Well, in this case, the code had been in a loop. The loop got large, so moved contents out, but then the loop control words no longer worked. In this case, the loop control words made it clear the derived sub, where the code in the main loop would go next. I used those constants (and have done similar in the past) for exactly those reasons -- even though the loop contents were no longer "inline", due to the use of such constants to indicate what they did in their previous loop the code was actually easier to understand -- more clear.

If I had come up with a new or different set of constants, I don't think it would have been as clear. As for constants being upcased? That's so 80's. :-) Seriously -- all caps hurt my eyes and fingers typing -- I'll usually capitalize something, but there are many cases where doing so hurts perception and understanding.

An example would be several math constants (displaying these on pm in code blocks, is bugged, so you'll have to use your imagination a bit: ( But the 3 chars are Φ (Phi), ɸ (phi) and π (pi) ).

  sub Phi () { .5 * (5.**.5 - 1) }   sub Φ  () { goto &Phi }
  sub phi () { .5 * (1. + 5.**.5) }  sub ɸ  () { goto &phi }
  sub pi  () { 4 * atan2(1, 1) }     sub π  () { goto &pi }
[download]

If you upper-cased these, how would you tell Phi from phi. In the greek names (on the right which display as characters on modern systems), they are also the upper and lower case version of the greek phi. Pi, uses a lower case 'pi' just as it is in the english text version.

Anyway, it's all about context... ;-)...