# Talk:Solve the no connection puzzle

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

## Video

According to the video, A and B, G and H are also connected. Fwend (talk) 13:46, 11 April 2016 (UTC)

## REXX formatting seems off

Near the top where it starts "aid=...", the lines after that seem mid aligned? (But I am viewing on my tab, but then again, it is usually OK)? --Paddy3118 (talk) 03:48, 5 October 2014 (UTC)

The REXX program no longer has that statement, perhaps you were viewing it whilst it was being updated?   Does the current version still have "off" formatting?   It looks OK to me, but then, it always did view OK to me at this end. -- Gerard Schildberger (talk) 06:48, 5 October 2014 (UTC)

Hi Gerard, I was talking specifically about the second section of code starting with the do-while loop that is all indented around sixteen spaces? <lang rexx>/**REXX program solves the "no-connection" puzzle (with eight pegs).*/ @. = @.1 = 'A C D E'  ; @.5 = 'E A B D F' @.2 = 'B D E F'  ; @.6 = 'F B E G' @.3 = 'C A D G'  ; @.7 = 'G C D E' @.4 = 'D A C E G'  ; @.8 = 'H D E F'

```                  do nodes=1  while @.nodes\==;   _=word(@.nodes,1)
subs=0              /* [↓]  create list of node paths*/
do #=1  for words(@.nodes)-1
__=word(@.nodes,#+1);  if __>_  then iterate
subs=subs+1;           !._.subs=__
end  /*#*/
!._.0=subs          /*assign the number of node paths*/
end   /*nodes*/
```

</lang> --Paddy3118 (talk) 08:06, 5 October 2014 (UTC)

(No L in Gerard).     Yes, the   DO NODES=1 ...   starts in column 20, the   DO #=1 ...   starts in column 31, the indentation structure depending on what (I think) looks better (and not using any hard and fast rules) and conveys the logic structure and/or purpose of the DO loops.   I sometimes use a great-indent (as used here) for some smaller DO loops, but normally I only indent (each) DO loop and SELECT structures two spaces. -- Gerard Schildberger (talk) 08:46, 5 October 2014 (UTC)
Hi Gerard, what is the indentation aligned to that makes it look better? Is see nothing immediately above the "do nodes =1" and nothing immediately below the "end /*nodes*/" that continues the visual alignment as well as no indication that that block is a sub block of the preceding code line? --Paddy3118 (talk) 10:47, 5 October 2014 (UTC)
The whole point of my indentation is to indent (offset) the code to not continue the visual alignment with the code before and after the (greatly) indented code (that is, to make it very distinctive from the "mainline" code).   Another alternative would be to write a subroutine for the indented code (which has its own drawbacks).   The indentation of the DO loops is itself an indication of what it is, a block of code using a DO loop construct (and in this case, with another DO loop within it, also greatly indented).   Some programmers don't indent the DO statement at all; that isn't my style of coding.   I prefer to use indentation instead of relying on the DO and END statements standing on their own --- to me, the indentation (in this case, great indentations) serve that purpose better.   It's all very subjective, of course, and I tried to keep the reader of unfamiliar code forefront in mind.   I'm a firm believer in matching (perfectly aligned) indented code (as opposed to Egyptian indentations).   Matching/aligned indented code (which is evident in the programs that I code) makes it easier for persons unfamiliar with the code to peruse the program logic.   Using (matching/aligned) indented code helps the reader to bypass (eyeballing) that section of code on re-examinations of the program, making it easier to "jump or skip over" the DO loop construct on subsequent re-readings.   The idea/concept of (program logic) indentation is a very volatile and can be an argumentative subject.   Any discussion on program code indentations (akin to a religious or political debate, editor wars come to mind) can be very confrontational when being discussed or its use validated. -- Gerard Schildberger (talk) 20:02, 5 October 2014 (UTC)
Hi Gerard, if you see a lot of this type of indentation style in REXX code from other authors too then that makes it idiomatic code which is the best sort for RC. Thanks. --Paddy3118 (talk) 03:55, 6 October 2014 (UTC)