Nothing much



#12

I just learned of this site and wanted to post an acknowledgment of how nice it is to see it. I still use my HP-25, it still works as well as ever. That RPN never caught on is always a mystery to me, but HP calculators have been the best of the best since the beginning -- which I remember first-hand.


#13

I can make RPN-phobic people mad by reminding them that the usual calculator is mixed mode, in that it is inconsistent.

You sometimes press the function before and sometimes after the number.

Sine, for instance, is pressd after the number, but addition must be selected before entering a number.

I point out that RPN calculators are at least consistent. Gets them every time :-)


#14

Hi;

I cannot agree more with you.

I teach HP12C Use & Program classes, and I remind the students that it exist algebraic NOTATION / EXPRESSION and math operations. When we want to take note of some math operations we use algebraic notation in the form of math expressions. If someone understands that notation, he will be able to reproduce the math operations.

RPN deals with math operations, not algebraic notation. After the HP28S we, HP users, made contact with a new way of using algebriac expressions: as objects in a RPN-fashion structure. We could enter as many algebraic as possible into stack levels, use them, combine them, and if we wanted to know what sort of expression would come from a particular RPN sequence, we could just fill the stack with variable names and use RPN operations. The resulting object would be an algebraic equivalent to the RPN sequence. Amazing!

I thnk that sometimes an Algebraic expression allows faster understanding than an RPN sequence (I usualy take as much time as needed to depict RPN sequences in others programs, so I can understand what sort of procedure was used), but RPN sequences are intended to be math consistent, NOT algebraic consistent.

Anything (for sure) else to add here? Comments? Flames?

(Cheers)


#15

In my opinion, RPN is nothing more than an incredibly efficient method to enter an algebraic expression into a calculating device. For online manipulation of data structures such as Real & Complex numbers, matrices, or - in case of RPL - complete expressions displayed in algebraic notation, RPN is perfect. Over 10 years ago, I wrote a small application on the Atari TT which took an RPN input to build an algebraic expression for later use in other programming languages such as C, to provide a simpler and less error-prone way of writing expressions. And today, I'm still missing a similar tool to enter expressions in, let's say, Word or Excel. Take the 49G as an example: in relatively few keystrokes, you can easily build a large expression - and I'm not talking about this equation writer... but now I'm diverging...

Regarding the drawbacks of RPN/RPL: As mentioned in an earlier thread in this forum, RPN/RPL are candidates for 'throw-away programs'. Key it in, use it, and forget about it. Here I'm clearly in favour of a mixed mode calculator which has an 'algebraic', i.e. standard infix notation option for programming. The 32SII is pointing in the right direction: it allows the use of equations, which are more readable than an RPN keystroke sequence! Unfortunately, they're also slower in execution and larger in size, which is an issue when you've got merely 384 bytes available (hah, my 15C beats this!).

So, my 'dream calculator' would have RPN entry, RPN keystroke programming for the quick'n dirty tasks, and an algebraic programming model for the serious programs. All in all, this sounds like a 42SII :-))

Pascal

#16

Welcome, Ken. Maybe you could share a memory or two of the early days in the memory forum. As far as RPN not catching on, I can think of a few reasons off hand:

1. Price competition. Most algebraics were cheaper than HP's.

2. Since TI went with algebraic, when they sold calculator chips they were algebraic as well. The result was that RPN was 'swamped out' by all the algebraics hitting the market. The only other company i'm aware of that made RPN calculators was National Semiconductor but their implementation was not the same as HP's and had more limitations. I don't know if National put their RPN calculator chips on the open market but I don't think they could have competed with TI even if they did.

3. Reluctance of most people to scale the learning curve of RPN. Somewhat disappointing but not surprising when you add in points 1 and 2. When scrolling multi-segment or dot matrix LCDs allowed entire formulas to be input and edited at once that pretty much finished it in the publics mind, I think.


#17

In my opinion, RPN or algebraic is not a chips (microprocessor) issue. Most microprocessors have features like stack support and register load instructions which don't prevent RPN at all. In fact, while it can be a matter of discussion, RPN may result quite natural to implement on them. RPN or algebraic is a matter of software (or firmware, if you prefer). The calculator application of the HP LX 100/200, which runs in a 8086 architecture, is a nice example of dual mode operation.


#18

<In my opinion, RPN or algebraic is not a chips (microprocessor) issue. Most microprocessors have features like stack support and register load instructions which don't prevent RPN at all. In fact, while it can be a matter of discussion, RPN may result quite natural to implement on them. RPN or algebraic is a matter of software (or firmware, if you prefer). The calculator application of the HP LX 100/200, which runs in a 8086 architecture, is a nice example of dual mode operation>

Yea but I was talking about the 'calculator-on-a-chip' devices that TI and others sold in the '70's. They had display driver, keyboard driver, calculator rom - all on one chip. All the manufacturer had to do was supply a keyboard, display and power supply.


#19

The TI57 was exactly like this: one chip and... nothing but a 9V battery clip, keyboard and LED array (display). I had one in 1979... Wow!


#20

I recently started collecting TI calculators and one of the first ones I got was a TI57 with a bad battery pack. The pack contains a convertor from 2.4V to 9V - but I didn't know what the output voltage was. I connected the calculator to a bench supply and slowly increased the voltage, occasionally switching the calculator off and on in case it needed a clean power supply edge to reset. I got up to about 16V with no results except random LED segments sometimes lighting, so I figured the calculator was dead. Later I got another TI with the same battery pack which was working. I discovered to my horror that TI used red for the negative supply wire and black for the positive! Around this same time TI had their 9900 microprocessor, in the documentation they numbered the address and data lines opposite from everybody else (0 was used for the most significant bit). But the story has a happy ending: after the abuse I gave it, the TI57 still works completely! I can only guess that between the protection diodes in the chip and the LEDs connected to so many pins possibly acting as shunts, the IC never saw an excess voltage.

#21

steve;

re: your paragraph #1 & #3 - sad but true. re: your paragraph #2, here is some info from our own dave, joerg woerner, ball & flamm and others.

there were a lot of rpn calcs out there back in the "golden age" of calculators. there were 7 or 8 makes made in the former ussr and most had lots of models, a couple in germany, one in japan, one in taiwan, of course sinclair in "jolly old" made two. ti even made an rpn translator rom for the 58/58c/59 series so people who were saddled with a ti could use 67/97 programs. someone from japan mentioned another translator rom for the pannasonic hhc here at the forum a few months ago. and yes nat.semi./novus did sell their chip for the mathematician at least once; "fate sa" of argentina made a pretty nice box, put it around the n.s. chip, and called it the microcifra cientifica. sadly-the trig accuracy still sucks. add this to about 6 or 7 other brands from the us of a. and you get over 50 models.

we are left with the conclusion that since aos entry requires no thinking, and people chose it even though there was an affordable rpn under every other rock; most people must not want to or be able to think.

- d


#22

<<there were a lot of rpn calcs out there back in the "golden age" of calculators. there were 7 or 8 makes made in the former ussr and most had lots of models,>>

That's nice to know but I don't think the Soviet calculators at that time could have influenced the European/Asian/US market one way or the other.


<a couple in germany, one in japan, one in taiwan, of course sinclair in "jolly old" made two.>

This is good to know, but i'd have to know more about the specific models to tell whether they would have been a good or a bad influence on the acceptance of RPN. The NS models for example were so crappy, comaparatively, that they may have had a negative influence.

<<ti even made an rpn translator rom for the 58/58c/59 series so people who were saddled with a ti could use 67/97 programs. >>

Yes I know about that. Do I detect embers of the HP/TI wars still burning here?....

<<someone from japan mentioned another translator rom for the pannasonic hhc here at the forum a few months ago. and yes nat.semi./novus did sell their chip for the mathematician at least once; "fate sa" of argentina made a pretty nice box, put it around the n.s. chip, and called it the microcifra cientifica. sadly-the trig accuracy still sucks.>>

Neat. That's my point; did these cheapo models help or hurt?


<<add this to about 6 or 7 other brands from the us of a. and you get over 50 models.>>

Whoa. 6 or 7 other brands? We are talking about scientifics here, right? What are the names of these other brands? I know there were some 4 function units that used essentially RPN but I don't think we can count those since the market for 4 function calculators is too broad. I'm focusing on scientifics because there people had the opportunity and incentive to see the advantages of RPN in a significant way.


Forum Jump: