WP 34S: missing vector cross product



#17

I saw that the cross product is still missing in the vector library, so that's a good chance to write my first WP34s program. ;-)

Here's the pure code which fulfills all conditions required in vectors.txt (at least I hope so) - you just have to give it a label of your choice:

STO C
RCL* B
RCL T
RCL* T
-
x<>y
STO I
RCL* T
RCL A
RCL* D
-
x<>z
STO J
RCL* A
RCL B
RCL* I
-
R^
STO L
Rv

Just 20 steps, not so bad for my first program ... :-)

Franz


#18

Give me a break, I only added vectors.txt yesterday.

I'll add your code instead of writing my own :-)


- Pauli

#19

The vectors.txt file is updated. vectors.wp34s will be in due course I expect. I think that 112 steps is pretty lean for this lot :-) Thank to Franz for the final piece.


- Pauli


#20

I must say that writing this cross product was really a nightmare:

taking care of where all those stack values are moves during the calculation, avoiding stack overflows and finally fulfilling your conditions about storing the 2nd vector to L,I,J (and keeping the 1st one on the stack) increased the number of my grey hairs. ;-)

And with all these troubles I had an idea which would greatly simplify such calculations (with so many parameters on the stack):

What about having a 2nd copy of the 8 stack registers?

One command could duplicate the original stack into this 2nd copy, and then you could just do all calculations with these '2nd-stack' registers on the usual stack: since the order (and naming) of the 2nd-stack won't change, it would be much easier to adress the stackregister names.

Maybe this 2nd stack registers could be named with a dot preceding the names (.X, .Y, .... .C, .D) and adressed with the sequence RCL XEQ (because XEQ is almost the only key which is still free to use after a STO/RCL), which would then display "RCL .__" and wait for any stackregister name (X to D).

Of course this should also work with STO XEQ, and there should be 2 commands to copy (or backup) the normal stack to this 2nd stack and vice versa (maybe STO STO and RCL RCL).

Well, just a dream of me - but it would indeed make programming much easier ...

Franz

PS: I know this could be done with the usual registers Rxx (and STOS), but I also know that you don't like changing any registers during 'usual' calculations within a program.

Edited: 23 July 2011, 7:22 a.m.


#21

Not possible: no RAM left.

We do have save and restore stack operations: STOS & RCLS.
They save/restore the stack to the numbered registers only.


- Pauli


#22

Quote:
Not possible: no RAM left.

Ok, and what about a few more stack manipulation commands?

It would be nice to have something like Rv and R^ not for the whole stack but just for a part of it (e.g. only do a roll-down/up on the first 3 registers X,Y,Z).

Edited: 23 July 2011, 8:43 a.m.


#23

While we are thinking of stack commands :-) how about something like these?

Y <> Z

Y <> T

Z <> T


#24

Quote:
While we are thinking of stack commands :-) how about something like these?

Yes, not a bad idea - my HP-48GX has much more (useful!) stack manipulation commands than are available on the WP34S.

But the major problem is certainly the limited space on the WP34S, wheras the 48GX doesn't suffer from such space restrictions. ;-)
#25

And the other four stack levels??


- Pauli


#26

True true... but I do think at least these for the standard 4 levels would be useful.

How about a general <> function?

Is that possible?

X<> already prompts for the other location. Is it possible to code for both? :-)


#27

It isn't possible. The op-codes used for commands don't have enough space to hold two register addresses. So it would have to be a separate command for each at about eighty bytes per -- even supporting just swaps of Y, Z and T with anything will eat into our feeble memory reserve. It is possible I could streamline things and reduce the space use a little by reusing code profusely. I don't think supporting the specific individual functions will be any smaller.

I did have a general rearrange the (4 level) stack plus last X in a completely general manner instruction but it was lost. It replaced a lot of the traditional stack operations rather nicely but would have been really confusing for users.

There is hope however, define them yourself and stash them in a flash segment. For example:

    LBL'Z<>T'
X<> Z
X<> T
X<> Z
RTN

or

    LBL'Z<>T'
cpx Rv
x<>y
cpx R^
RTN


- Pauli

#28

Cleanest is to implement "stack-dup and stack-pop". Upon entry, your code does a stack-dup; this gives you a copy of the current stack and status. Do what you will (including calling other program segments.) When you're done, you explicitly write results into the lower stack and status registers, and then execute a stack-pop, which throws away all of your unsaved values, leaving the calling code with your results. Stack-pop gets really messy if you modify the structure of the "top stack".


#29

You will need a register area from the numerical registers for this. DUP is already there: STOS. Recalling only part of the stack can be done with R-CPY. R-CPY needs X for its argument so you need to take care of this in an extra step.

Pauli, a command ROLLv with an argument for the number of rolls might be helpful.


#30

Quote:
Pauli, a command ROLLv with an argument for the number of rolls might be helpful.

Number of rolls as in the number of times Rv is executed or the number of registers being rolled?

I suspect this suggestion will upset the RPL haters :-)


To be honest, I've never needed either nor even wanted the capability and I've done more than a little RPL programming in my time.

- Pauli


#31

Pauli, I meant the number of rolls to perform.


#32

Quote:
I meant the number of rolls to perform.

Well, I would say Rv and R^ only the first n stack-registers is more important. Performing n rolls (as you suggest) needs just pressing the same key (Rv or R^) n-times, but imagine the following situation:

You need the value from stack-register A during your calculation, and after that it's not needed anymore. In this case a "R^ A" would be very comfortable, wheras you would have to key in the following sequence to achieve the same:

x<>y
x<> Z
x<> T
x<> A
And you know how many keystrokes this would need

(especially the "x<> _") ... :-(

Franz


Edited: 24 July 2011, 3:56 p.m.


Possibly Related Threads...
Thread Author Replies Views Last Post
  [WP-34S] Unfortunate key damage with update to V3 :( svisvanatha 5 869 12-10-2013, 11:37 PM
Last Post: Les Bell
  WP-34S (Emulator Program Load/Save) Barry Mead 1 488 12-09-2013, 05:29 PM
Last Post: Marcus von Cube, Germany
  DIY HP 30b WP 34s serial flash/programming cable Richard Wahl 2 683 12-04-2013, 11:14 AM
Last Post: Barry Mead
  HP Prime: adding an entry to a vector Alberto Candel 12 1,323 12-02-2013, 01:18 PM
Last Post: Alberto Candel
  HP Prime - vector question bluesun08 3 474 11-18-2013, 07:26 PM
Last Post: Han
  [HP-Prime] - Missing confirmation before deletion Jean-Michel 11 1,142 11-14-2013, 05:47 AM
Last Post: Thomas Radtke
  WP 34S/43 ?'s Richard Berler 3 641 11-10-2013, 02:27 AM
Last Post: Walter B
  My FrankenCulator (wp-34s) FORTIN Pascal 4 660 11-09-2013, 06:18 PM
Last Post: FORTIN Pascal
  WP 34S Owner's Handbook Walter B 5 846 11-09-2013, 05:34 PM
Last Post: Harald
  wp 34s overlay and programming. FORTIN Pascal 6 857 11-08-2013, 01:28 PM
Last Post: Nick_S

Forum Jump: