EXC/ x<>Rn for stack efficiency

[Matt Agajanian]

(06-19-2017 05:44 PM)GrampaDave Wrote:  

Quote:Herman the moonshiner

But HP had better names, whose sly humor made them memorable, at least to native English speakers.

Radio Engineer Ann Tenor
Arctic explorer Jean-Claude Coulaire
Logistics specialist Justin Tyme

(I confess I made up that last one, but it's exactly in the HP style.)

Yea, these were what added to the fun & enjoyment of learning how to use our HP calculators.

On another note, as advanced and to compete with the SR-52, I wonder why HP did not include an EXC function on the 67, but waited until the HP-41.

[Thomas Okken]

(06-27-2017 09:42 PM)Matt Agajanian Wrote:  

(06-19-2017 05:44 PM)GrampaDave Wrote:  But HP had better names, whose sly humor made them memorable, at least to native English speakers.

Radio Engineer Ann Tenor
Arctic explorer Jean-Claude Coulaire
Logistics specialist Justin Tyme

(I confess I made up that last one, but it's exactly in the HP style.)

Yea, these were what added to the fun & enjoyment of learning how to use our HP calculators.

On another note, as advanced and to compete with the SR-52, I wonder why HP did not include an EXC function on the 67, but waited until the HP-41.

Because the HP-67 had "merged keystrokes," meaning every instruction was one byte. PANAMATIK posted this very enlightening link the other day: http://www.panamatik.de/ProgramCodes.pdf

With the HP-41C, HP let go of the one-byte-per-instruction thing, and that enabled them to do what TI was doing, i.e. allow register numbers from 00 to 99, and add new functions like EXC a.k.a. X<>.

In the HP-42S, they took this even further, adding RCL+ etc., and alphanumeric variable names.

[Matt Agajanian]

Yes, I saw that table when it was posted. Maybe I'm missing something in your explanation. How would [f] [x<>] n (or [x<>] n) consume more than one/two bytes? HP had room for [STO] +, -, x, / n. So. wouldn't the exchange function consume one/two bytes?

ADDENDUM

Okay, upon inspecting the HP-67 table, I think I see what you're saying.

Each instruction had to be given a one-byte hex code. After all was said and done, if I read the 67 table correctly, there are only four vacant spaces. So, there was no way to give 11 (0-9, (i) ) op codes for register exchange functionality. And, limiting exchange to four registers doesn't make sense. Right?