RPN Question - Printable Version

RPN Question - Printable Version

+- HP Forums (https://www.hpmuseum.org/forum)
+-- Forum: HP Calculators (and very old HP Computers) (/forum-3.html)
+--- Forum: General Forum (/forum-4.html)
+--- Thread: RPN Question (/thread-9363.html)

RPN Question - Gamo - 10-25-2017 11:11 AM

To divide x by 2
ENTER 2 ÷
What other keystokes will do the same that use only functions and No numbers except number to enter ?
No (÷ x + -)

Gamo - I got answer

RE: RPN Question - Paul Dale - 10-25-2017 12:10 PM

The WP 34S possibly has the shortest solution at two functions:

Code:

yards->m

m->fathom

More general is a statistical approach:

Code:

CLSigma

Sigma+

CLx

Sigma+

x-bar

Neither the arithmetic operators nor digits are required in either sequence.

Pauli

RE: RPN Question - Gerson W. Barbosa - 10-25-2017 12:29 PM

(10-25-2017 12:10 PM)Paul Dale Wrote: The WP 34S possibly has the shortest solution at two functions:

Code:

yards->m m->fathom

Also, if losing T is not an issue:

Code:

ENTER 

||

Gerson.

RE: RPN Question - Paul Dale - 10-25-2017 12:43 PM

My first thought was the || function but I then made a mistake trying it and moved on.

Pauli

RE: RPN Question - jebem - 10-25-2017 01:05 PM

(10-25-2017 12:29 PM)Gerson W. Barbosa Wrote: Also, if losing T is not an issue:

Code:

ENTER ||
Gerson.

Am I wrong, or it is the Electronics Engineer in you that came up with that elegant answer? Smile

RE: RPN Question - Didier Lachieze - 10-25-2017 01:10 PM

For small numbers (between -99 & 99) you can do: [10^x] [√x] [LOG]
You can extend the range to [-230.25, 230.25] with: [e^x] [√x] [LN]

RE: RPN Question - Gerson W. Barbosa - 10-25-2017 01:33 PM

(10-25-2017 01:05 PM)jebem Wrote:
(10-25-2017 12:29 PM)Gerson W. Barbosa Wrote: Also, if losing T is not an issue:

Code:

ENTER ||
Gerson.

Am I wrong, or it is the Electronics Engineer in you that came up with that elegant answer?

Here is a funny example of different coding for the same problem, according to the authors’ professions:

http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/archv015.cgi?read=74095

(See link in message #3)

BTW, my solution for that one was “3 XOR” (valid for the HP-42S only). No, I am not logic circuit designer :-)

Gerson.

RE: RPN Question - Gerson W. Barbosa - 10-25-2017 01:37 PM

(10-25-2017 01:10 PM)Didier Lachieze Wrote: For small numbers (between -99 & 99) you can do: [10^x] [√x] [LOG]
You can extend the range to [-230.25, 230.25] with: [e^x] [√x] [LN]

Nice solution for small arguments! I’d thought of “LN SINH”, but this will work only for very large arguments.

Gerson.

RE: RPN Question - Gamo - 10-25-2017 01:54 PM

Didier is right my intended answer is [e^x] [√x] [LN]

Someone also suggest ENTER 50 % is very good idea.

RE: RPN Question - emece67 - 10-25-2017 02:21 PM

(10-25-2017 12:29 PM)Gerson W. Barbosa Wrote:

Code:

ENTER ||

This seems hardly beatable to me. Other approach still in the wp34s may be:

Code:

LB

DEC X

2^X

limited, of course, to arguments >=0.

Also:

Code:

1/x

RCL+ X

1/x

This last one keeps T intact, updates L in the right way, and also works with arguments <=0:

Code:

STO L

STO- L

INC L

INC L

RCL/ L

Regards.

RE: RPN Question - Craig Bladow - 10-25-2017 06:32 PM

Solution for HP 41CX

Code:

<   (Enter this in Alpha mode)

ATOX

NUMBER

P-R

RE: RPN Question - Dieter - 10-25-2017 08:59 PM

(10-25-2017 12:10 PM)Paul Dale Wrote: The WP 34S possibly has the shortest solution at two functions:

It could have the definitely shortest solution with one single function call if (if!) the HALF function proposed back then would have made it into the 34s firmware. ;-)

Yes, I still would appreciate such a function, and I think it would nicely complement STO+X or RCL+X for doubling X.

Dieter

RE: RPN Question - Craig Bladow - 10-29-2017 02:43 AM

(10-25-2017 02:21 PM)emece67 Wrote:

Code:

1/x RCL+ X 1/x

Since the 41C doesn't have RCL+

Code:

1/x

ST+ X

1/x

RE: RPN Question - Dieter - 10-29-2017 07:44 AM

(10-29-2017 02:43 AM)Craig Bladow Wrote:
(10-25-2017 02:21 PM)emece67 Wrote:

Code:

1/x RCL+ X 1/x

Since the 41C doesn't have RCL+

Code:

1/x ST+ X 1/x

Note that this method may cause roundoff errors: 7 → 3,499999999.

Dieter

RE: RPN Question - Craig Bladow - 10-29-2017 09:42 PM

(10-29-2017 07:44 AM)Dieter Wrote: Note that this method may cause roundoff errors: 7 → 3,499999999.

I believe we have found machine epsilon for the 41, 1x10^-9.

Since HP could have achieved a smaller epsilon if they had used binary vs. bcd format in the HP-41, there must have been some other tradeoff or legacy design that led them to use BCD.

RE: RPN Question - Mark Hardman - 10-29-2017 09:47 PM

[Image: attachment.php?aid=211]

RE: RPN Question - rprosperi - 10-29-2017 11:49 PM

(10-29-2017 09:47 PM)Mark Hardman Wrote:

Thank you.

I'll leave it there...