Hello Chris Dean,
Maybe you are right concerning to the flaw. But I have my own reasons to consider the behaviour of the sigma function of HP 17BII+ a flaw. If it is not a flaw, we have to admit that the function sigma of the HP 17BII and HP 19BII is not correct. But this hypothesis is false because the sigma function works fine in these calculators.
The following equation runs in the HP 17BII and HP 19BII without problems.
SUM = SIGMA(N:1:N:1:N)
With the HP 17BII, the variables N and SUM appear in the menu. The same happens with the HP 19BII. Also, the functions L() and G() works fine in these calculators. But, in the HP 17BII+, the variable N does not appear in the menu and I have to do magic to get the same result I can get with the L() and G() functions with the HP 17BII and HP 19BII. Besides, in my opinion, the HP 17BII+ had to mantain compatibility with the HP 17BII and HP 19BII. Because it did not mantain compatibility, I had to re-write all programs to run in this new calculator.
In the HP-27S/19B Technical Applications Manual, HP says that the "counter variable" of the SIGMA function is a "local variable" and thus it did not appear in the menu. If this is true, the SIGMA function in the HP 17BII and HP 19BII were not implemented correctly. Of course, this is FALSE because, as I said, this function works flawlessly in these calculators. So, this USELESS RULE that says the counter variable of the SIGMA function does not appear in the menu led to the FLAW in the SIGMA function of the HP 17BII+. Happly, the HP engineering did not implemented this "technical rule" (sic!) in the HP 17BII and HP 19BII.
I think there were a flaw in the implementation of the sigma, L() and G() functions of the HP 17BII+. Otherwise, these functions would work as fine as in the other calculators it replaced.
Regards
Iracildo