Thank you Skip, I was finding myself really resisting your array solution until you mentioned the application to more complex J sentences. Then I thought the sliding matrix would still be confusing, but I could see the power in representing consecutive operations.
What if instead of sliding the matrix along the operators right to left, we highlighted the token (either argument y or the current operator; x and y and the current operator for dyads) and projected the current value of the result below the line. Thus the values of y (and x) would stay unchanged and we could see the result transformation in place as the highlight moves down the line. I'll try to get an example up today, but I think this is very promising. Thank you so much for your persistence with this idea. Cheers, bob On -Mar17-2010, at -Mar17-201011:55 PM, Skip Cave wrote: > Bob, > > I still think leaving the plus alone on the left, and moving the right > argument as a ghost over on top of the plus, shows what is happening > more clearly, while keeping some semblance of the original J code. This > can also extend to vectors and matrices. With vectors and matrices, you > don't have to replicate the plus to line up with all of the numbers. > Just center the vector or array of numbers over the left plus, and then > make the conjugation, solidify the array, and slide it down below the > original > > I believe that all of the videos should display the initial equation as > close as possible to the actual J code, except with the numerical vector > or array replacing the variable name. Then the right argument should be > ghosted and moved either over the operator, or the left argument > depending on whether you are showing a monadic or dyadic function. The > result can then be shown, un-ghosted, and slid below the original > equation. In all cases, the original equation is left alone. with the > final result shown below it > > One interesting ramification of this process is that more complex J > expressions can be demonstrated using the the same paradigm. Starting > with the right argument, it is ghosted and moved to the left, the first > function is performed, the result solidified, and then ghosted again, > moved to the left and the next function performed, solidified, ghosted, > moved, etc. When the final answer is arrived at, it is solidified and > moved below the original equation. This can also clarify the > right-to-left evaluation process as well. > > This "ghost-moving-solidifying" process can be used to demonstrate what > is going on with all of the primitives, as well as more complex J > expressions. In all cases the start is the J code, which should never > change. > > Skip Cave > > bob therriault wrote: >> Hey everyone, >> >> The longer and larger version of Conjugate is now posted at: >> http://www.jsoftware.com/jwiki/Vocabulary/plus >> Comments regarding speed, movement, content and clarity are welcome >> (especially from newcomers). Is this a useful animation? >> >> Cheers, bob >> >> On -Mar12-2010, at -Mar12-20108:17 AM, bob therriault wrote: >> >> >>> I love that quote Catherine, >>> >>> Going back to your frustration at the 'hiding' of the Conjugate (+) >>> operation in vectors and matrices, I realized last night that the solution >>> may not be in showing the way the individual instances of Conjugate work on >>> each atom of a matrix, but by using more scalar examples initially to show >>> a wider range of inputs and outputs for Conjugate. >>> >>> Today, i'll work at putting together a 'Club Mix' of the shorter animation. >>> It'll have more examples, be larger and run a little longer. This might be >>> a bridge between the short reference version and the longer tutorial >>> version that may include interaction and/or voice-over. If both versions >>> were kept on the reference page the user would have a choice between a >>> quick check, or a more extended description. >>> >>> Cheers, bob >>> >>> On -Mar11-2010, at -Mar11-20104:14 PM, Catherine Lathwell wrote: >>> >>> >>>> To quote Ken, when I first talked to him about the doc project... "Maybe >>>> all >>>> that Art School is good for something after all!" :) >>>> >>>> shhhhhh.... don't tell I didn't know how J does minus! :) >>>> >>>> C >>>> >>>> On Thu, Mar 11, 2010 at 7:00 PM, Devon McCormick <[email protected]> >>>> wrote: >>>> >>>> >>>>> Catherine - APL uses the same character for the same thing - it's just >>>>> positioned differently. >>>>> >>>>> BTW - your feedback is particularly valuable - please continue... >>>>> >>>>> Regards, >>>>> >>>>> Devon >>>>> >>>>> On Thu, Mar 11, 2010 at 6:28 PM, Catherine Lathwell <[email protected] >>>>> >>>>>> wrote: >>>>>> >>>>>> What's the deal with the _ character? Is this how J works these days? >>>>>> >>>>> I'm >>>>> >>>>>> not really up on the J vocabulary, so I didn't understand conjugate >>>>>> >>>>> before >>>>> >>>>>> this... so it took me a couple of runs to figure out that it's *minus >>>>>> >>>>> *you >>>>> >>>>>> mean. My first few rounds, I was scratching my head trying to figure out >>>>>> what the heck you need an underscore in your number for. >>>>>> >>>>>> ... >>>>>> >>>>> -- >>>>> Devon McCormick, CFA >>>>> ^me^ at acm. >>>>> org is my >>>>> preferred e-mail >>>>> ---------------------------------------------------------------------- >>>>> For information about J forums see http://www.jsoftware.com/forums.htm >>>>> >>>>> >>>> >>>> -- >>>> Catherine Lathwell >>>> http://www.aprogramminglanguage.com >>>> ---------------------------------------------------------------------- >>>> For information about J forums see http://www.jsoftware.com/forums.htm >>>> >>> ---------------------------------------------------------------------- >>> For information about J forums see http://www.jsoftware.com/forums.htm >>> >> >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm >> >> >> > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
