what is
> going on, with minimal extraneous effects. As far as I'm concerned, you
> have created the basic design template for video presentation for all
> the dyadic primitives.
>
> A small nit: It might be better to have a slightly longer delay before
> solidifying the sums. It is happening so fast now, it doesn't show the
> add process as a discrete step. It just all runs together.
>
> Bob said:
> The only changes that I would make is that I would reduce the 3X3
> matirix to 2X2,
>
> Skip replies:
> a 2x2 matrix is fine, but I kind of like the 3x3. I have no logical
> rationale for that, however. It might be even better to use different
> length dimensions in each step, to show that there is nothing special
> about 3.
>
> A few more options to continue the video could be:
>
> a 3-long vector plus a 2-long vector, giving an error, demonstrating
> length mismatch.
>
> A 3x3 matrix plus a single scalar, where the additional places are
> ghosted before adding. This would show that the reshape/duplication of
> scalar arguments extends to higher dimensions.
>
> A 3x3x3 array added with another 3x3x3 array. This gives you a chance to
> show off your 3-D modeling skills designing 3x3x3 blocks of billiard
> balls with numbers. You probably need a slightly oblique view of the
> 3x3x3 block so the viewer can see all the dimensions.
>
> A 3x3x3 array added with a scalar, where the scalar is ghosted to a
> 3x3x3 array before adding
> .
> I believe that this simple video sequence of summing processes gives the
> newbie a great introduction into J array processing. A similar video
> should be created for each primitive for the NuVoc. For some, a visual
> demonstration of the principals is much more compelling than any amount
> of text discussion. I have always felt that the best J programmers keep
> visual models of what goes on, in their heads.
>
> This style of video gets the novice off on the right conceptual foot for
> any.J primitive It should probably be required viewing for the newbie
> when they arrive at a primitive's page in the NuVoc.
>
> We also need to create a similar video for the monadic forms of the
> primitives.
>
> Skip Cave
> .
> bob therriault wrote:
>> Hi Skip,
>>
>> I came to the same conclusion as you suggested below. I just posted the
>> results on Jwiki Plus (+) NuVoc:
>> http://www.jsoftware.com/jwiki/Vocabulary/plus
>> It's the last video in the list. The only changes that I would make is that
>> I would reduce the 3X3 matirix to 2X2, but let me know what you think the
>> next iteration should look like! :)
>>
>> Cheers, bob
>>
>> On -Mar9-2010, at -Mar9-20102:36 PM, Skip Cave wrote:
>>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
