More HP 33S algebraic bugs? Try these..

[Matt Kernal]

Inspired by Gene's findings in his post "HP33S question:POLL - Try this in algebraic mode", I found a few more strange results on my S/N: CNA4100... 33S.

** WARNING ** The following key sequences involve CLEARING your 33S's memory. Only those who DON'T CARE IF THEY LOSE THEIR CALCULATOR'S CONTENTS should try these steps **

1. Switch to algebraic mode, and make sure the stack/display is clear by pressing the <C> (ON) key.

2. Using Gene's number, press: 24 then press the Minus key.

3. Your display should have "24-" in the top line and "24.00" (assuming FIX 2) in the bottom line.

4. Now clear the calulator's memory by pressing the green <leftshift> key, followed by the <clx> key (you know, the <backspace> key just below TAN). Using the big "wobble" key, navigate to: 3ALL, and press <ENTER>. At the "CLR ALL? Y N" prompt, navigate to "Y" and press <ENTER>

5. Finally, press: 24 then press the ADD key. What do you see? On my calculator, the top line is "-24+" and the bottom line is "-24.00". What's up with that?!?

Here's another one:

1. Do step 1 above.

2. Using Gene's number, press: 24 then press the Divide key.

3. Your display should have "24/" (/ is the divide symbol) in the top line and "24.00" (assuming FIX 2) in the bottom line.

4. Do step 4 above.

5. Finally, press: 24 then press the Divide key (or the Multiply or Minus or Plus key). What do you see? On my calculator, the top line is "0/" and the bottom line is "0.00". BTW, if you had pressed the any of the other keys (Multiply, Minus, or Plus), the top line would be "0x", "0-", or "0+", respectively, and the bottom line is "0.00").

I know these are really, really obscure key sequences, but aren't these bugs? I wonder if the original Walmart units are the same/different?

Matt

[John]

The HP41 had bugs that happened in obscure situations too.

Not every bug is serious enough to warrant a big deal. They are often curiosities, nothing more.

[Les Bell]

Quote:

---

I know these are really, really obscure key sequences, but aren't these bugs? I wonder if the original Walmart units are the same/different?

---

I wouldn't say these are really bugs, so much as inelegantly handled errors. The problem is that when you press "24 -" the calculator is now in an intermediate state. It expects the next thing you input to be a (what? term? - it's so long since I studied parsers, recursive descent, etc) number, or something that can be evaluated to a number, like sin() or open-paren. When you clear the memory right in the middle of an incomplete expression, this is the equivalent of a syntax error - you shouldn't be expected to do that.

Obviously, the calculator is not clearing the algebraic stack state correctly for some reason. It's annoying to us, because in RPN, the calculator maintains all state information in the form of the numbers on the stack, and there are no pending operators or other state information hidden away anywhere - so that type of error can't really arise.

Best,

--- Les Bell

[http://www.lesbell.com.au]