NEW HP 12C Platinum !!!



#2

Guys,

HP just released the “new” HP 12C Platinum, it is RPN + Algebraic switchable and with 400 memory steps (vs. 12C 99 steps)! It is available at S$185.00 (approximately US$100) please go to http://www.educalc.net/default.asp?tid=147432 . I have spoken to the Sales Manager, according to her we should be expecting the “new” HP 32SII soon too :-)


#3

Searching for "12c platinum" at www.hp.com reveals the manual for it!

Frank.


#4

Thanks a lot for the info, but downloading and reading the
manual it's kind of disappointing, because:

  • It seems the only changes made are that there's now up to 400 program steps, instead on up to 99, and an algebraic mode, but

  • Although it seems you can have up to 63 or 64 registers
    for data storage, there's no way to STO or RCL more than 20, from 0 to 9 and .0 to .9. HP hasn't provided any indirect capability (i) at all, so you can't use registers above those 20 neither manually nor in a program. They can be used just to store cashflows (up to 30 different cf)

  • Further, you can do storage arithmetic only on registers 0 to 4, no extensions to allow it at least for registers 0 to 9.

  • Although the manual describes its programming capabilities as "powerful", no improvements here at all. No subroutines, labels, flags, nothing. Still only two logical tests. No inserting program lines, still abysmal editing facilities.

  • No new functions at all, in particular no scientific functions like trigonometrics.

In short, a real platinum opportunity missed once again. Had HP made some minimal improvements, such as adding scientific functions, subroutines, labels, and some way to use the extra registers, this could have been a dream machine for nearly everyone, business and scientific users alike. Wouldn't you buy a Voyager-series, classical RPN calculator with 400 program steps, basical scientific functions, and decent programming ? Yes, but sadly this isn't it. Many of us will buy it regardless, but it's a real pity when one considers what it could have been.


#5

I suspect that the "target market", financial and management types, wouldn't use trig or other "scientific" functions very much, and would do precious little programming on their calculator.

Maybe HP's idea is to "keep it simple", and they may well be right in this case.

Having a precious metal in the product name should be especially appealing to that market segment.

Regards,
James


#6

Even TI in their TI BAII Plus puts all trig functions AND hyperbolics in a $30 calculator.

Wish HP would wake up that people want access to financial AND scientific functions. If not, TI will continue to clean their plate in the financial calculator market. For $30, three BAII Plus calculators can be bought for one 12c Platinum.

#7

. . . but the simple increase in memory is suggestive. If they produce the 32SII in a decent package (better than the 30S, but o.k., maybe the original Pioneer quality is too much to hope for) with no more changes than 4x memory (i.e., ~1K) I will be pleasantly suprised.

No, it wouldn't be the extensible, I/O enabled 41 replacement that many people dream of, but having a good shirt-pocket RPN offering again routinely available would be a big step forward.

#8

From the manual:

"The status indicators turned on at the end of this test include some that normally are not displayed on the HP 12C Platinum."

Does anyone know what these are? Are they an indication od what a future calculator might do, or are they a left over from the last calculator to use this display?


#9

Well, the original 12c has annunciators that aren't used (USER for instance).

Might be the same thing.

Gene


#10

But the 12c platinum has the "alg" and "rpn" anounciators that has never before existed on a segmented hp lcd display, has it? This seem to indicate that the display is new...


#11

You are right, but that may or may not mean that the USER annunciator will be used for another model.

After all, the 12c has been made for over 15 years after the 11c/15c went the way of the wind and it's continued to have the USER annunciator, so it's hard to tell. Maybe there'll be another calc, maybe not.

The ALG feature is fairly useless. There are NO parentheses.

You can't solve 1 - ( 2 / 3 ) = without massive rearranging in ALG mode.

That's nuts.


#12

It may not be the most efficient in keystrokes but 1-(2/3)

1
[Enter]
2
[Divide]
3
[Enter]
[R\/] (roll down)
[-]
[x><y]
[Enter]

You have to keep in mind the 12C platinum is a algebraic with a stack


#13

Manufacturers need to either:

1) Implement basic hierarchy where multiplication and division are done before addition and subtraction (really, this ought to be required before HP sells any calculator),

or

2) Add parentheses.

If option 2 is taken, it becomes a real chore to solve many basic problems and in many cases requires people using ALG mode to begin in the middle of the problem and work outward. Hmm...that's what we all do when we use RPN.

Algebraic's argument is that you key a problem the way it is written.

Not true without #1 and/or #2 above.

My preference would be, if you're going to put ALG on a calculator, to do both!

Gene

#14

The display may be new, but they may have made the decision to keep the old annunciators, even the unused ones. If they had turfed them, they would have also had to find all places in the code where they were used (such as in the self-test code). Eradicating that code might have required a more extensive test. Adding new functions doesn't have anywhere the same risk, ironically, as deleting old ones.


#15

Hi Patrick, guys;

I read part of the Platinum PDF and it's indeed mentioned that after the selftest, some annunciators may not be available in normal operation. Should we take it as an unchanged excerpt from the original HP12C's manual and there are no such USER, G, RAD annunciators or they have actually been kept and squeezed so that the new low-battery, ALG and RPN fit in the same area? I had this doubt and was not sure it was actually to be mentioned, but given the circumstances I would like to ask Gene: if you have the Platinum in hands, should you check for this particular fact, please? Also, were digits kept as for their original shape, size, format? (or will my TTF files be still useful?)

Thanks.

Luiz C. Vieira - Brazil


#16

After a self test, annunciators are:

RPN ALG USER f g BEGIN D.MY C PRGM

No GRAD.

Digits seem to be "slimmer" than the original.

My basic opinion:

If this were the same price as the original 12c, I'd still buy the original.

Reasons:

1) 400 steps may be nice, but I already don't program with the limited programming capabilities in the 12c...more steps doesn't help me much there.

2) Solving for financial functions is slower as is the entire responsiveness of the machine. I'd become frustrated very quickly.

3) ALG mode is useless without parentheses or hierarchy.

4) Color scheme is outlandish to have in office meetings...people will think it is a toy while the old 12c looks "respectable".

#17

Hi;

I tried to find the manual but I was not successfull. Can anyone post the final link so I can download it? Or even send me a copy to my e-mail address.

Thanks a lot.

Luiz C. Vieira - Brazil


#18

Try this:

HP-12C Platinum Manual

Best regards


#19

Hi, V.A.;

somewhat hard to find this little fella...

Best regards.

Luiz C. Vieira - Brazil

#20

For a calculator that does not respond to keystrokes:
1. Insert a thin, pointed object all the way into the reset hole near the battery
compartment and then remove it.
2. ...

If it fails - heavy metal f*ck

<G>.....Mike

#21

Found this Domain while looking through hp-calculators.com website:

http://platinm1.tempdomainname.com  which appears to be equivalent to: http://hp-calculators.com

And these links takes you to a couple HP photos of the 12C Platinum:

http://platinm1.tempdomainname.com/images/products/12cplat-large-2.jpg
http://platinm1.tempdomainname.com/images/products/12cplat-large-1.jpg

which also works through the hp-calculators.com domain

(ie.  http://hp-calculators.com/images/products/12cplat-large-2.jpg ).


I can't however, find any photos of the other new calculators Steven Soto mentioned (so far, that is).

Matt


#22

I note that it still has a 'CLx' button, rather than a '<=' button. Does that mean it still has the 12C's out-of-date feature of requiring you to erase all your input and start again when you make a keying error?

(I also note they've printed 'financial calculator' right on the front of the case. Just in case the owner forgets what it's for, I guess :-) .)


#23

Type 123456789 instead of 123456788 and you must retype the whole number.

I said it before...if the new 12c was the same price as the old, I'd still buy the old, more memory or not.

Since the new one will be $30 more...

Gene

#24

It looks quite American...


Neil Armstrong would have used this on the moon.

Miss Liberty has this tucked under her left arm.

A calculator more suited to pork bellies than Euros, from the look of it...

;-)

(To complete the American decor, the back just HAS to have the following label added:

"Calculator batteries must not be eaten.
Do not operate vehicles or heavy machinery while calculating.
Not intended as a substitute for a trained mathematician.

...").

The more legal boilerplate, the better!

#25

and

HP-15C Palladium
HP-32SII Iridium
HP-17BII PureGold

Other possibility:
HP-42S 'Heart of Gold'

Ok, the last one was a Hitchhiker's joke...

Raymond


#26

Right now they only have Unobtainium models. ;-)

#27

The HP49G should be the 'gold' version if you are refering to the HHGTTG (Hitch Hiker's Guide to the Galaxy). Why...

Well if you wrap it in a slice of lemon the effect is very similar to the result of a Pan-galactic Gargle Blaster. i.e. a big headache after reading the user guide and trying to get the thing to work, also the lemon reference is quite good for this product!

PS - does anyone know what the hardware is like and if it is suitable for a scientific model (if HP ever get their head out of the sand).

#28

If it is online, I'll share my thoughts later today.


#29

the old 12c was short only a few lines of memory from being able to have a good trig program. there is a copy floating around somewhere on the museum. i think it was done by series expansion. the program works; it just needed more room to do decent output. now theres room. and a scientific comming out soon.

some people would complain if they got hanged with a new rope.


#30

Hi;

a great contribution from Viktor Toth, that can be found here, at the Articles Forum. If the Platinumn Version has plenty of program steps and its structure is the same (GTO nnn?) then existing programs need only a GTO nnn review and enhancement. I think all Financial professionals will be proud showing the new "breed". It's gonna be a salesperson dream... as a tool and as a promotional sales record.

Luiz C. Vieira - Brazil


#31

It too looks like a toy. The original 12c will continue to sell rather well.

I don't think the financial people will like all that well. My 2 cents (as a financial person)!


#32

I agree. It's not as ugly as some of the other recent models (notably the 49G) but it's certainly a big step down from the original Voyager series.

I wonder if HP will ever again release a new calc I wouldn't be ashamed to own?

#33

"the old 12c was short only a few lines of memory from being able to have a good trig program"

Have you seen this ?


HP-12C Trigonometrics


#34

I mean the VERY first line. :-)


#35

I don't get it ... what's so curious about the first line !?

AFAIK, this is an article published in Datafile some time ago. Being very fond of chess, I grabbed the whole web site to try the positions on my chess software, and this PDF file was there, which I remembered having seen before, perhaps included in a CD-R compilation or something. I then tried the program and found it delivers, too bad there seems to be no more PDFs there ...

BTW, Gene, do you have an HP-12C Platinum at hand ? Can you verify if this 12C trigs program runs fine on the new machine, and if yes, whether it does run faster ? If so, I might give the 12C Platinum a try as a daily-use, always-in-the-pocket RPN calculator.


#36

I have a platinum at home and will try it out.

#37

Anyone put this somewhere else?


#38

i couldn't get it to copy from the site but i have a version of a shorter one by victor toth. if you want that, write to my email and i'll send you a copy.

#39

Thanks for your interest. I'm just setting up a calc-related web site which will feature a number of my articles and programs and the one you mention was temporarily placed in another chess-related web site of mine, while making some tests.

This and all the other articles and materials will be generally available in a near future, but meanwhile, if you're still interested, please provide a valid e-mail address and I'll send a copy of it to you.


#40

I would like to have the trig functions page.
A local STaK (Suomen Taskutietokone Kerho=Finnish Pocketcomputer Club) meeting will be held in Helsinki 17th May Saturnday 13:00, where I will introduce the Platinum and the old style 12C (3 batteries) for comparison.
I would like to load the trig fn programs into the new largeer program memory before the meeting. BTW: A simple adding loop runs about 5,83 times faster than before.


#41

VP is correct. A simple loop doing nothing but + instructions runs about 6X faster on the new 12c platinum.

But, solving for everyday functions (Y^X, e^x, and the financial solution to the interest rate i) all take longer.

Have no idea why they would let this be the case.

Gene


#42

It may sound odd, but what I'd bet is going on is that you have HP working to replace the most successful calculator they ever had, facing the phase-out of a part central to its function-- the Saturn processor.

We heard rumors, back just before they were folding the ACO division, that the Saturn was long-of-tooth and would be replaced in short order.

We know that, now that HP *has* no calculator manufacturing facility of its own, it is relying on an outside manufacturer to realize its designs, INCLUDING the current 12c.

The outside manufacturer has made calcs of its own, and probably has its own choice of machine-architecture.

Now there is current stock of 12c's and their parts-- enough to sell them probably for a few years yet. But if you know the well is gonna run dry, you start now: you have your contractor make the old design, but concurrently, you bring out its REPLACEMENT design.

And I'd bet that there is a reason it's "no longer fun" to be producing Saturn chips. Maybe the Saturn chip is supplied by Agilent, and they no longer want to extend a contract for a part whose volume and margin do not support continued production.

Or, assuming the ability to MAKE them stayed with HP and did not walk off with Agilent, it probably represents the only example of that PARTICULAR technology HP still makes; a production line with one customer, one purpose and no future past that.

HP, under those circumstances, would see that the Saturn-producing facilities were underutilized, while facilities for making inkjet cartridge nozzles (to pick an absurdity) needed expansion and so forth.

There comes a point where a particular part should be either outsourced or replaced-- and so you say to your outside contractor, "Can you make a version of this calculator using YOUR CPU parts instead of ours?"

Voila! A 12c-like "12c Platinum" that is not the SAME at its most fundamental-- it has had a BRAIN TRANSPLANT.

The "operating system" of one machine was ported, as best as could be, from one CPU to the other. But, Saturn was kind of a special beast-- and another calculator or CPU chip would have a different mix of instruction-set, of registers, of I/O setup and timings.

It is probable that the new chip being used on the 12c Platinum has a lot to recommmend it-- it is probably clocked faster and has wonderful features not implemented back when Saturn was current. But, as will be seen in the performance of the "12c Platinum", there are probably things the Saturn was set up to do in very few instructions, which the new CPU can only duplicate by doing many more instructions.

So, what Gene is mentioning about certain operations is possibly a result of an inherently faster processor having to run around doing a lot of stuff in the background that the Saturn was designed to do Natively.

If this is correct, the ONLY impact it will have on the user is the calculation times will be different. That is a shame, if it is slower on commands that are used a lot. But it may be the only thing that COULD be done-- as 12c Saturn version is phased out gradually, in favor of 12cP.

Once they have proven the viability of 12cP, the only thing to keep them from a 15cP or a 16cP is their assessment of market prospects. But they may already have filled the "slots" of their market strategy with different form-factors: hence whatever Steven Soto saw as a 33s, etc.

Anyway, that is my guess. Anyone out there able to substantiate or disprove any of my speculation?

:-)


#43

Hi Glynn, guys;

For sure, another processor and algorithm.

If we take the fact there are other O&M brands being held by HP design team, and based in the fact we had a large search for RPN related material (remember?), it's not hard to accept that, at least as a pilot project, this is probably the first RPN-derived HP product form the other already known brands (I do not have them in mind: anyone, please? I know the first letter is "K"...). This way, existing algorithms like the base four operations and some other features are kept and, in some case, reduced to accept the new RPN demands (smaller than any AOS-related). Otherwise, the not-existing features, like all HP12C financial resources and how they affect and are affected by the automatic stack, might be programmed, and they had to take place as SW structres, not HW existing features. Let's take as an example the HP42S and the HP41. In the HP41 there were sixteen basic system registers with fixed, pre-defined arddresses. This does not happen with the HP42S, mostly because the "stack registers" will demand memory enough to contain a zero, a non-zero number, a complex number, a sis-character ALPHA string or a matrix descriptor, and in this case, each "object" demands different memory space. It's a waste of memory to keep the maximum space needed to hold the largest object times five (let's not forget the LASTx register). So, the stack registers will occupy space as its needed. This is accomplished in the HP42S by software controll (and it is a Saturn-based machine), while in the HP41 its a matter of existing memory space.

I wonder what processor did HP use in the HP12C Platinum, and I guess it's one of the existing new processors used in the new algebraic machines. Hey, gene, don't you wanna have a look inside yours so we can have a confirmation? AHN????

My US$ 0.02.

Luiz C. Vieira - Brazil


#44

<<note: I don't really know more than a wisp of what I'm talking about, so these are only wild darts in the air...>>

I wonder, Luiz;

Say we had a cpu with 56-bits, as is discussed in the link on cpu technology for the Voyagers.

XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX XXXX <- bits

and so had 10-BCD number plus 2 Exponents and also signs on both number and exponent.

Now, say we have developed routines that use these registers and memory of this size, and we can do amazing maths with them, achieving a certain precision of result.

But, imagine that our boss says: "you must use a different part now-- this one!" and hands us a RISC cpu with five or six times the registers, and the ability to act directly on data in situ in each register, not just on the one Accumulator you had before; but they are all only 32-bits wide.

Ack!!!

Your task, Mr. Vieira, is to modify your routines to make the same results, the same precision as before-- on the new architecture. Certainly you can do it...

BUT, you can see that now your data representation is not a native fit to one register. It must be split, and two registers (or, rather, sets of dual registers) must keep the numbers.

Your new processor is faster, clocking maybe ten times as many cycles each second. But the splitting of your data-representation into pieces COSTS YOU many of those cycles-- and VERY costly in terms of time is an operation in which you must bring data in to a register PAIR from memory or copy it out from a register PAIR to some other place.

But these operations are central to certain maths. Solving For something tends to require a lot of swapping data back and forth, I would figure. And because the data-representation makes us handle two registers as if they were one big one, the benefits of our nifty new 32-bit RISC-based chip kinda get chewed away, most noticeable on certain complex operations.

I bring this up because many of the cpu "cores" used in designs of integrated circuits these days were designed with RISC instruction sets (good for our cause) and 32-bit data-paths (uh-oh) because large data-paths are complex and make it harder to produce reliable VLSI designs. And, besides that, nobody making the core design ever thought: "Gee, we need to support really long data representations, because somebody will do BCD math and need a 10-digit mantissa and exponents and signs to go with it..."

Most cpu designs had floating-point manipulations and C-language libraries in mind; so asking it to replace fully, functionally and FASTER, an odd little chip designed specifically for the BCD math task is like outfitting a race-horse with two barrels full of water and biggger shoes and bringing it to be your desert-camel. Enough training and it can work-- but the camel's whole plodding nature was in it, while the best we can expect the horse to do is awkwardly adapt and keep up as best it can.

I wonder if maybe the "core" of our 12cP's main chip is an ARM7 or a MIPS 32-bit one, or something similar. It would easily, EASILY explain a performance degradation on certain math ops despite 20+ years of cpu speed and performance "progress".

--- just more speculatin'--- glynn.


#45

Hi, Glynn. thanks for the nice words!

I posted this message a few hours ago, but I revised it for errors and it's been posted again, O.K.?

Hey, Glynn; Let's first talk about a few important things: internal core, data bus and SW precision.

Remember the ancient 8086 and its partner, the 8088? The 8086 was introduced before the 8088, and the 8086 is a 16-bit data bus processor, while the 8088 is an 8-bit data processor. After introducing the 8086, Intel realized that all commercially existing 8-bit architecture should be changed to accept the new 16-bit data bus processor, and industry would take a bit long to incorporate it in product lines. We know industry takes some time (nowadays it's shorter) to incorporate new technology and make it a consuming line. So, Intel introduced the 8088 later, and it was a lot better sold. The 80286 took a better advantage.

In both cases, either the 8086 or the 8088 were able to run, say, C- and PASCAL-related compilers, and both could handle a lot more than 8 or 16 bits processing. And dealing with BCD data and straight binary representation is just a matter of handling data. In both cases, BCD and straight binary, either the 8088 or the 8086 already had their "hidden aces": both offer (and almost all general purpose processor, too) carry-bit for BCD and straight binary math operations. The carry bit, also found in the HP16C (the actual Das Kleine Wunder), allows "expanding borders" when dealing with "long-integer" or "double-precision" bigger guys. Just the carry bit. And you may know that the HP16C handles 64-bit # (not data) but also goes to 128-bit # when performing [DBL÷], [DBL×] and [DBLR] (DouBLe Remainder after division]. How come? It uses Y an Z registers to hold a 128-bit input data and X and Y registers to hold a 128-bit result, if Word-Size is set to 64, the maximum available.

But we know voyagers deal with 56-bits wide data numbers. Why is the HP16C able to handle 128-bit data?

Software. And when you bond internal design to software, or even better, you "design" internal structure having a software goal in mind, you custom chips will show exceptional performance when running the software they were molded to run. And that's what we've been seeing in almost all HP (I mean "H" "P", actually) calculators: exceptional performance with custom "chipset". Dealing with RPN custom chips means designing chips with stack registers availability or stack manipulation "made easy", as we found in both 8088 and 8086 when talking about BCD coded numbers and carry-bit handling. Imagine Voyagers chips (each R2D2 for each model) already have stack manipulation "made easy", register arithmetic "made-easy" (except for the HP16C, that does not offer register arithmetic and I wonder that would be hard to handle variable size registers with arithmetic abilities...). Even in specific cases, like HP15C matrices operations, optimizing them based in internal architecture would be a matter of knowing internal architecture enough to make the best of it. And you may be sure internal Voyagers design had in mind not a general purpose chipset, instead a number-dealing, KBD-LCD I/O restricted processing unit. And I imagine it is a single eight-bit unit, while a heavily optimized four-bit would also do the job alright.

Problems? Commercial problems! HP would never do what the company that sells the microprocessors used in the HP9G (what the name of it, d.. it? K...) is doing: offer their chips to be used elsewhere. They are too much restricted, or too well focussed, that using a Voyager processor to hold an algebraic operating system would mean a complete redesign, perhaps a new project... just because they are not general-purpose, VLSI, 32-bit processors.

What called my attention is that they have probably ported existing HP12C software and created sort of morph-coding layer (does anybody know where is Linus Torvalds?) to convert HP12C's code to internal ARM7 or similar. That would, indeed, degrade performance if the conversion needs to emulate stack state for each single operation. That would mean a lot faster single operations and exponentially slowing-down loops and the like. If you use specific internal resources in programs, like solving cash-flow problems in a loop-controlled situation, your program will spend a lot more time in the Platinum than in a regular 12C. And simply increasing clock will reduce battery life, so HW and SW balance must be a lot well managed.

There are other circumstances, but I believe these are the ones that directly affect overall performance. I'd like to invite others to come to discussion, as you wish and want, so we may largely consider new technology. My last in-deep research goes a bit beyond RISC internals, but I briefly read about new technology a couple of years ago. As you may notice, Glynn's knowledge is fairly updated if compared to mine.

This is my not-enough US$ 0.01 contribution.

Luiz C. Vieira - Brazil


#46

Hi Luiz, glynn,

Very good posts. I agree with most of what you're saying, subject to a few comments.

First, I don't see mapping 56-bit words onto two 32-bit words as a big deal. These processors have to handle 64-bit integers and 64-bit floating-point on a regular basis. (More about f.p. below.)

Calculators will not be using StrongARMs or other 32-bit CPUs anyway. Reasons: 1. They cost too much. 2. They use too much power. They are really overkill for simple calculators, anyway. (My definition of simple: anything less than a 41 or 48.) For small calculators the overriding concerns are do it cheap and do it with less power. This means using the simplest CPU possible, one on which the CPU core, ROM, RAM, and I/O will all fit onto the same chip, then run it at as low a clock rate as possible. Getting it all on one chip kind of rules out 32-bit CPUs.

If I had to design a 12CP I'd use an 8-bit single-chip micro with RAM and ROM on board. There are many product families out there that would be suitable. These chips are all CMOS and are optimized for low power operation (standby or sleep modes, for instance). And they have enough horsepower to get the job done.

There are a couple of possible ways to write the software. The easiest way would be to write a simulator for the HP CPU and run it on the target machine. You could then in effect copy the entire 12C ROM, add the (few) enhancements and run it in simulation mode. (You would have to do something special about I/O.) The end result would be a little on the slow side, and memory utilization would not be so great, but it would get the job done quickly and cheaply. I have to wonder if this is what happened here. It would certainly explain the slow performance for some tasks.

A better approach would be to start from scratch, write optimized code for BCD arithmetic, digit access, etc., and rewrite the algorithms to really take advantage of what the target processor has to offer. The nice thing is that you no longer have to worry about the restrictions of the HP architecture, such as the limited number of registers.

In the future maybe scientific calculators may move away from BCD arithmetic and adopt binary floating point. Binary arithmetic is a better match for computer hardware anyway; it would be faster too. I heard somewhere that the 30S already does this. The main thing going for BCD is the comfort factor; it's nice to know that 100*0.01 = 1 exactly. Maybe we're not ready to give that up yet.

Comments, criticism, etc. (usually) welcome.

- Michael


#47

For the BCD arithmthics you can also use specialized 4 bits processors, for example Atmel's MARC4...


#48

Hi Chris,

The MARC4 looks interesting, especially the FORTH programming aspect.

It seems fitting somehow to implement an RPN calculator with a stack-oriented language on stack-oriented hardware.

- Michael

#49

I've read that for business applications, BCD math is preferred.


#50

Actually, it isn't just preferred, it's mandatory. Imagine that all quantities are in dollars and cents, and then try this line in (for instance) Visual Basic 6.0:

Print 0.10 + 0.10 + 0.10 + 0.10 + 0.10 + 0.10 + 0.10 + 0.10 + 0.10 + 0.10 - 1.00


Upon running it, you don't get zero dollars, as you should, but:

-1.11022302462516E-16


If you don't get zero after cashing in ten cents as few as ten times, and then forking out one dollar, just think what this kind of behavior can do to long series of additions and subtractions, normally intermixed, as is usual in accounting. You would find it extremely difficult to reconcile you cash activity.

#51

Saturn-compatible CPU's were used in the Pioneer and Clamshell series, and 38G/39G/40G/48xx/49G,
and of course in the famous HP-71B,
but not in the Voyagers.

The Voyager series used a variation of the nut CPU,
which was used in the HP-41 series.
The redesigned 12C (NOT only the Platinum) uses a replacement of another kind.
AFAIK it's a hardware nut CPU emulation.

Seems the 12CP could use some code optimizations;-)

Raymond

#52

The 12C doesn't use the Saturn either. From what I've read, it uses the last of the 56-bit serial CPUs descended from the original HP-35 chip set. This was the same design used in all the Voyagers as well as the 41. By that time, the existing architecture was seen as unsuitable for more advanced calculators and Saturn was designed for the next generation of machines.

But I suspect your main premise is correct -- the 12CP CPU probably isn't an HP calc CPU. It doesn't really matter since there are no user-accessible hooks to machine-level programming or resources.

Saturn's future will really depend on what HP does with the 48/49. Migrating to another CPU will leave behind a lot of the low-level stuff beloved by 48 fans.

- Michael


#53

LOL!!! That's ONE way of insuring my bet is correct-- it can't be one because it never *was*!!

Of course Raymond and Michael are right, as I would have known by reading about the 12c cpu technology on this site.

I somehow thought the Voyager's cpus were 64-bit as well. How disappointing. I feel 8-bits poorer knowing that some of your cpus are bigger than my cpu.

I'll probably need years of psychotherapy for this.

;-)

#54

If a boss told me to implement 12c algorithms for a new hardware...

First, I would praise the Lord for being so lucky.

Second, as someone already said, my first option would be to implement a simulator for 12c cpu and hardware. I would do that because of the 12c financial algorithms - they were first implemented in the classic series, improved in woodstocks, and further developed in the original 12c. This was not a single person task. Being a prudent engineer, I would love to develop all 12c OS myself, but I would humbly accept that three generations of firmware developers did a great job, and use their work to the fullest extent possible.

Checking if the simulator approach was used on the 12cP requires a "forensics algorithm" equivalent for the financial functions. If results are exactly the same in 12cP and 12c, then I would guess a simulator is used.

Further considerations:

1. I doubt any original 12c firmware developers are available for doing the work on the 12cP

2. My *bet* is that 12c original source codes are not available anymore.

3. I guess HP used a third party engineer team for the 12cP. If this is true, the the 12cP probably has a lot of "features" unexpected in a HP RPN calculator. Some of these "features" (humpf...) seem to be detected by the beta testers.

My conclusion is:

Yes, there will be new HP RPN calculators.

No, they will not be the products we expect in terms of overall quality, features, documentation, etc.

What do the beta testers say: does the 12CP stand up for the famous HP RPN calculator quality ?


#55

Hi;

after reading Renato's post, it came to my mind (sometimes it happens...) one question.

In all Voyagers there is a software reset (it also "rotates" the X-register contents) that must be used when calculator does not respond to keystrokes. It is triggered with the sequence [ON]&[first-row,fourth-key]. In the HP10C, 11C and 15C it's [ON]&[yx]. In the HP16C it's [ON]&[D]; in the HP12C it's [ON]&[PMT];

Gene, does this "reset" still exists? I did not find it in the PDF I have here. It should be necessary to "pnch the keys" to see what happens. As we may notice, reset is now available as a metal contact in a hole at the left side of the battery, and this is new to the HP21C Platinum.

May we take some information form this test...

Luiz C. Vieira - Brazil


#56

Sorry, didn't notice the question.

I was able to do it on my original 12c, but I can't get any combination / quick key pressing, etc. to do this on the 12c platinum.

#57

Luiz,

Do you find that the 12C truly rotates the X register value, I don't see that on mine. Starting with 1.0, for example, I get 0.0 after a few resets. However, on the 15C and 16C I get back to the original value after 28 resets -- a true rotation. An early 11c that I have doesn’t do this reset at all but a later one does rotate after 28 resets. The 10C's that I have work more the like the 12C.

This "feature" seems to be very version specific. Has anyone else played around with this? The intermediate results (before it cycles around) that you get on the 15C can be very weird!

-Katie


#58

Hello, Katie;

It's a good point of yours. I read about the "rotation" in an article about the HP15C's extra resources. I have it printed (somewhere) and the e-address is also saved (somewhere else), but I'll try to find it.

I have only new HP12C (Chinese breeds) I did not actually see a display rotation. I remember testing in a few Voyagers at the time I did not have any HP12C. Anyway, I got my chinese CN1180xxxx, ordered a [FIX]9 ([f][9]) and tried the following:

123456789  [ON][PMT] 
First time the display goes blank; after presing [ON][PMT] again I got:
-4567980  12
BEGUIN
If I try [ON][PMT] again I get:
9.99999999 99
It seems numbers are normalized to MAX_VALUE when something different than expected "runs" throught unexcpected, I guess. I wish I had others to try.

Thank you for the warning, Katie.

Luiz C. Vieira - Brazil

#59

I agree with and enjoy a lot of what you've said, Renato, but for two things I'd argue:

1. If HP is not prudent enough to archive expensive work like the development of software for a product like the 12c, then they are quite unusual. Even Apple, famous for disavowing that they ever MADE some of the products that carried them to fame, are aware of the whereabouts of the source code of every operating system and every machine Boot-rom they ever produced. They may never release any of it to hobbyists, but they know which file-cabinet to go to.

HP, I am quite sure, insisted on standards in reporting and archiving and, whether it came from a lone programmer, or a team, it was likely subjected to close scrutiny by mathematicians and logicians at the time, just so an algorithmic mistake did not sully their reputation.

Of course, code has to be commented to be useful for later researchers, and the comments are whatever the programmer feels is most important to be mentioned as he/she writes the code. Hence, later eyes, later programmers (I agree with you that the then-programmers of the 12c are likely retired now) often have a next-to-impossible task of decoding somebody else's code. So, even if HP has every printout they ever did of the source, that does not necessarily mean that it would be easy to port it from one product to another without a bit of thrashing about.

But I am pretty sure that, if I were a VP of product development at HP, and I said, I want the original commented source on our venerable still-in-production 12c, it would be on my desk THAT DAY.

If it weren't, and someone told me it had gone to the bin, there would be some serious, nay, VIOLENT repurcussions that would make even Klingon underlings cower in fear and trembling.

Such an unthinking breach of stewardship as tossing out developed intellectual property on a still-produced product, is simply inconceivable and intolerable in a modern corporation.

2. "Forensics", as you say, can tell us a lot about the specific internals of software written for a given machine. You can tell, for instance, if certain algorithms were used, and if rounding or certain precision was taken into account at given stages. But you've implied that forensically, we would be able to tell if a "12c simulation" was written to run under 12cP hardware, by examining to see if the two machines returned the same results.

I would say that the objective of ANY programming effort, whether a new implementation or a simulation, emulation or a code-translation layer, would be to achieve the same or BETTER results than the original. It would simply not do, to have 12c and 12cP DIFFER in what they achieve mathematically or operationally, unless it corrects a perceived deficiency or "bug". So, I am NOT sure that exact matching of result tells us ANYTHING substantial about the nature of the code within the 12cP.

You COULD say, and be absolutely sure of this: if HP's new financial calculator calculates a penny different on a Billion-dollar account than the standard 12c, whether or not it makes sense to the programmer, HP is going to have to do a lot of explaining; both old and new users are going to pitch a fit. Accountants don't like to think their math accrues error along with interest. And both old and new machines are still sold by HP. "Oh, yes, our 12cP finally gets it right, for $20 more, but we still sell the old one for the financial tyros". LOL!! It *would* be a bit embarrassing, don't you think?


#60

Glynn,

Great post.

1. About source code availability. I agree - chances are that you are right, and that I got carried away in a line of thought. Anyway we will never know for sure. I forgot to mention that this line of thought started by looking at the "improvements" of the 12cP over the 12c - my conclusion is "someone could not or wanted not mess with the original code".
Most "improvements" look like patches, instead of real source code modifications.

2. About forensics. I agree that two different algorithms *can* produce exactly the same results. But in the real world, when dealing with complex algorithms, if two programs deploy exactly the same answer, chances are that the two program are exactly the same.

3. Good point about precision of the 12cP x 12c. Messing with the precision of such a sucessful product is not advisable.


Renato


#61

I have to agree with glynn too; HP is too organized a company to lose source code to a current product. (Though I suppose it's possible.) Also, there are the engineer's original notebooks, sure to be full of useful background info, memos etc. (Notebooks are thye property of the employer, so they stay even when the engineer goes.)

I assume that it wasn't a question of the code being patched, just a desire not to mess around with something that was already good enough; hence, just incremental improvements.

As for accuracy, why not compare results to a calculator with more digits? (And does the manual specify accuracy limits, like the HP-35 or 45 manual did?)

- Michael

#62

How about this...I'll pull out my PPC ROM manual and run through some of the test cases for accuracy to see how the two 12c's compare.

So far, they have agreed in the last decimal point with everything I've tried.

I'll see what I can do.

#63

I've sent the HP-12C trigs article in PDF format to your e-mail address (less the DROP_, of course). Best wishes for your meeting.


#64

Hi, Valentin;

I'd send you an e-mail to express my happiness when reading your name in these posts, but I could not find your e-mail address. Forgive-me if I should not do it publicly, but I could not help myself avoiding to mention it somehow.

By my own, let me say: welcome.

Luiz C. Vieira - Brazil

To remove this post, the passord is 12345.


#65

Hi, Luiz. Thanks for your very kind words, the reason I do not include an e-mail address is I've suffered very bad spam experiences in the past which I'm not eager to risk repeating by making public my e-mail address again.

Including 'removable' bits in a posted address only serves to stop spambots at best, but it's totally useless to stop 'spamhumans', if you know what I mean.

However, if you can provide an e-mail address of yours, where I can contact you directly, I'll send you my own.

Best regards.


#66

Hello, Valentin;

When you mention 'spamhumans' I mostly read 'spam with (bad) actitude'. The worst case of invasion is the conscious invasion.

Please, be my guest: lcvieira@quantica.com.br

As you mention you are one of the Voyagers' fans, we have some (maybe a lot) to talk about. Except for the HP10C, I have at least one of each, even some with different "guts".

What I wanted most to express has already been expressed previously in this thread, but if you don't mind sending me your personal e-mail address, I'd gladly save it amongst others. And please, feel free e-mailling me anytime.

Best regards.

Luiz C. Vieira - Brazil

#67

In addition to the 12C Platinum, can you tell us (or point us to) more info about the other NEW caluculators Steven Soto mentioned in his post below (see HP-33S thread)?

Thanx,
Matt

#68

There is an article in the latest edition of Business Week (Wildstrom)and also an article in May edition of Strategic Finance on the 12C Platinum. Both articles are tributes to the 12C staying power over the years. Good articles.


#69

Here is the link to the Strategic Finance article.

http://www.strategicfinancemag.com/2003/5tools.pdf


#70

Hello, Gordon;

can you help me in one of both ways?

1 - If you can, please, send me the PDF file ( lcvieira@quantica.com.br )
2 - post the page address instead of the PDF file address so I can download it; my browser refuses to download the PDF file and shows a plug-in error; I cannot see the file.

Thank you very much.

Luiz C. Vieira - Brazil


#71

Hello Luiz,

there is a trick I use in such situations: Create a HTML file on your local system which contains a link to the PDF-file. When you open that local file with a browser, you can right-click the link and select "save target as...".

Holger


#72

I'll do that.

Thank you.

Luiz C. Vieria - Brazil

#73

http://www.businessweek.com/magazine/content/03_19/b3832018.htm


#74

That page is so busy loading ads, it doesn't have time to load the article!


#75

No problem here.
Just disable Active Scripting and JavaScript,
and most of the other nice stuff,
then the page will load in three seconds.

Raymond

#76

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=3023467042&category=11712


#77

Just for the record...that is NOT being sold by me. :-)

Mine is not in the package any longer.


Possibly Related Threads...
Thread Author Replies Views Last Post
  HP 12C Platinum Programming v. Gold Face Dean Lewis 10 1,082 11-03-2013, 07:30 PM
Last Post: Kimberly Thompson
  HP12C Platinum PC software activation fails Russell Clinton 0 373 07-02-2013, 02:32 PM
Last Post: Russell Clinton
  HP 12c Platinum Emulator Activation Server Problem Danny Farley 5 586 04-09-2013, 09:13 PM
Last Post: Danny Farley
  Good deal : HP 12c Platinum for half-price Mic 1 403 01-21-2013, 03:51 AM
Last Post: Vincent Guilbault
  Simple financial problem with a HP-12c Platinum Lapin 3 496 11-11-2012, 09:33 AM
Last Post: Lapin
  HP 12C Platinum CHS bug? M. Joury 23 1,724 08-11-2011, 06:45 PM
Last Post: Gerson W. Barbosa
  HP 12C, 12C Platinum & 15C iOS App Walter Lam 2 428 06-02-2011, 01:25 PM
Last Post: Andrés C. Rodríguez (Argentina)
  Calculator for CFA; HP 12C Platinum or TI BA II Plus Professional Sujith Abraham 23 1,669 03-22-2011, 11:46 PM
Last Post: Martin Pinckney
  Self tests in the HP12C Platinum+ Jose Ernesto 3 435 02-08-2011, 02:31 PM
Last Post: Lyuka
  HP 12C Platinum Screen Anthony (USA) 3 433 02-05-2011, 12:19 PM
Last Post: Anthony (USA)

Forum Jump: