Little curiosity: why the fourth stack register is called "T"?
#1

Hi all.
I was reading some old issues of the HP Solve newsletter, with historical presentation of RPN and its comparison with the other methods.
I have a little curiosity, is there a special meaning for the letter "T" used for the fourth stack register? I can understand the choice for X, Y and Z, but why T?
Thanks.

#2

Hi,

this may be completely incorrect, but when I got my first introduction into the wonderful world of RPN calculators back in 1983 by my maths teacher (bless his soul) I was taught that 'T' stands for 'Top'.

Best,
Glenn
E=mc² +/- 3 dB

#3

Top

#4

I don't know if this is definitive, but the HP-33 manual says:


Quote:
The "oldest" number is stored in the T– (top) register. The stack is the work area for calculations.

#5

T is for top!

#6

have a look to http://www.hpmuseum.org/hp9100.htm

#7

The T register was introduced with the HP-35.

To my knowledge it means "the Top register".

Sylvain

#8

Temporary register.

#9

Thanks Nina,
this perfectly answers to my doubts.

Thanks also to all the others who gave their opinions.

#10

Visiting a museum, you may leave the cafeteria and walk through the other rooms. Please do so. You will find lots of interesting exhibitions and information. Coffee's not everything.

d;-)

#11

Interesting to read that register T is the Top stack register. It is so the way it is visually depicted in the manuals. In truth, the HP calculator stack is inverted. The true Top stack register is the X register since it receives new values and pops off old ones!!!

Namir

#12

Quote:
Temporary register.

That is how I remember it. It is "temporary" because some internal calculations -- transcendentals, I believe -- can corrupt it.

#13

Quote:
Visiting a museum, you may leave the cafeteria and walk through the other rooms. Please do so. You will find lots of interesting exhibitions and information. Coffee's not everything.

d;-)


Yes, in this period, also thanks to the nice emulators available, I am revisiting many things concerning old calcularors, and this museum is very useful.

#14

Quote:
Interesting to read that register T is the Top stack register. It is so the way it is visually depicted in the manuals. In truth, the HP calculator stack is inverted. The true Top stack register is the X register since it receives new values and pops off old ones!!!

Namir


It depends on your point of observation I think :-)
For example the 15C Owner's Handbook, Edition 2011, at page 33 has a nice illustration of the main stack concepts, including "lift" and "drop", and they consider T the top register and X the bottom.

#15

And this is one of the big differences between the classic RPN calcs and the HP 48 RPL stack concept.

In the HP 48 (and the later machines), the topmost stack level is level 1, which is the level most operations work on.

As with the various other stacks in the HP 48, one pushes objects _onto_ (or on top of) the stack, not inserting on the bottom, and lifting the whole stack, as on RPN machines.

The HP 48 stack concept is more suited for large stacks with many levels, whereas the RPN stack concept is suited for limited stacks, with the special case of a sticky T register (or stack level).

Additionally, if transferred into reality, who would want to push things under an exiting stack of things? Normally one would place a new object on top of a stack.

However this also could explain why some RPN afficionados have difficulties with the HP 48 stack concept;-)

Don't get me wrong. In the past, I used many stack tricks, like T replication and L backup and so on, but these tricks only make sense with the limited stack of an HP-41 and before, and are hardly transferrable to other platforms.

#16

Obviously everyone here agrees to the assumption that "T" means something like "Top of stack". My explanation: HP (who produced a 3 level stack machine first) ran out of characters typically denoting coordinates in an euclidian space. They needed a "forth dimension", and in advanced physics "T" denotes "time" in a 4 dimensional space.

#17

Richtig ! Obviously... "T" for Time.

#18

Quote:
Obviously everyone here agrees to the assumption that "T" means something like "Top of stack"

Not everyone. :-)


Gerson W. Barbosa said:

Quote:
Temporary register.

and I said:

Quote:
That is how I remember it. It is "temporary" because some internal calculations -- transcendentals, I believe -- can corrupt it.

I looked at the HP-35 operating manual -- see page 17-- and confirmed my memory that transcendentals use the T register. So, I stick with "T" meaning "temporary".


-- Rich

#19

Is the 35 the only model where transcendental modify the T register or does that happen on other models as well?

#20

It looks like the HP-35 was the only one, judging by the next HP calculator:

Keystrokes            Display

HP-35 HP-45

1 ENTER^ 1 1.00
2 ENTER^ 2 2.00
3 ENTER^ 3 3.00
30 SIN .5 0.50
Rv 3 3.00
Rv 2 2.00
Rv 2 1.00

#21

Quote:
Interesting to read that register T is the Top stack register. It is so the way it is visually depicted in the manuals. In truth, the HP calculator stack is inverted. The true Top stack register is the X register since it receives new values and pops off old ones!!!

Namir


I must disagree. X is on the bottom. This allowed HP to do away with the trouble-prone springs used in other stack implementations and introduce their brilliant gravity-operated stack :-)

Edited: 30 June 2013, 6:20 p.m.

#22

Quote:
In the HP 48 (...), the topmost stack level is level 1, which is the level most operations work on.

As with the various other stacks in the HP 48, one pushes objects _onto_ (or on top of) the stack, not inserting on the bottom, and lifting the whole stack, as on RPN machines.

On display of any RPL model, however, level 1 is at the bottom and the stack builds on top of it. Maybe this inconsistency contributes to the user friendliness of RPL :-? Compare the HP-42S with y above x, matching the RPN stack image.

d:-)

#23

Quote:
On display of any RPL model, however, level 1 is at the bottom and the stack builds on top of it.
No, level 1 is the nearest reachable one, so it's in front of the other levels, but the HP 48 has the ability to show the levels behind the nearest, and they had to visualize it somehow, so they chose to display the topmost stack level near the keyboard for ergonomic reasons.


Quote:
Maybe this inconsistency contributes to the user friendliness of RPL :-? Compare the HP-42S with y above x, matching the RPN stack image.
Maybe you know that the HP-42S is an RPL machine internally, with an RPL data stack and return stack and (nearly) everything else which makes it an RPL machine.

Number operations are performed either in the CPU directly if applicable, else the RPL data stack is used for temporary results, which then will be copied to the simulated XYZT "registers" which reside in system RAM.

BTW the predecessor of the HP-42S is the HP-28S.

The RPN UI is the shell for the underlying RPL kernel, and the FOCAL execution engine is an emulation the HP-41 built-in execution engine, written in RPL and Saturn assembly.


And if we talk about inconsistency, the HP RPN calcs are good examples, because they mix postfix and prefix (e.g. SF 02). RPL machines are way more consistent in this regard, since nearly every internal basic function or command actually works with postfix operands.

On the HP 48, you _can_ use postfix, infix, or prefix notation where applicable if you like.

And yes, the HP 48 user interface is very user friendly. However one has to learn to work with it, as is the case with traditional RPN calcs, too.

Discussions about RPN vs. RPL fill books, however it's much of personal taste which UI to prefer. I'm comfortable with both, and stack operations are easy with both.

#24

From memory, on the HP-45 and -65, the use of transcendentals over-writes R9 rather than the T stack register.

Best,

--- Les

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

#25

I always thought that the register naming was

X,Y,Z = 3 spatial dimensions

T = time dimension

Don't think I read this anywhere. It was my own "obvious" conclusion upon encountering HP RPN and Einstein in high school way back in the late 70s.

Edited: 30 June 2013, 11:46 p.m.

#26

Quote:
Hi all.
I was reading some old issues of the HP Solve newsletter, with historical presentation of RPN and its comparison with the other methods.
I have a little curiosity, is there a special meaning for the letter "T" used for the fourth stack register? I can understand the choice for X, Y and Z, but why T?
Thanks.

I always thought temporary - I could be wrong.

Eddie

#27

I finally found it!



file: Museum DVD/HP_Museum_2/cd16/ads/entvseq.pdf



Ad: "ENTER vs ="





How the operational stack works


#28

I always considered the "T" designation to stem historically from the original HP-35 machine's use and overwrite of that stack register for trigonometric functions. The HP 9100A/B had X, Y, and Z stack registers, so calling the added HP-35 register "T", for its use in trigonometric functions, has always made sense to me ever since I first learned about the HP-35 in 1972.

Later machines did not use stack registers for trig calculations, so then it became necessary to invent some other basis for the "T" designation. In any event, it's certainly not a "temporary" register. Its contents are maintained as the stack drops in subsequent calculations...contents remain unchanged for all later operations until the stack fills with some replacement in T. Its content is thus usually the most permanent of the stack registers.


Edited: 3 July 2013, 12:54 a.m.

#29

Yes, permanent for stack dropping actions, but quite temporary for stack lifting operations. Moreso than any other stack register since once the value is pushed out of T, it is gone forever. Guess they should have wrapped around the alphabet and called it the "A" register to avoid all this discussion 40+ years hence :-)

#30

Quote:
I finally found it!



file: Museum DVD/HP_Museum_2/cd16/ads/entvseq.pdf



Ad: "ENTER vs ="


Hi Sylvain,
nice reference, thanks for posting.

#31

Quote:
Guess they should have wrapped around the alphabet and called it the "A" register to avoid all this discussion 40+ years hence :-)

Or kept going forward and called it the "[" register.
#32

Quote:

Or kept going forward and called it the "[" register.


You did want to propose "Ä", didn't you?

d;-)

#33

I was curious and I could not satisfied myself with

only one description reference for the T register.

So I re-read all the classic series manuals and here is what I found.

HP-35 manual: Top (p.6) and Trigonometric (p.17)

HP-45 manual: Top (p.21)

HP-46 manual: not defined

HP-55 manual: Top (p.10)

HP-65 manual: not defined

HP-70 manual: Top (p.12, p.81)

HP-80 manual: not defined

HP-81 manual: not defined



Sylvain

#34

These only say the T register is on the top of the stack, not that T stands for Top.
Let me repeat Nina's link above:

http://www.hpmuseum.org/hp9100.htm

In case one doesn't mind to follow it, let me quote the curator's words on this matter:

"When people first encounter HP's RPN calculators, they often wonder about the stack labels X, Y, Z, and T. The HP 9100 had a three level stack with registers X, Y, and Z. X was called the keyboard register since numbers were keyed into it. Y was called the accumulator since the results of arithmetic on numbers in X and Y ended up in this register. (The entire stack was displayed.) Finally Z was called the temporary register. When HP later used 4 level stacks, the temporary name stuck and the new level became T."

Best regards,

Gerson.

#35

Yes you are right, it is circumstantial at best but it's all I could get from the manuals.



Now, I was not part of the HP engineering group or close from one of them at the time
and I assume the curator was, so I rest my case.



Temporary it is! 8-)



Best regards,



Sylvain



Possibly Related Threads…
Thread Author Replies Views Last Post
  HP Prime CAS curiosity bluesun08 11 4,707 12-10-2013, 01:03 PM
Last Post: Han
  HP Prime Matrices curiosity bluesun08 0 1,479 12-09-2013, 06:44 PM
Last Post: bluesun08
  HP 50g - select characters on the stack, copy/paste Sean Freeman 7 2,609 11-20-2013, 07:11 AM
Last Post: Sean Freeman
  Prime: Placing more than 1 item on the RPN stack in a single program? John Colvin 4 2,204 11-19-2013, 08:59 AM
Last Post: Miguel Toro
  emu48 - copy stack doesn't work (as expected) Thomas Radtke 2 1,923 11-11-2013, 02:19 PM
Last Post: Thomas Radtke
  HP Prime Stack operations from within a program John Colvin 1 1,340 11-08-2013, 09:45 PM
Last Post: Helge Gabert
  Another Prime curiosity Michael de Estrada 0 944 11-05-2013, 11:52 AM
Last Post: Michael de Estrada
  HP-80 CHS Exponent Curiosity Max Stone 4 1,828 10-22-2013, 02:39 PM
Last Post: Max Stone
  setting up subroutines that are called many times... Geoff Quickfall 4 1,793 10-18-2013, 03:41 AM
Last Post: Geoff Quickfall
  Prime: Anyway to refresh stack? kris223 5 2,069 10-16-2013, 05:09 PM
Last Post: kris223

Forum Jump: