It was not accurate for me to say that the trains-only technique eliminates the "inside, outside" design decision. I think I was not mistaken in saying that teaching under such emphasis would not assist people in reading existing code that has relied on conjunctions to compose verbs. And, of course, such composition could (and do) occur in any tine of a train, or in the absence of a train.
--Tracy On Wed, Nov 30, 2011 at 9:57 AM, Kip Murray <k...@math.uh.edu> wrote: > f ([: h g) e > > On 11/30/2011 8:34 AM, Tracy Harms wrote: > > I agree with R.E. Boss that the language is crippled, not improved, by > > avoiding composing conjunctions. > > > > On Wed, Nov 30, 2011 at 2:37 AM, R.E. Boss<r.e.b...@planet.nl> wrote: > > > >> > >> > >> IMO the role of @ is essential and by removing it you amputate J. > > > > ... > >> > > > > What stands out to me as the greatest loss is that a very common > phrasing > > decision is eliminated. I think of this as the "inside, outside" > decision, > > and it's about where to write a monadic verb that takes the result of a > > fork. It can be put outside the fork using [: > > > > [: h f g e > > > > or it can appear inside the fork using @: (or another compositional > > conjunction) > > > > f h@:g e > > > > There are situations where "h" and "g" are more readily thought of > > together. In those situatons, using a composition in the center tine is a > > boon. There are other times where "h" seems nicer posed against the train > > as a whole, so Cap works well. There are no simple rules for resolving > > these differences, and they're mainly aesthetic. > > > > Programmers who learn tacit form will be up against this pattern sooner > or > > later. I can't imagine anybody becoming even modestly fluent without > > learning to see v2@:v1 as a single verb, and so recognize the "inside" > form > > as a single fork, while the "outside" form is two forks. > > > > My strongest misgiving about the idea being called "Simple J" is that it > > will not fit facilitate reading. In my own learning process I read far > more > > J than I wrote, and in the early stage most of what I read was not within > > the range of what I could have written. An introduction to the language > > that omits compositional conjunctions will not prepare people for reading > > the J code that is actually out there to be read. It won't adequately > help > > somebody who's learning the language in order to make changes to an > > existing J application. And it might be taken to suggest that the > > difficulties encountered in learning J are due to programmers not having > > kept things "simple" by using only verb trains. Difficulties will arise, > > but that won't be their cause. > > > > --Tracy > > ---------------------------------------------------------------------- > > 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
