Moved from the thread "[Jgeneral] Looking for links - online or off - between J and natural language (example: English)."
Feedback regarding this screencast: http://bobtherriault.wordpress.com/2010/10/10/animation-of-visual-j-interface/ Hi Bob, Responses in line... > From: bob therriault > Sent: Wednesday, 13 October 2010 13:16 > > Thanks for the great feedback Ric, > > I'll give you my responses/rationalizations, keeping in mind that at this > early stage there will be many changes. > > 0) The representation of the noun argument (as opposed to a noun that might be > attached to a conjunction) is certainly up for discussion. At one point I > thought I would focus more on the sentence itself without the argument, but > thought that having the intermediate values was more useful and that of course > required an argument. Changing the linking to a curved arrow would certainly > differentiate the application to an argument from the result of a verb. I think that is important to include an argument/arguments - otherwise it will be confusing as to whether some verbs in the sentence are the monadic or dyadic instances. > 1) I am guessing that I put the argument on the bottom and the result on top > through the influence of the Roger Stokes diagrams that I had seen in Learning > J. I don't have any particular attachment to either direction, but must admit > that I find the sideways 'tree representation' of J verbs difficult to wrap my > head around. As an aside, since I am in Canada and I think you are in New > Zealand, it would stand to reason that we would invert each others views ;) That's one way of looking at it, on the other hand I imagine it would be even more logical for you northern-hemispherians to start reading a page at the top than for us down here ;) > 2) I chose yellow for / to represent an adverb working on a verb. The verb + > is blue (I felt it was working dyadically within the single argument through > the influence of /). The result of the +/ does get represented as a monadic > verb when it is projected into the fork. I felt that the composite verb +/ > should be broken out into yellow (adverb) and blue (dyadic verb) rather than > representing them in green (which is the view that the fork has of them). I agree with breaking out +/ into the yellow (adverb) and blue (dyadic verb), my comment is more about the fact that the background colour for the whole construct is also yellow. The +/ construct is not an adverb and so to me it is confusing that it has a yellow background. > 3) I have always thought of trains as a type of implicit conjunction. I think > it came from Henry Rich's J for C programmers [1], although to be accurate he > refers to forks and hooks as modifiers. I agree that the result is a verb, but > then the result of u...@v is also a verb (and @ is a conjunction). Forming > the > train as a conjunction, lets me imagine verbs taking on different roles > according to their position, the same way a true conjunction would affect they > way they interacted with the arguments. If the fork 'conjunction' was part of > another conjunction, it would be represented as a verb in the same way that +/ > was represented by green square inside it even though it is composed of + and > / . Believe me, I'm way out on a limb with this one, so I really look forward > to more discussion. I see what you mean, the fork construct (and trains in general) acts as a type of conjunction - combines other parts of speech. However I think the confusing thing for me is that you are using the background colour of a shape to show/represent more than one thing - sometimes it represents the part of speech of the word(s) it contains, other times it says something about how the words it contains are combined. For example if you were presenting the construct u...@v then the @ would have a red background and the u would have a green background and the v would have either a blue or green background (depending on whether there were both left & right arguments or not) but the u...@v construct as a whole is (as you say) a verb and therefore for consistency to my mind it should have a blue (or green) background. If you want to differentiate forks perhaps it would be better to use the shape of the container or the type of border instead of the background colour? > Cheers, bob > > [1] http://www.jsoftware.com/help/jforc/compound_verbs.htm#_Toc191734374 > On 2010-10-12, at 3:01 PM, Sherlock, Ric wrote: > > > Bob, > > I think you're right that a simple, attractive, interactive parser for J > sentences may ease the introduction of newcomers to some of J's concepts. > > > > Here is some feedback: > > > > I like the visual "explosion" of a placeholder to its components, however I > find the similarity of the visual effect linking the gray graphic > (representing the Noun) to the monadic verbs +/ and # confusing. Perhaps > something more directional (like an curved arrow) could be used to show where > the noun is inserted into the sentence. > > > > IMO it would be more intuitive if the diagram worked from top to bottom > rather than bottom to top (iow the arguments were fed in from the top and the > result produced at the bottom). > > > > Why does the exploded graphic of +/ have a yellow background? It is a > monadic verb in this sentence, so shouldn't it have a green background? > > > > You describe a fork as a conjunction. Do you mean this in the same way as > say @ is a conjunction? If so then I think this is incorrect. According to the > dictionary, a train of any length is a verb not a conjunction [1] in which > case the box with the red background should also have a green background (it > is a monadic verb). > > > > [1] http://www.jsoftware.com/help/dictionary/dictf.htm > > > >> From: bob therriault > >> Sent: Wednesday, 13 October 2010 09:45 > >> > >> Hi Devon, > >> > >> If the topic of representing array thinking comes up, the last three posts > on > >> my blog express some fumbling about I've been doing in the last two weeks. > >> > >> http://bobtherriault.wordpress.com/ > >> > >> Most recent is a screencast of a J 'training-wheels' environment that may > be > >> achievable with the new beta. > >> > >> Cheers, bob > >> ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
