Tracy -- thank you for taking the trouble to wrestle with problems like this. It wasn't in my mind to ask this of the J community.
Primarily I shall use my record of Chat Forum traffic as a checklist when revising the printed book. This is where Chris Burke directed the APWJ business. I understand the management does not encourage cross-posting. But it will help me if any APWJ problems aired on the Programming Forum get a 1-line pointer in Chat to warn me they're there. Having said that, when I come to do the final revision I shall of course cast my eye over Programming Forum as well. Ian (copied to both Chat & Programming) On Thu, Apr 2, 2009 at 11:41 PM, Tracy Harms <[email protected]> wrote: > I probably should have posted this to Programming because it's getting > technical. > > > ---------- Forwarded message ---------- > From: Tracy Harms <[email protected]> > Date: Thu, Apr 2, 2009 at 3:38 PM > Subject: Re: [Jchat] At Play With J: Boggle > To: Chat forum <[email protected]> > > > Thanks Bill, Roger, and Kip for the assistance. > > In the Boggle article I see some additional difficulties. Some are > around the he paragraph that reads: > "Here are four successive items from pc: " > > The noun referred to there (pc) is defined as a local noun within a > larger explicit verb (paths). It would thus not actually be available > for inspection in the manner shown in the article. > > (1) I propose changing the copula to pc=: within the definition so > that using the array on its own makes sense. I think an end note may > be appropriate on that change. > > The results of that execution don't fit the context indicated. The > prior J sentence was '#3 paths 4' but the results indicate that the > context (state) is due to '#4 paths 4' > > (2) I'm not sure how to correct this. It may be best to alter the > prior example so that it sets that context. Or perhaps it would be > better to add '#4 paths 4' as a second example. > > Late in the paper is "an incomplete table of the number of paths of > given lengths in grids from size two to five." > > (3) A minor option: We can add the missing two values if we wish, and > perhaps notes on execution speeds as of this later date. (I get > 1110000 for #5 pathb 8 but an out-of-memory error for #5 pathb 9) > > I'm more concerned by the discrepancy of results between the two verbs > that are intended to be identical, paths and pathb. The only input > where I've seen matching output is 4 (paths -: pathb) 4 . This > warns us that one of the two functions is inaccurate. Identifying that > error and deciding how to correct in in the reprinting go beyond what > I can do alone. > > (The most visible aspect of that discrepancy is that the result > reported in the '#3 paths 4' example does not match the result shown > for the same values in the table.) > > Plenty to think about here. > > 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
