Programming exercise (RPL/RPN) - Reciprocal Fibonacci Constant

[Thomas Okken]

(02-17-2017 02:09 PM)Gerson W. Barbosa Wrote:  

(02-17-2017 08:03 AM)Werner Wrote:  How can the 42S code be 28 bytes, 41-compatible, yet the 41CV code is 33 bytes?

I think it's strange too, but that's what I got on my HP-41CV with CAT 1 and the printer on, after PACKING. Perhaps I should do a manual byte count.

The HP-42S' byte count doesn't include the END, which of course is cheating, because that END (or .END.) really does take up three bytes.
That still doesn't explain the discrepancy, though, and there is another factor that should make the 41 version one byte shorter than the 42S version, not two bytes longer: on the 42S, numbers are always preceded by a null byte, while the 41 only inserts a null byte between consecutive number lines.