Hi,
numbers of the same magnitude are displayed differently while using the same display/mode settings.
How to reproduce:
Set DISP to:
ALL
RDX.
Set MODES to:
DEG
RECT
QUIET
CRES
KEY
Settings not shown here are not activated.
Enter:
CLST
34
55
/
5
SQRT
1

2
/
The result shown is:
Y: 6.18181818182e1
X: 0.61803398875
Background: There is an OFFIMG which shows the "Fibonacci Spiral" and I wanted to determine (just for fun) what the delta is between the value shown with the spiral and the "real golden cut...the most irrational number of the universe"
Both numbers are of the same magnitude.
Both are the results of arithmetic calculations.
And both are displayed differently.
I dont understand, why.....
And if there is no reason, this may be a bug.
Tested with my DM42 with firmware 3.3
Cheers
Meino
BUG? or my missunderstanding: Number display
BUG? or my missunderstanding: Number display
DM 42  SN: 00373, Firmware release v.:3.18. / 3.18. as compiled by SwissMicros
Re: BUG? or my missunderstanding: Number display
The 2 numbers are not of the same magnitude, though approximately the same, but this result is interesting.
Note that a real HP42S and Free42 (Android) both also do the same thing, so it's not a bug.
bob p
DM42: β00071 & 00282, DM41X: β00071, DM10L: 071/100
DM42: β00071 & 00282, DM41X: β00071, DM10L: 071/100

 Posts: 713
 Joined: Tue May 02, 2017 3:48 pm
 Location: United States
 Contact:
Re: BUG? or my missunderstanding: Number display
On the real HP42S, ALL mode always shows all significant digits, in fixedpoint notation if it can, and scientific notation otherwise.
The peculiarity arises from the fact that, for numbers with magnitude less than 1, it will show a zero before the decimal, and that zero counts towards the maximum of 12 digits that are shown of the mantissa.
Thus, 3 1/X shows 3.33333333333e1, but 0.33333333333 (eleven threes) is displayed without exponent.
Since Free42 has more than 12 mantissa digits, things are slightly different, but basically, what it does is round the mantissa to 12 digits, and then it applies the 42S rules.
And so: 5 SQRT 1  2 /, rounded to 12 significant digits, is 0.618033988750, and because of that final zero, it can be displayed in fixedpoint mode with no loss of precision compared to SCI 11 mode.
The peculiarity arises from the fact that, for numbers with magnitude less than 1, it will show a zero before the decimal, and that zero counts towards the maximum of 12 digits that are shown of the mantissa.
Thus, 3 1/X shows 3.33333333333e1, but 0.33333333333 (eleven threes) is displayed without exponent.
Since Free42 has more than 12 mantissa digits, things are slightly different, but basically, what it does is round the mantissa to 12 digits, and then it applies the 42S rules.
And so: 5 SQRT 1  2 /, rounded to 12 significant digits, is 0.618033988750, and because of that final zero, it can be displayed in fixedpoint mode with no loss of precision compared to SCI 11 mode.
Re: BUG? or my missunderstanding: Number display
Thanks for the explanation Thomas. I'll bet this is one of the odd little quirks of the 42S that you didn't guess right on the first try... Who would? But with your explanation, it at least makes sense now.Thomas Okken wrote: ↑Sat Feb 24, 2018 3:16 pmOn the real HP42S, ALL mode always shows all significant digits, in fixedpoint notation if it can, and scientific notation otherwise.
....
bob p
DM42: β00071 & 00282, DM41X: β00071, DM10L: 071/100
DM42: β00071 & 00282, DM41X: β00071, DM10L: 071/100

 Posts: 713
 Joined: Tue May 02, 2017 3:48 pm
 Location: United States
 Contact:
Re: BUG? or my missunderstanding: Number display
You offend me, sir! When I bought my HP42S, lo these many years ago, I set it to ALL mode as soon as I knew it existed (i.e. within minutes), and then it was only a matter of time (very little time) before I went "WTF is this thing switching to SCI for when I do 3 1/X?"
I got a bunch of little details wrong at first, but this wasn't one of them.
Re: BUG? or my missunderstanding: Number display
I salute you sir for your inquisitive nature!Thomas Okken wrote: ↑Sat Feb 24, 2018 5:36 pmI got a bunch of little details wrong at first, but this wasn't one of them.
There is just no knowing just what a truly OCD mind will do next, but I personally take great comfort in that!
bob p
DM42: β00071 & 00282, DM41X: β00071, DM10L: 071/100
DM42: β00071 & 00282, DM41X: β00071, DM10L: 071/100