HP-33S this week


I'm sure someone that preorder will receive their calculator today... Who will be the first to comment the calculator???



According to the email I received from Walmart and their UPS tracking service, my HP 33s is scheduled to arrive at my door tomorrow :-) Can't wait to see what HP has turned out to replace the HP 32s, now that I can't use my 48gx or 42s on the upcomming NCEES exam this April. It's kinda strange that Walmart is selling these and HP still doesn't show them on their web site? Anyway my fingers are crossed that its as good and dependable as my other HP's.


Walmart says my order is still processing and has not shipped. They took the money out of my account on the 12th but have not shipped yet.........interesting....


Apparently Frey's (Wilsonville, Oregon) had some, on or before Saturday, as there was an empty hook with an HP 33S label and $54.99 price tag. They were out of stock on that one, and the 17BII, 12CP too. I saw my first 48GII ($99.99). The silver colored case was nice, but in terms of key colors they chose, kind of reminded me of the last-iteration of the 32SII.




Did you see the re-released HP48GII?



It think it must be the re-released 48GII. I've never seen one before now (the originals were taken back by HP to address battery drain, and bigger issues).

I wasn't impressed. For $40 dollars more, the 49G+ is a much better value, in terms of additional technology, convenience, and features (except I'd rather have the cool silver body color, over the gold :-)



Wal-Mart emailed me that the 33s had been shipped on the 12th, and that it "should arrive before 2/15/2004".

But even now UPS Tracking shows only that "electronic order information received" -- no details of package scanning, and so no indication that the item has passed into UPS control, let alone is actually in transit.

So, while it looked like progress, there still seems to be reason (for me at least) to not expect too much too soon.


My UPS Tracking shows that the item was entered into their system "Orign Scan" Feb-13th, and Departure Scan Feb-14th, delivery to my town Feb-17th. So I hoping all is above board and correct, thats all I have to go on.


Is this the way most new calculators are released from HP, by mail order first. Then later making them available in stores. It seems odd that as soon as they are announcing the release they would not be available sometime very soon afterward in stores.


I've seen that Walmart claims to have shipped some 33s calculators, but I haven't seen where HP has announced it.

Not sure why walmart says they are shipping them, but HP hasn't said "Here's the 33S!" yet.



It's not unusual for HP to start selling a model before it officially announcing it. For example, I was able to buy the HP39G+ (early version with battery-draining ROM bug) even though it is officially not released.



I think he meant that the previous poster incorrectly assumed HP had announced the 33S. They haven't yet.

The poster said "Since HP announced this, why isn't it available..."

But they haven't. Walmart appears to be shipping them, but they haven't been announced yet (that I have seen anyway).


The user manual is available online, and a product description with picture somehow became available last year some time.

What, exactly, constitutes an "announcement"?

I suppose it makes sense to fill a few supply pipes before creating demand. (Not everything can be marketed like Cabbage Patch Kids, right?)


Well, the manual is found only by directly searching for it - it's not publicly linked to the calculator site.

Go to www.hp.com/calculators and you do not see the 33S listed.

If it were announced, it would be there, I think.

The information last year (datasheet) did not come that I'm aware of from an official HP announcement.

So, to be announced, I would think HP would have to have made a public announcement. Seems fairly simple to me. :-)

The point was that since HP does not seem to have announced this product yet, there is no reason to complain about it not being available in all outlets yet.



Well, now the UPS Tracking site shows the package "In Transit" and to be delivered by/on 2/17/2004 (as in, today).

So, that sounds like independent confirmation to me. Short of Wal-Mart sending the wrong item, I think I can expect to see the 33s (for better or for worse!) tonight.


Are you sure they are not sending you 33c? (Thirty three cents)

Seriously though please post here when you get it.


If you don't mind me asking, when did you place your order with Walmart? I placed an order last week (when they put the 33 back on the webpage), and my status still shows "processing".



FYI, I ordered as soon as I read on this board that Wal-Mart was offering the pre-order online -- 15 Dec., 2003.


In case you're interested, the item itself seems to have originated in Portland:

DATE            Time         LOCATION            SCAN          


Congrats.........unfortunately they cannot meet demand........

Thank you for shopping at Walmart.com.

We're sorry to say that your order has been delayed due to
a delay in our fulfillment process. Unfortunately, we
were unable to ship the item(s) you ordered below by the
date we originally gave you. We are working with our
suppliers to fill your order as soon as possible, but we
are currently unable to provide a revised shipment date
for your order.

We will continue to attempt to ship this order. However,
you can cancel the order at any time up until the date the
item(s) are actually shipped.


Functionally, it's a good, silent keyboard, with no missing keystrokes noted.

It "boots" in RPN.

The MEM command shows 31,277 .

The vinyl case is like an open-topped little brother to the HP-49G+ case -- same materials & feel.

The manual is a paperbound, 3/4" thick thingy, much like the 32sII manuals.

Soft rubber non-skid material lines the lower sides, and the back feet are tipped with non-skid rubber. The keyboard face appears to be brushed aluminum, though I think the buttons under the display are painted silver. It's slightly longer, wider and thicker than my 32s, but about the same weight.

Negatives: The keyboard is incredibly busy, and the chevron style may accentuate that impression. And the decimal points in the display are almost invisible -- it would have been nice (from my nearsighted point of view) to have utilized a digit position and put a big fat dot in the middle of the character, rather than the teenly little pixel lost between digits.

But then, we knew it wouldn't be perfect, now didn't we?

I haven't programmed it yet, and I'll have pictures up tonight (~1800 PST) if nothing else is going on . . .


Negatives: The keyboard is incredibly busy, and the chevron style may accentuate that impression.

So it still has the lousy slanted keyboard? I had hoped (but not expected) that the delay would be due to a complete keyboard redesign...


Well, I'm glad at least someone has received theirs by now. I would assume that HP should be sending more to distributors as soon as possible. I'll keep my fingers crossed.


I had hoped (but not expected) that the delay would be due to a complete keyboard redesign...

Murphy's law: "Blessed are those who expect nothing, for they shan't be disappointed."




Good to know that it doesn't "clack" "clack" "clack". How much pressure to register keystrokes? Tactile feedback? Which keyboard does it feel most like?



HP33S checksums for Katie Wasserman's "99 Digit of Pi on an HP 32SII" program:

P=66  CK=BCC8
Q=51 CK=5A0B
R=51 CK=DC04
X=78 CK=36FB
A=36 CK=55DD
B=96 CK=12C2
M=39 CK=7400
N=57 CK=9826
D=39 CK=1F79
E=57 CK=5C9F
Execution time is 8 min 24s, that's ~20% faster than on a HP32SII.



FWIW, here are a few quick pictures:

A 3/4 view

The back and vinyl case

The battery compartment open

Side-by-side with my trusty 32s

A couple more things I've noticed: It has a Reset hole in back, and there are two screws (not heat stakes) holding the top end together from inside the battery compartment. (I wonder how long it will be before I try to take it apart?)

In response to a question, the keypress feels much like my 32s -- with maybe a bit more travel, slightly more pressure required, and a more definite "click" (though no louder).

Edited: 17 Feb 2004, 7:04 p.m.


Mabye some will blast me, but i think the look is not bad at all, no? :-) No purple/blue faceplate for this one!!! I can't wait to get mine! I live in Canada, No Walmart site to get it ;-(... The display look "grighter" than the 32s.

Edited: 17 Feb 2004, 7:23 p.m.


I got the hp 33s, the purple, green, and yellow/gold colors remind me of the colors of Mardi Gras. Especially the Mardi Gras in New Orleans where the babes dangle from the balconies flashing their boobs. ;-)

Whoever thought of the color scheme must have been a party animal during Mardi Gras....


What kind of batteries are in that battery compartment?

The data sheet I saw for the calculator said it uses 2, CR2032 Lithium "coin cells". Are they under covers or something?




They're each under a rectangular metal cover bearing the following warning (in red): "Warning! Turn off before changing batteries. Please replace batteries one at a time."


So, does it show both stack levels in the display in RPN mode?


In RPN mode it does pretty much what you'd expect. It behaves like the HP-42s display, but actually appears more like you'd imagine a two-row 32s/sII style display, with discrete dot-matrix characters -- it's not a full-screen graphical dot-matrix grid.

In PRGM mode, two program lines are displayed -- when you enter a new program statement, the lower line displayed is pushed to the upper, and the new statement is entered on the lower display line. The "up" and "down" functions of the four-function arrow key are used to scroll up & down in PRGM mode, and for BST & SST in normal ("run"?) mode. Scrolling up past the beginning or down past the end wraps around to the end or beginning (respectively) in PRGM mode.

In PRGM mode it will NOT let you enter duplicate labels. Too bad. Of course that's allowed in my 97 and others -- it would have been very useful with all the memory this thing's got.

I notice that filling the variables with non-zero values does not change available memory -- they're permanently allocated somewhere. Also, there aren't any half-bytes being accounted for in program length counts. (And for the program mentioned above, each line seems to have taken three full bytes.)

Menus are a bit funny -- instead of being tied to the top row of keys, they're arrayed in a two-dimensional grid (two lines of two or three selections each) with an underline cursor. You move the cursor with the four-function arrow key, and hit "ENTER" to select. (Each menu entry is also tied to a numeral, 1-6 or whatever, and you can select directly via the keyboard.)


Oh Wow. It's even uglier than I expected -- and I was expecting a lot of ugliness.

The side-by-side picture with the elegant 32S really shows how appalling this new little monstrosity is.


ugh, the 32s is ugly


ugh, the 32s is ugly
The 32S looks like a scientific instrument that would be right at home in a laboratory or on an electronics test bench. That's what makes it beautiful. The 33S looks like a stylish fashion accessory that would be carried by someone who wants to impress the "cool" people. That's what makes it ugly.


You're absolutely right. The 33S is the kind of "with it" gadget that should have a loop in one corner, so it can hang from a wrist while its owner parades the boulevard.

Next from HP: The 33SII, with "skins" of iridiscent colors (50 designs to choose from), keys arranged in a spiral or starburst pattern (buyer's choice), and voice feedback: you press the square root and hear "Mmmhh, square root, dear" in a sexually insinuating female voice.

If any of my predictions come true, I'll kill myself.

-Ernie (12345)


You mean you will "DELETE" yourself??? I'd rather stick to calculator parlance!


If any of my predictions come true, I'll kill myself.

Please don't! We need all the people with good taste we can get. :-)


Actually, it looks much nicer than I had expected. Guess I just had not made my mind up to dislike it from the start.


I kind of like it too. I'm looking forward to receiving mine; soon, I hope.


The 33s is better-looking and better-made than I feared. The additional memory is EXACTLY what the 32s line needed, and the two line display is a real improvement. But the glitzy, crowded keyboard is indeed a drawback (though it functions perfectly). Sure, it's not much more crowded than that of the 32sII (or even the venerable 67's), but I've always prefereed the 32S to the sII, and from my standpoint, the new design represents another step in a wrong direction.

The increased memory is indeed super, but the limit of 26 labels and 27 variables will inevitably grate on our sense of balance. Why provide all that tanalizing RAM, and then hobble it with usage restrictions? I can't help but imagine how much nicer it would have been had they simply extended variable naming and program labelling to TWO characters. (Then, of course, we'd all be whining about lack of serial I/O!) I keep repeating the mantra: "More" memory is the important part -- don't get hung up on whether you can easily use it all!

The layout of the keyboard is less than logical (I think) from an RPN user's perspective. And the small ENTER key is just plain tragic. These drawbacks are in practice (IMHO) far more significant than the weird chevron layout has turned out to be.

I do think the 32s display is noticeably clearer, and from a wider range of viewing angles, than is the 33s'. The 33s seems to have the LCD dots set somewhat above the reflective back, and so a shadow is cast and blurs the character images slightly from any direction but straight on. The 32s doesn't seem to exhibit this problem at all(!?!?) It's amazing that significant problems with keyboards and LCD displays that had been solved long ago seem to somehow resurface in newer models. I'd guess that this speaks to the lack of continuity in the H-P calculator division.

And, yes, I've already taken the 33s apart. It turns out that, in addition to the two exposed screws in the battery compartment, there are four screw holes in the back of the case, each plugged by a little cylinder of soft gray rubber with some self-stick on its back. Unsnapping the back after removing the six screws was a little harder than expected, and the result was distinctly non-impressive: a black-painted circuit board with two lacquered-over (or whatever) chips. There really isn't much to see in there:

HP-33s Internal View 1

and (not very different) View 2.

The PCB is attached to the calculator face by heat stakes, exactly like that of the 49G+. As the keyboard works so well, however, there didn't seem to be much gained by taking that apart. I took the pictures and put the thing back together. (It still looks and works fine.) The significant thing is that its case, at least, is put together with screws, which is how most of the TI's are, and how the 49G+ should have been.

All in all, it's GREAT to have a relatively-high-quality RPN keystroke programmable available again. The benefits do largely counteract the negatives, but the net effect is (to me, at least) that of a "usable alternative" rather than a certain-to-be-loved addition to a long line of amazing products. We'll see, I guess . . .

I want to wish good luck to those of you awaiting delivery. Mine (serial # CN40400032) was shipped from Portland, OR, so WalMart.com must have located a distribution center at a port of entry near China. You're probably waiting for the next container ship to arrive, and I wouldn't be surprised if it turns out Homeland Security has had a role in the delay.

Edited: 18 Feb 2004, 10:38 a.m.


Thanks Paul. Looks like a winner!

Other than the ENTER key, I think all key positions are essentially the same as on the HP32SII.

That's why the LBL is above the + key, etc.

The ENTER key moved (and looking at the pictures, I TOO wish it were double wide and in the middle), but most function locations are the same as the 32SII. Take a look at the 32SII and you'll see what I mean.


I'm looking at my 32s -- I don't have an sII at hand at the moment -- but I don't think those two differ significantly vis-a-vis what I consider the drawbacks of the 33s' layout.

On the 32s, the stack manipulation keys are a nice tight group: ENTER/Last x, Roll Down, and X<>Y are all close together, and near +/-, E (enter exponent) and left-arrow (edit-delete).

One the 33s, Roll Down is in the second row; moved away from X<>Y, +/-, E and left-arrow in the fourth; and ENTER/Last x is way down in the last row. Quite apart from the unfortunate size of the ENTER key, the scattering of these functions seems unnecessary and illogical.

It's nice, I suppose, that R/S and XEQ are together, but they're occupying "prime real estate" that would more logically have been taken by the more frequently-used ENTER and Roll Down. (Admittedly, my "more frequently-used" claim applies to RPN mode . . . )

The breakup of the stack manipulation group is, I think, far more deleterious overall than has been either the reduction in ENTER key size, or even the glitzy key shapes.

But, of course, this is purely subjective. Overall, I repeat, I'm happy that a good, RPN scientific keystroke programmable is available. If the glitz helps keep 'em coming, I'll live with it.

I do hope, however, that it's not too very long before a sober, rational, HP-33r ("retro") model is the latest thing!



apart from the diffent design, two-line display and bigger memory, are there any differences in its math capabilities compared to the 32sii? E.g., can you calculate complex arcsin(1.5)?

How is algebraic mode diplayed? How are complex numbers represented in algebraic mode?

Thanks in advance


Good questions. Unfortunately, I'm probably not the person to answer.

Playing with it, it seems that, in RPN mode <L> (left-shift) CMPLX acts as a prefix for operations, causing those operations to work on the stack registers as two pairs (x,y & z,t for binary operations) or one pair (x,y for unary). As the display shows x & y, I suppose both parts of the complex result are visible. Specifically, <L> CMPLX SIN seems to calculate and return a result, but <L> CMPLX <L> ASIN appears to be ignored.

In ALG mode (this is the first time I've tried anything this way), the top line of the display is turned off, until a series of values and operations are entered. Then the top line shows the formula, and the bottom line shows some sort of intermediate result. I am acquainted with this sort of behavior (having used a TI-83+SE quite a bit), but I don't seem to be getting the hang of what's going on with the 33s' ALG behavior. (I'm sure it's logical, I just don't know what the logic is.)

But in ALG, <L> CMPLX seems to set the calculator into "complex mode", with "0i" displayed on the top line and "RE=" on the bottom line, and the up/down functions on the four-function arrow key (the big middle key below the display) used to toggle the bottom line between "RE=" and "IM=" values. There must be some way to enter the real and imaginary parts, but I don't know what that is. Here again though, SIN seems to calculate and return a result, but <L> ASIN is ignored.

I obvously don't know what I'm doing here, and haven't read the manual, and so must admit I can't communicate more effectively for you than the above.


Paul --

Thanks for the photos and discussions for an interested audience.

Several people (Axel and Tizedes) have asked whether the 33S is able to perform certain advanced mathematical operations that the 32Sii could not. I think that you had the answer to Axel's question, but didn't quite commit.

I'd wager a moderate sum that the answers to both questions are "no" -- itcan't do inverse functions of complex-valued arguuments, and can't use INTEG within SOLVE or vice-versa, and it can't perform double integrals.

What people need to understand is that the upgraded models incorporate very little effort in terms of ROM improvements for more mathematical and operating functionality. Genrally, KinHPo has added AOS (due to market demand) and a few simple functions, along with more RAM to the replacement models. Most of the deficiencies of the legacy models still exist.

Thus, the 12C Platinum has more programming memory, but the user still can't insert/delete instructions.

The 33S has much more memory, but there still are only 26 labels and 27 registers -- just like the 32Sii. There is no particular reason to believe that KinHPo upgraded the pidgin complex-number functionality of the 32S/ii (apparently based on the 41C Math Pac); or, that they troubled themselves to figure out how to make SOLVE/INTEG work together in a nested manner under the 32Sii's paradigm of RPN routines and equations (let alone AOS routines).

They've removed the limited-RAM constraint (thanks to plummeting costs), but to fix the other problems would require skilled work.

-- Karl S.


I think you've got it exactly right. The 33S is basically a "32SIII"; the only significant changes are:

1. Two-line display (a plus)

2. Far more memory (a plus, but not as much as it should be, since there are still few labels and variables)

3. An ugly keyboard with an undersized and poorly located ENTER key (a minus)

4. Algebraic option (a plus for some people)

I expect the 33S, like the 32SII, will actually have 33 storage registers, not 27. These are registers A-Z, plus the i-register, plus the six statistics registers. Still not enough though.


Thank you Paul & Karl,

that's what I had expected. I'd probably wait for the 43S.



I'm think I'll collecting more money, and I'll buying an N-GAGE... :D Or a better idea: I'll back to my lovely HP15C... No more Y2K design...

Thank you again!



You're probably right about the SOLVE and Integrate functionality. After having played with it a bit this morning, and having (briefly) compared the manuals, I see only one really new related feature: the equations list. (And I don't remember whether such was available in any form on the 32sII.)

Early on, when I/we were complaining about the 32K memory but only 26 labels, someone pointed out that the equations list would be another way to make use of available memory, and that seems to be the case. Equations are entered (with, admittedly, limited editing capability) in "Algebraic" form. SOLVE and Integrate, when operated with the equation list displayed, operate on the current equation (rather than a program designated by FN=). For users of SOLVE and Integrate, this should be a truly useful improvement, and one which reduces their need for programming.

As far as the "skill" involved in creating the 33s, I don't think the effort should be denigrated. I think that H-P's goals (however flawed we may think have been the choices that led to them) have been admirably realized. Both the 33s and the 49G+ (two new models with which I've been acquainted) represent flashy (O.K., "consumer electonics") packaging of aging H-P functionality, implemented as emulations using off-the-shelf CPU technology. They're "leveraging" (or strip-mining, whichever you prefer) their code base to get new product out.

In trying to re-start the H-P calculator operation and at the same time take on TI and Casio, this first step seems a logical one, and I'd say, of the models I've seen, the 33s may be the best-executed part of that strategy so far. I'm sure they had PLENTY of "skilled" work to do with this first step, without taking on fundamental enhancements to the most complex functions.

Remember, this all had to be sold to the corporate board (or someone), and has to succeed if it's to have any legs at all. And the HP-33s looks to me like it can be a real winner. Take a look at it and try to find something functionally equivalent in a package of similar quality.

I don't think it's far-fetched to say, "If you want a 43s someday, buy a 33s today." (For your high school kid, at least . . . )

(And no, I'm not paid by H-P and have no connection to the company!)


What do you mean by an equation list. Can you store equations as a variable, or is there simply a scrolling list of equations(as in you can have many different equations at one time and scroll through them; the 32sii had that)

As far as the many complaints about the memory issue, the reason HP included lots of memory is because of the complaints of too little memory on the 32sii. 384 bytes is not enough memory. There is now more memory; in fact, more memory than anyone could ever use. I for one am tired of getting out of memory notices when I try to store a variable or solve an equation (I tend to run my calculator at about 340 bytes full, and if clearing all the variables and deleting equations does not free enough memory, I clear the memory) I don't use the 32sii as a program vault like graphing calculators are used. The programs I have on it I write for a specific reason (such as currency conversion for a trip to mexico, reimann sums for checking calculus answers for a week, etc). The longest time I have had a program on my calculator is about 3 weeks (maybe a month) and that was my "guess the random number game" (which is strangely addicting)
The 33s's massive memory bank is there so that you can't possible use it all. If you want to have 128 mb of memory on a calculator, buy a 49g+ and a SD card. Also, it probably cost HP about the same amount, if nto less to include a 32kb chip of memory rather than a 1.5 kb chip of memory. 32kb is a very common memory size for calculators, and so Kinpo would have experience with using them.

Take the 31kb of memory as a blessing, not as a crutch. It will keep the out of memory message away (I wonder if they even included the out of memory screen?)



The equation list is the latter -- a scrollable list that starts out empty, and into which you may enter as many equations (presumably) as memory will allow. If I've got this right, you enter "EQN" mode, scroll to the equation desired, and hit "SOLVE" (or Integrate). (That's the way it seemed to work for me.) It looks like a really nice feature for calculator customization.

Re: memory, I agree: "More" is the idea, and the thing to keep in mind. No doubt, had they extended label and variable names to, say, one alpha followed by one digit, much larger programs would have been made possible. But then we'd want serial I/O. (One thing leads to another . . . )

Having all that memory that seems unusable apparently offends some "sense of "balance" in design. But again, "more" is available, and that will keep me happy.

And yes, the "MEMORY FULL" message is referenced in the manual, and does indeed show when (for example) ~10,500 program statements are entered.


It's true that you probably won't get the "MEMORY FULL" message on the 33S. But that's because you will get the "DUPLICAT.LBL" message long before the memory can be filled with programs.

In fairness, the increased memory of the 33S (relative to the 32SII) is an improvement. But it's a bit deceptive for HP to market the 33s as having 31KB of memory; in practice, the lack of labels means that only a few of those KB are actually usable for programs


If you had ever used a 42s with long variable names, you would cry. Since you have not, you rejoice at the extra RAM. However, Hp only had to take a small step further, to give us a classic. Instead, we get a 33s. Now it may someday become reguarded as a classic, but the 41c and 42s (though even the 42s has shortcomings) were immediate classics, without reservation.


I'm not sure I would ever have called the 32SII a classic to begin with, and the 33S is a replacement for the 32SII, not one of the classics. There really doesn't appear to have been an intent to re-engineer one of the classics but to replace what they could no longer make...plus add a couple of small things (algebraic, for example).

I still hold out for a 43S. I can hope, can't I? :-)



I guess that in the ALG mode, complex numbers are easier to handle...


Someone pointed out via email that one needs to use parentheses (both are shifted keystrokes) to enter the real and imagined parts of complex numbers. So, from a keystroke standpoint, I think the RPN handling of complex numbers is much more straightforward. (Of course, I'm an RPN user, so it would probably seem so to me in any case ... )


Here's what real good complex-number capability would provide a +i* key to enter the imaginary part of a complex number. It would work like EEX: Once +i* is pressed, all subsequent digits are entered into the imaginary component; backspacing would remove digits one-by-one until "i" itself was removed.

Neither "(, SPACE, )" (28/48/49), "ENTER, I" or "Re<>Im" (15C), nor "ENTER, CMPLX" (42S) would be necessary.

Does anything so straightforward exist on a calculator?

Speaking of CMPLX:

On the 42S, it would have been nice if that were a menu for a family of functions (as on the 28C/S), rather than merely a complex-number assembler/dissembler.

On the 32S/32Sii/33S, it would have been nice if CMPLX worked like "SF 8" on the 15C, providing access to all the mathematical functions with complex arguments, rather than serving as merely a "Simon says" prefix.

My thoughts,

-- KS


It is kind of confusing remembering every time which of the numbers of your stack corresponds to the imaginary part. The algebraic mode shows clearly the two parts (even with a marked "i"). I'm also an RPN user (10 years with my hp48gx...I know, it is a luxury that can't be compared...but....)but certainly for operations with complex numbers I'll switch to ALG.
These are also other things that I expected from a just released HP calculator:
- If they handle operations with complex numbers, why I can't see complex numbers as a result of SQRT(-1) or Arcsin(-5)?
- If the 15c could handle vector and matrix operations, why the 33c lacks of that? I know that you can program 3*3 matrix inverse or determinant at most but labels are really few for 32k, and the variables also. The basic programs that the manual shows uses most of the labels for things that the calculator should have "in-build": linear equation solver, vector operations (dot, cross, polar to rect and rect to polar), simple matrix operations (inverse, determinant)
- Casio "simpler and cheaper" calculators have matrix and vector capabilities. Of course they are not the RPN realm but they make the job (see casio fx-991ms).



this calc is can to SOLVE and INTEGRATE at same time? And what about multiple integrals?

Thank you!



Thanks for the pics Paul. Not only are they detailed but arty in an "Alien" sort of way. I'll try to get one at Frys next time i pass by (a 33s - not an alien) and maybe give my 41 a break. Wally World will be my last resort.

I'd like to share one with you. It's an internal of that other RPN company's new RPN. Remember when there was stuff to see in a calc? These new ones can be photographed on a scanner. http://www.msdsite.com/forums/upload.php?&upload=zoom&pid=315


These new ones can be photographed on a scanner.

I'm not sure about the cheap contact-image-sensor scanners made today, but any good CCD scanner like the HP ScanJet 3C/4C/6100 will work fine even for things that are far from flat. I was expecting the focus range to be fairly tight, but was pleasantly surprised that I was able to get good scans even of PC boards with some 1/2 inch tall components.


Nice photo of the Aurora 1000 innards.

Has anyone got a good photo/scan of the HP 12C Platinum innards?


Hey! It's encouraging to hear that it's a nice little unit, although I am one of those who was more than startled at the faceplate design.

I'd love to have one, but I wonder if I should hold out for a "33R" ("retro") version! Anyone want to lay odds on whether HP might be keeping an ear or eye or two on the pulse of us geeks, nerds, and what-have-yous expressions of preference for a more "traditional calculator" case design? Could I dare hope for a 34C style case??

This makes me think of PCs today, if you build your own- the latest rage is in the blinky LED bedecked more plastic than metal cases. I just built my own just over a month ago, and I went "retro" with a staid, steel case my brother-in-law gave me because he (a grown man) wanted one of those blinking LED plastic jobs... and get this- none of the LEDs are red! They're blue! I know they're new, I know they're hip, I know until GaN chips we never had 'em, but ugh, red (or even green, yellow, or orange) seem a little less... harsh and stark.

If anyone wants to know or if it makes anyone feel better, I went out of my way slightly to get a clicky, though plastic keyboard (Focus FK-2001) which is VERY affordable... and like almost all PC keyboards, the ENTER key is still in the same place and much bigger than the rest!



I'd love to have one, but I wonder if I should hold out for a "33R" ("retro") version! Anyone want to lay odds on whether HP might be keeping an ear or eye or two on the pulse of us geeks, nerds, and what-have-yous expressions of preference for a more "traditional calculator" case design? Could I dare hope for a 34C style case??

Recent events have demonstrated, beyond any doubt, that HP isn't listening. I can earnestly hope and wish for a rebirth of the HP-67 or 41C, but it ain't gonna happen this side of eternity. Wishful thinking is one thing; pipe dreams are quite another.

Sorry for the cold water bucket, but this morning I'm feeling more cynical than usually.

Ernie (12345)


That's not cycnicism, it's truth. HP long ago stopped listening to us "calculator geeks" and started listening to the bean counters.

The sad fact is, I would be willing to pay twice as much (more, actually) for a NEW 42S (or a carefully thought our replacement) than the retail price of the 33S. I guess there just aren't enough of us out there to have a voice in their marketing decisions. They seem to just want to do everything on the cheap, but I for one am willing to pay more for truly high quality tools.



"On the cheap" is the new American consumerist motto. Cheap high volume buy-today-trash-tomorrow consumerist junk. HP just responding to the (lack of) character of American consumers and investors. Cheap sells, baybee.


Now THAT is cynicism.

Take care.



The question is listening to whom?

I am not sure that it is a profitable venture to build something the 1000 calculator geeks (like myself) would pay $250 for.

The broader market just won't pay that $$ for such a thing, IMO.

Disparage bean counters if you like, but products must have a financial success probability close to 100% or it is too risky to do them.

(Who really wants a 43S !)



1000 calculator selling at $250 sounds like something that a chartity-geek organization would do.

HP is already behind in the calculator market. Rolling out expensive new calculators that will be easily outselled by well-intrenched TI calculators is not the kind of success story HP is after!

My girlfriend works for a local university bookstore and her store outsells TI calculators (the cheaper ones sell even more) compared to non-TI machines.



I certainly meant no offense to any bean counters out there. I recognize that the prevalent corporate mentality today is that of minimizing risk to near zero; but that is not the type of thinking that leads to great inventions or advancements in technology.

I lament the seeming demise of the attitudes of the founders of companies like Ford, Harley-Davidson, IBM and of course Hewlitt-Packard. These and many other large companies which (at least at some point in their history) manufactured and sold products based on new and/or untried technology would never have gotten off the ground if their founders hadn't been willing to take enormous personal and financial risks.

How much risk to their overall financial well-being would HP really be taking if they re-issued or updated the 41C or 42S or some other classic and offered them at $200+ each? I suppose it could lead to some top executive not getting a big enough bonus that year to be able to afford to pay cash for his new BMW750il or Cadillac Escalade; and that would be a modern American tragedy. (Darn it, I was trying for cycnism and I got sarcasm instead... I'll try again later).

Take care.



No offense taken!

But, that's just not how it works today in large companies.

Companies in many cases don't want to be bothered with small divisions and products anyway.

HP will not make appreciably more money with calculators today nor appreciably less.

I consider it a good thing they are still making them, but it is not because these drive an appreciable difference to the bottom line (or else, I would think they'd have their own page in the annual report) :-)

A company that makes $5 billion in revenue often won't bother with a product line bringing in a whole lot less than that (and I have no idea how much revenue the HP calculator division makes).

HP almost killed the calculator area when ACO went down, remember.

Some people will look at things like the HP33S keys and say, "I'll never buy something like that". Ok, then, I would say good luck on your journey. The old keys probably will never come back. I personally don't ever see the double-wide ENTER coming back. I just don't see the change coming. (Note: entirely personal thoughts here - no "inside" information of any type).



Really, I'm not shure if it's a disadvantage not to have a wide enter key. The Human is very easyli getting used to something - it's like unerasable burned in our brains. But i think it's only about getting used to the "narrow" enter key. There is no important reason (please don't eat me for this.. - it's only my opinion) why it should by wide - every number has to be hit with the same precission as the enter key - there is no order of importance betwen 0-9, x, -, +, :, enter and some others keys. The only thing i could mention it would be usefull for:blind typing - but with the enter key on the bottom you'll find it blind too, would I believe.

So imagine there had never been a wide enter key - and suddenly the 33s would have one. Try to be honest to yourselfs - wouldn't we all think:"hey are they crazy putting in a wide enter key, what is it good for, we would prefer one more key instead!"

Fact: im used to the wide one about 14 years now, too. Maybe I will revise what i've written in this post as soon as I've got a 33s ;-)

1234 to eat this post


The HP11-16 series ENTER keys weren't very wide either!


..you're right there. ..ok, let us talk about a "long" Enter key ;-)

If you turn them 90 degrees clockwise the key will also be wide but the display narrow and the key labels confusing to read ;-) ;-) ;-)

It would be interesting to know why the engineers decided to make the enter key longer than the other. Im often thinking about reasons for it. But maybe it mainly was a marketing strategy, tho show that it's a RPN calc and to show that it's a Hewlett Packard from long distance.
It could explain why they changed it since the 49G(I believe it was the first,not shure) - because all new ones aren't really HP anymore(does someone not agree?). The "old" HP engineering team maybe never would have changed it because it belonged to the image, spirit and appearence of the HP RPN calcs..




The "old" HP engineering team maybe never would have changed it because it belonged to the image, spirit and appearence of the HP
RPN calcs..

That is exactly why I object to the new, smaller ENTER key. It's clear HP is trying to distance itself from it's old image of a technology-driven, "We're the best, if you can't appreciate RPN you're not smart enough to be our customer" kind of company. My feeling is if they're too ashamed to put a big, prominent, in-your-face, "THIS IS AN RPN MACHINE!" ENTER key on their calculators, then I'm too ashamed to be seen using one of them.


These discussions of 'Red Guard Capitalism' get my goat.
The problen IS the bean counters, and I haven't the slightest respect for them.
They are busy destroying inventiveness on ALL levels.

Oh gee, IBM ONLY made XXX million profits instead of XXX.0001 million profits.
Oh gee, POOR HP only made......etc.etc.
It has NOTHING to do with profit;
It only has to do with endless greed, monet grabbing and RIDICULOUS payments to top management.
It's time the beancounters were sent home before all inventiveness goes out the window in the name of super-profits.


It would be nice if some people with your passion were listened to as much as some others (passionate about sure profits) apparently are. It don't seem to be so.


Now, do we throw out the baby with the bathwater? Is all lost? Is there no value whatsoever in the product we're discussing?

You're entitled to your opinion.

Good luck sending "the bean counters" home. I'm sure it will make for interesting reading in the Don Quixote Journal.


You know, outright dishonesty is not actually an acceptable part of a true capitalist system; nor is unchecked greed. I am not particularly enamoured of the heretofore unseen levels of personal greed, moral cowardice and unethical behavior which (in my opinion) are at the heart of the United States' current situation regarding corporate risk taking and profit making. I thought I had made that pretty clear in my previous posting. However, until a better system is developed, we need to patch up the one we have and limp along with it.

Winston Churchill once said something like, "Democracy is the worst form of government on the planet... except for all the others." I would paraphrase him as follows: "Capitalism is the worst economic system on the planet... except for all the others."

I'm BUYING the 33S regardless of how greedy HP the Corporation is; regardless of how ugly or disfunctional someone else thinks it is; regardless of whether it shows my poor taste in consumer products or whether it proves that I suck or whatever. I WORK hard for the money I EARN, and I will SPEND it as I see fit. If someone else chooses not to support HP and to not purchase their products then that is also fine. That is the beauty of the capitalistic represenative democracy in which I reside; freedom to choose.

Whew.. Well then...

So, Paul, any more opinions (good or bad) on this new HP calculator? I read your posting about the ease with which the LCD can be scatched. What happened? Any ideas on rigging up a "screen protector" for it?

Take care.



Hi Wayne,

Don't get me wrong! I am all for positive capitialism!
The term 'Red Guard Capitalism" has been coined as a description of the lousy kind of capitalism that has absolute laws like 'profits at any cost' like some eccentric 180 degree reworking of Mao's texts.
Profit and growth are healthy things.
But it can go too far.
It should be enough to produce a good product, put people on the payroll, pay your taxes, and invest in the future.
If profit is left after that- and your growth is faster than inflation- then you should be happy.
Today's company growth is too fast, and the foundations get lost.
No, Paul, I can;t change it!
And I will still buy a 33s!
But something is going to have to change in the future to avoid more bubbles bursting in the economy.


Can't say I disagree with that.



We live in strange times.

Parents upset with the Honor Roll in school because it might make the other kids feel bad.

Increasingly restrictive regulations and "outcome measurements" in scientific medicine, yet practically no regulation whatsoever for the "alternative" medicine "practitioners" (and a huge push for insurance funding). Most of the educational meetings have stopped because someone might get dinner.

Almost nobody knows how to service the complex heating and air systems that exist in large buildings.

The hospital has replaced it's nurses with "Medication Technologists" with two weekends of training, then...I kid you not.... sends a letter reprimanding doctors for being impatient with the new staff for not being able to give or receive a basic medical report, and telling us it is "an opportunity to educate our new colleagues". (I thought that was what school was for...)

It's not about calculators that work well. It's about calculators that look cooler than the ones next to them on the shelf. I'm guessing few people will ever do more than add 2+2 anyway, (which now requires a calculator.)

Is that cynical enough, Wayne? <grin>

No wonder I admire the classic calcs. Marvelous engineering by people who seemed more interested in quality than "regression towards the mean", which is the new push in society.

P.S.... I've decided what alternative medicine I'd like to practice... trephenation. It's thousands of years old, so it must be valid! I'll see if I can get paid for it, and I plan to offer it to the Governor since he's so in favor of funding alternative medicine and removing all licensing requirements for alternative practitioners....


I already have enough holes in my head, thank you.

And yes, it was just cynical enough.



Far be it from me to be pedantic, but...

Do not confuse accurate description of circumstances with cynicism.

For instance the statement "HP are producing pretty ordinary calculators for an elitist market" is debatable but not cynical. The statement "it is waste of time waiting for the avaricious bean counters at HP to produce a decent calculator" is cynicism. The difference is in the imputation of motive.


I agree with you (see my previous post, above: "That's not cynicism"). However, Michael's post certainly did not lack cynicism. Of course, it also did not lack accurate descriptions of circumstances. As I said, "just cycnical enough".

Take care,



There are about 50 governors who might benefit from this procedure. To which one were you referring?



Iowa's governor, Vilisek, wants to pass a bill making it unecessary to be licensed to practice alternative medicine and ensure equal access to insurance payments.

I think I'll use my HP-25 to calculate the best placement for the holes....

Or maybe I'll perform the procedure on myself, so I won't care as much about the news....

I guess I'm just stuck in the past...



I would have guessed Massachusetts. Iowa is a "flyover" state. Your governor should have more sense.

Although I should not be insulting anyone's governor. I live in Virginia, under the Mark Warner regime.



I'm next door. Our governor (Johanns) vetoed an emergency mental health funding bill the day after his son tried to commit suicide (public information).

Years later, he appears on the front page of the paper touting his plan to "fix the mental health system" (that he watched fall apart).

To me, it's like setting fire to a building, rescuing some people, and being a hero. What about setting the fire? "That was a long time ago, I've done great things since then!"

Sigh. THANK YOU, YANI, For reminding me that this isn't cynicism as much as truth....

Happy Programming, (HP)


Lecture for Sunday

The glass is [ 60 COS ] full

As many preachers, I will talk about things I have not seen by myself. I rely in the inspired words of brother Paul, perhaps one of the first few to saw them, and one who generously shared his enlightened knowledge with us.

After a long wait, the 33S is with us. It shares the “33” number with older ancestors as 33E and 33C, and the “S” symbol with many respected RPN Pioneers like 32 and 42. Its family roots are so established. While the words “Hewlett Packard” seem not to appear at first view, at least the HP symbol is there, giving us some matter to hang our beliefs on.

I need to tell a short personal story in order to show why I’m convinced that the glass is half-full, and not the other way around. When my older daughter was 13 (now she is 19), I bought one of the last 32 SII available in Buenos Aires (old color scheme), to be used in her secondary school. Now she is an engineering student, and a second-generation RPN fan who cannot use regular calculators anymore. Worried about the possibility of having made her dependent on an almost extinct notation, I later bought her a second 32 SII (alas, new colors) just as a backup to ease the upcoming shortage pains that seemed bound to affect all RPN users.

When my second daughter entered the last years of primary school, I wanted to give her a scientific calculator, to introduce many of the mathematical ideas she will later be learning. The only thing available was the 30S, not RPN, poor quality, bad keys, no manual… Trying to avoid the already mentioned RPN shortage pain syndrome, I assumed with little comfort that perhaps algebraic notation was adequate enough for her. Now I have a better option for her.

For years we asked in this Forum for something probably close to a 43S. The 33S is not that machine, and has a busy keyboard with a strange chevron layout. But, at least, it is RPN, its keys seem to be better than those on the 49 series, overall quality seems acceptable, it comes with a thick manual, and is affordable enough to fulfill HP's internal success criteria about volume sales (yes, business are a part of it, whether we like them or not).

As many people said before, at least is a RPN programmable we can buy today and use without risking to lose, wear or damage our priceless 41, 42, etc. And I think that the price/functionality ratio is more than acceptable; perhaps the best of all current HP calculators.

Older models will not be remanufactured again. It is plain impossible (the old components and chip manufacturing processes are no longer available) and, further, the economics of so doing are more than dark. And, let us not forget, old classics also had their share of defects and limitations; things that our love for them should not prevent us to see.

Just as examples:

The 41 is one of my favorites, but mine would be working was it not because of internal bad contacts, damaged flexible backplane and broken plastic screw posts. How nice would have been for it to use standardized connectors for peripherals! (granted, there were few such standards at the time). Were 64 registers enough for the basic 41C?

My 25 is also a favorite but... How many Woodstocks have died because a battery pack connection and charging circuitry which were less than excellent?

My 42 is a third favorite but... How many Pioneers are impossible to repair due to an almost unopenable case? How many have problems because of plastic dome keyboards? How many 42 users needed a serial I/O which was purposefully omitted? Was a good idea to enter Alpha symbols via just six softkeys?

How many Champions lost their battery doors, or suffered from damage on the flat, flexible cable that connect both halves? Was the 28C useful with just 2K RAM?

My 48G+ is not a favorite. In any case, for 48 and 49: Are RPL and the infinite stack good things for a calculator? (my opinion is negative) Wouldn’t be nice not to keep nagging to the user all the time, just as bragging about error message variety? Wouldn’t be nice to implement a NULL function for mispressed keys, as in the HP41?

How many Spices had bad chip contacts or loose battery connectors?

Was it reasonable to charge almost U$S 1K (1975 dollars, I mean) for a 100-nonmerged steps HP65 with almost no editing functions nor conditional test capabilities?

(I don’t have many issues with the first Classics and Voyagers, but there may be some)

Moral 1: Like it or not, all families and models had their share of defects and problems. And, while we may miss the old models, the old company, the old times, etc., we should concede that things were not as perfect as our fond memories may suggest today.

Moral 2: Embrace the 33S, try to make the best from it, keep asking HP for that elusive 43S (if they have successfully migrated the 32 SII functionality to current hardware, they may intend to migrate the 42S someday...), for a better keyboard layout, for double letter labels. I think that, looking from the 30S to the 33S, HP has put at least some attention to our voices.

Moral 3: Let’s hope that some certification authority will not ban the 33S based on equation lists or full-of-prompts-programs that can be written in almost 30 Kby RAM… (cross our fingers)

Moral 4: Electronic auctions prices should drop to reasonable levels for old models, allowing bona-fide collectors to keep their hobby running.

Moral 5: Creative electronic wizards may attempt to turbocharge the 33S: Shadow memory initialized via USB? Display signaling and keyboard matrix switching for simple I/O in a 97S style? There are reasons not to discard these possibilities.

Moral 6: Time passes by, so let’s enjoy your lives and the good and affordable things available while we can. As I said once, “HP was what it was at the time when we were who we were” The world (and we all) have changed a lot in 25 years, some for better, some for worse. Things are not (and will not be) the same again.

May you all be blessed

(Feedback welcome!)


I agree,
HP is definitely going in the right direction. This reminds me of studying business applications of derivatives in calculus. We talked about how it is okay to lose money for the first 3 years as long as they are at a relative minimum.
HP made great calculators, and started out at a fairly high position. They then got higher (41, 15, etc) and then eventually dropped down, hit rock bottom with the 9s/9g. They then started going up again (the 12cp is better than a 9g), and apparently are increasing more now. Realize that calculator quality is increasing. All we have to do is wait for that next relative maximum (whenever that may be). Perhaps it will be for my generation

-Ben Salinas



Let's hope that the first derivatives (for quality, functionality and "HP calculator worthiness") continue to be positive, and that the second derivatives will not get negative (at least for a while)!

Thank you for your feedback.


Good post. However

'Was it reasonable to charge almost U$S 1K (1975 dollars, I mean) for a 100-nonmerged steps HP65 with almost no editing functions nor conditional test capabilities?'

The HP65 had a large number of conditional tests. They are in blue on the second row of keys.


You are right, I intended to mean "few" conditional tests, as there were only two flags and a "not-so-complete" suite of conditionals. But "few" slipped from my own "thinking-in-a-foreign-language" and "text review" processes. Thank you for your correction.

And, by the way, the HP65 was indeed a wonderful product for its time. I was most impressed in 1975 when I read about it. What I was trying to state in my post is that all models had their share of "possible improvements".

Edited: 22 Feb 2004, 11:16 a.m.


Well said. Thank you.

Take Care,



Thank you, Wayne.



Thanks for a very inspiring sermon. It's uplifting and doesn't insult HP or its customers.

I agree with you on most of your points: yes, the old calculators did have their faults, and yes, market forces and other issues will prevent their revival. We can only look to the future and hope.

The 33S is not perfect but it looks good enough. If the weird key layout does not actually interfere with my ability to use the machine, I can live with it. Likewise, I'm not crazy about the small Enter key but I can cope. I'll probably get one so that I don't have to worry about losing one of my better ones. And if it is successful enough to inspire HP to build better follow-ups, so much the better.

I can answer one of the questions you posed:

Was it reasonable to charge almost U$S 1K (1975 dollars, I mean) for a 100-nonmerged steps HP65 with almost no editing functions nor conditional test capabilities?

The answer is "yes"! The people who bought it must have felt that what they were getting was worth the money. Let's face it, a fully programmable computer that could fit into your shirt pocket and save programs on magnetic cards is pretty cool. Especially for 1974. And don't forget, an HP-65 was used in one or two space missions as a backup computer. Surely they thought it was worth it!

There's no point in comparing today's better technology with yesterday's. Twenty yeard from now, people will look at our 2004-vintage calculators, laptops and PDAs and say, "What a ripoff! Who would buy such a thing?" Well, critics from the future, we would. They're the best we can get as of now, and while they can't perform <applications yet unborn>, they do serve us well right now, and are definitely worth the money spent.

Sorry that this is too long, but I'm evading homework.

Again, thanks for an excellent essay.

- Michael F. Coyle



Thank you for your comments. Perhaps I went a little too far with my criticism, specifically for the 65. It surely had some conditionals (just X vs. Y) and, as you said, people who bought it did so because it was worth that money (if not, they would't).

I'd like to say, again, that my idea was not to be rude with such marvels of the past (I much admired them at the time), but to point that no product is perfect; so there is always room for improvement and need for some adaptability on the user side.

For instance, we criticize the 33s keyboard (me too), the Enter size, the Enter vs. R/S position, the stack manipulation keys disjoint placing... But after discussing the 65 features, I realized that it used shifted keys for x<>y and roll up/down (the 67 did the same), which may also be regarded as non-optimal.

While I like the uncluttered HP41 layout, I am pretty sure that the presence of BEEP on the keyboard was mostly to raise market awareness about this feature, rather than a functional decision. And, also by the way, I felt disappointed when I discovered that my brand new 42S had no units management or conversions under the prominent CONVERT menu (only INT, ABS, FRAC, etc.).

Let's give the 33s a chance for what it's good; and let's hope that more "serious RPN" models will appear, even if based on a similar platform.

Thank you again. Best regards.

Edited: 22 Feb 2004, 6:45 p.m.

Amen, brother Rodríguez!!!

There are a lot of messages out there comparing the 33s to this model or to that model in HPs esteemed line of handhelds. I bought a 48G about ten years ago in college because that's what all the "cool" engineers had. Honestly, I used it in my beginning mathematics courses, and never really got much use out of it after that. When I graduated, I still used the 48G for a while, but it just seemed like too much for what I actually needed. It had some great programming potential, but it was slow and cumbersome to run. Many of the engineers in my office used the 32sII, so I dedided to give it a try... and I've been married to the line ever since. Simply put, the calculator was the best "fit" for my needs, and I think that HP recognizes this. There isn't one all-encompassing Swiss-army-like handheld calculator that will be perfect for everyone. The 32sII runs the numbers I need without being overly excessive in its functionality (like the 48G was). If I need to do something beyond its capablity then I turn to a dedicated mathematics computer program.

The only problem I've had with the 32sII is it's lack of memory. I have it maxed-out with just a handfull of programs, and would LOVE the opportunity to have just a litte bit more. Now I don't think that I need 32K's worth, but, hey, memory's cheap, right? I went ahead and ordered a 33s, and can't wait for it to get here. I've already got plans for it.

Aesthetically, there seem to be some people out there who don't like the new design... and they are as much entitled to that opinion as I am about mine. Personally, I think that it's kind of cool. My father was a product designer, and spent his entire professional life taking items like the simple calculator with its "box with bottons" appearance, and turning it into something a little different... even for those of us (engineers) who aren't accustomed to such.

In the end, I think that the 33s will suit me professionally... and perhaps personally... with the same level of quality that HP always has.


P.S. For those of you who are taking the PE exams and are stressed about NCEES's new calculator policies, don't worry. I took the Civil PE, and later the Structural SE exams with just my 32sII (and scored the highest in the state both times). You don't need anything more... trust me... you just need to study.


Stan --

I agree with everything you posted about the 32Sii and 48G -- I have both.

I also keep a few simple, handy programs and equations in permanent residence in my 32Sii, and sometimes find a lack of RAM to run an advanced function.

The 32Sii should have had at least 1kB of RAM, with permanent allocation for all 27 registers (A-Z, and indirect) using 216 bytes. That would have left plenty for other purposes. It was a bad idea to deallocate space for a register storing a value of zero. Just because a variable contains a neutral value doesn't necessarily mean that it is unused.

I may eventually buy a 33S for the same reasons you mentioned. However, I will not compete with those who are desperately trying to get one for the April exams, and I'd rather not patronize Wal*Mart if I can avoid it.


-- Karl S.


I agree. walmart is already getting too much of our paycheck---

I'll wait until a real calculator purveyor has them to buy one.

Side-by-side with my trusty 32s

now which one is more beautiful?


I quickly entered the following program:

A0001  LBL A
A0002 1
A0006 +
A0007 +
A9999 +

B0001 LBL B
B0002 +
B0003 +
B0421 +
B0422 RTN

It rapidly (~1 minute) counts up to 10,415 .

Lengths & checksums:

LBL  A   LN=30009   CK=97FA

LBL B LN=1266 CK=CC09

So, it seems that instructions take 3 bytes each, even such simple ones as "+" ? What about large numbers, say -1.23456789012E-123 ? Do they take 9 bytes ?

Not that it would be a problem at all, what with that much memory otherwise mostly unusable.

Best regards from V.

I inserted your suggested value into a small program and it used 15 bytes.

I haven't tested whether there are significant differences in the memory footprints of the various commands.

Hi Paul,

Thanks for checking it out. However, 15 bytes for a number as a program line seems incredibly memory-hog. For comparison, the HP-71B uses 7.5 bytes per number (i.e. exactly one half) and as few as 3 bytes if the number happens to be an integer from -99,999 to 99,999, IIRC.

Well, at leat this will help use somewhat those tons of sadly nearly unusable RAM.

Best regards from V.

Hi Valentin!

Or, a byte is worth a lot less now than it used to be worth. The old timers will say, "I remember when a byte could buy you a dinner; now it won't get you a cup of coffee."


Thanks to Paul for this great posting thread.

I gree with Bill. Thank you Paul for acting as a 33S "guinea pig". I definitely appreciate you taking the time to share your first impressions of this long awaited machine. I also appreciate you taking the time to answer everyone's (mine included) questions. This discussion thread has led me to NOT cancel my Pre-order of two of them.


I'm obviously having fun, talking about a favorite subject. This time I simply happen to be in a position to share. (Wouldn't it be great to be able to do this for a living?)

Thanks to all of you who participate!

(Now, to get back to work!)

I was rewriting a silly old program, using it as an example of how to program the 33s to minimize impact on its most limited programming resources: labels and variables. (For example, I inlined the subroutine four times, replacing four XEQ's to conserve the subroutine's label.)

In the process I discovered that, on the 33s: 3 bytes are used for each typical program statement, but 15 bytes are used for any program entry of a numeric value.

My first ported rewrite included nine separate invocations of 0, 1 & 10 (for shifting digits individually out of number strings, etc.) and took 300 bytes.

I then changed the code to "generate" the numbers (CLx = 0, then 10^x gives 1, and another 10^x gives 10), store them, and then take 'em from registers rather than re-entering each time. (It turns out that I was using the registers anyway, so it didn't have a negative impact in that respect.)

This approach took one more instruction (-3 bytes) but saved 9 * 12 = 108 bytes, for a net savings of over 1/3 in the program's size: 105 bytes.

The general question becomes, in the interest of conserving memory, how many (and which) different integer values may be generated on a 33s in four or fewer instructions without keying actual numbers? (Note: five instructions at three bytes each doesn't save anything over the original 15 bytes, hence the four-instruction limit.) To make things interesting, I'm trying not to use any more of the stack than the x register.

A few examples:

  CLx 10^x        = 1
CLx 10^x e^x IP = 2;
PI IP = 3;
PI IP X! = 6;
PI X! IP = 7;
PI IP X^2 = 9;
PI ->L IP = 11;

(CLx is an iffy case, as it destroys the x register content, requiring an ENTER if that's to be preserved.)

PI isn't the only constant available -- I haven't walked through the CONST list to see what's there yet. (In any case, I'd probably have trouble entering the symbols into a post here.)

You don't need a 33s to work on this. Knowing 33s memory is used at 3 bytes/instruction and 15 bytes per number entered is the key.

Of course, this is all surely moot on the relatively memory-rich 33s, but is nevertheless an interesting exercise for the obsessive memory maximizer.

Edited: 23 Feb 2004, 11:24 p.m.


For what it's worth' to avoid using CLx:

PI cos 1/x IP = 1;

PI 1/x 10^x IP = 2;


PI 1/x ln IP = -1;

PI sin 1/x IP = 18;

PI e^x IP = 23;

PI 10^x IP = 1385;

PI 1/x %CHG IP = -127;

I don't remember ever needing the numbers 18, 23,1385 or -127 in the body of a program before, but what the H***.

I'm sure there are many others, but I have to get back to work now.

Take Care.


A non-destructive alternative to CLx 10^x for obtaining "1.0" is: PI SGN . They're two commands each; the former is useful in a situation where getting rid of the x register content is desired, the latter when it should be pushed.

I think I came across a 3-command sequence to generate "2.0", but can't remember it right now. (I did try a few of the CONST values -- those and the units conversions should provide more than a few options.)

CLSigma Sigma+ replaces the x register content with 1.00, but stomps on 6 storage registers in the process . . .

Last night, I needed 27 as a base for indirect access to the statistical registers, so PI IP X^3 saved six bytes. (Yes, this is indeed silly!)

I too should get to work . . .

Hi, Paul:

Here's a couple of missing entries in the 1..10 range:

 PI E^X SQRT INT  ->  4
PI X^3 SQRT INT -> 5

I'll leave you the pleasure of finding out the last
remaining entry, 8 :-)

Best regards from V.

There's a CONST value symbolized by a bold "R": 8.314472000.

So, R IP = 8. (And follow with the cube root to obtain two.)


Do the conversion factors each use 3 bytes as well? If so, then...

PI ->HR ->cm IP = 8;

PI ->HR x^2 IP = 10;

etc, etc.

I'm still doing this exercise with a 32sii so I only have PI as a starting constant. My 33s should arrive tomorrow.

The unit conversion functions (if each only uses 3 bytes)would certainly open up more possibilities for this "pointless" exercise.

Memory conservation in programming seems to be a dying concern today what with Gigabyte computers and Megabyte handheld devices; but my FORTRAN instructor my freshman year of college used to deduct a letter grade if he could find a way to reduce the size of the homework programs we submitted, so I have always made an effort to conserve memory in any programming I do.

Take care.


I haven't found anything other than number values that doesn't use exactly 3 bytes of memory.

I imagine that, given a deadline and ~32K to play with, they chose the simplest implementation possible -- three-byte codes for every statement, and full-length numerics for numbers. Fixed allocation of the named variables and consistent-length program entries must have greatly simplified memory management.

The 32sII keeps track of half-bytes, but that's not surprising with only ~390 bytes available.

Memory conservation in programming seems to be a dying concern today what with Gigabyte computers and Megabyte handheld devices; but my FORTRAN instructor my freshman year of college used to deduct a letter grade if he could find a way to reduce the size of the homework programs we submitted, so I have always made an effort to conserve memory in any programming I do.

My FORTRAN instructor was the same way. When writing a program I may have spaghetti code all over the place, but once the program is working, I go back and streamline all the code. I have found that I can add extra things to the program to make it more helpful and still reduce the code by a third and sometimes by half of what it started out as.

Now if Microsoft would go back and get rid of all the spaghetti code in Windows, it might fit within 50MB instead of 500MB and be much more stable and secure.

GRAD; assume you can live with this for a while...

CLX; get 0

ACOS; get 100

x^2; get 10000

x^2; get 100000000

LOG; get 8

If the algorithms are as good as in the HP41 emulator I am using now, you can spare the final IP, as the result is exact.

But the byte count is greater than 15, and you still have to reset the angular mode...

Edited: 24 Feb 2004, 7:59 p.m.

(Assuming DEG angular mode)

PI; X=3.14

ATAN; X=72.34

SQRT; X=8.50

IP; X=8.00


Supposedly, 12 bytes, and just one stack level used (as in a normal number entry operation).

Obviously, I will not have a 33s for some time, so I'm unable to verify result (should be no surprises) and memory usage.

My idea here is to come up with sequences that will work regardless of MODES settings. So most trig functions are out, because one has to take the trig mode into account when utilizing the functions.

CLx COS seems always returns 1.00, regardless of the trig mode. Beyond that, I'm not sure that trig is useful in this context.

A couple of somewhat less-than-elegant methods to obtain eight are:

PI   x^3   ->gal  IP
CLx ->degF ->gal IP
These, of course, rely on 33s units conversions that are not generally available.

Overall, I've subjectively evaluated approaches against an implicit "hierarchy of elegance":

  approaches that use "pure math"
approaches that start with PI
approaches that utilize units conversions
approaches that start with some CONST value

One might ask, why is PI granted special status among constants? I don't have an answer -- maybe it's because it's so universally available on calculators?

Anyhow, thanks for participating in a little diversion. (I vaguely remember a similar topic coming up on this Forum in the not-too-distant past -- if anyone has a link, please share it!)

Not having a 33s on hand, I missed the x^3 key...

Me too.


Of course, this is all surely moot on the relatively memory-rich 33s, but is nevertheless an interesting exercise for the obsessive memory maximizer.

I remember the type all too well. We synthetic programmers _thrived_ on saving one byte (e.g., "E2" instead of "1E2" or "100") on the HP-41C.

Who ever said the Byte Grabber was a fool's toy?


Negatives: The keyboard is incredibly busy, and the chevron style may accentuate that impression. And the decimal points in the display are almost invisible -- it would have been nice (from my nearsighted point of view) to have utilized a digit position and put a big fat dot in the middle of the character, rather than the teenly little pixel lost between digits.

The keyboard actually is busy, I just noticed looking at your photograph next to a 32 that you have 1 row of keys more, 2 keys instead of 1 big ENTER plus the few sort of keys below the screen. Is this due to added functionality (const library, ALG mode, what more??) or does this help access to existing commands.

Also I support the big fat dot, like in existing 32s when I see your photograph.

Anyway I can't wait to retire my 32.


Remember to compare the 33s keyboard to the 32sII instead of the 32s. The keyboard isn't that much busier than the 32sII.

Just look at a HP-67 keyboard, it too is a little busy...

Part of the busy-ness of the keyboard are the redundant/silly functions. Unit conversions? Are they marketing this to the high-school crowd? Cube & cube root? Doesn't y^x cover that? etc, etc, etc...

It would also have helped if they had 2nd function labels printed on the keys like the 15C's "g" functions are.

Oh well, at least it's RPN...

Don't forget... This appears to be a replacement for the HP32SII and it HAD unit conversions so this will have them too.

Cube and cube root are new, but these are in fact to help with the high school crowd, IMO, that don't use graphing calculators.

Remember the complaints HP used to get when they didn't have an X root Y key? "Y^X can do that!" we'd reply, but eventually a root key showed up.

I have to agree with somebody else's observation that the design is supposed to make it look like a cellphone. The more I see Paul's pictures, the more I realize that it looks exactly like a cellphone. One of the Nokias I think. Given that mentality, we should probably be glad that it functions like an RPN HP at all. :)

When I manage to hook up with an 33s I can give my 15C a bit of a rest after a decade of service, so that it lasts longer.
My 11C mysteriously died after 7 years, and my 15C has a lame "3" key. The 33s can take over hazardous duty in the lab, while the 15C gets the position of honor on my desk. :)

The more I see Paul's pictures, the more I realize that it looks exactly like a
cellphone. One of the Nokias I think.
That's one of the major reasons it's so unappealing to me. I don't like cellphones in general, and I despise the Nokia phones in particular.

In 20 years, HP collectors will say "The HP33S was a great and beautiful calculator compare to this new HP-XX that looks like YYYYY 2024 cellphone model (or whatever will be "hot" at this time). I think for newer eye, the 33S will look nice and a HP-41 very ugly. Remember, when the HP-41 was made, this model also "look" like other hardware that were made at that time; the color scheme of electronic hardware in 70' and 80' and even "hardware fashion" were present in the HP-67, 15C or HP-41 so why not with the HP-33S. The goal is not to be beautiful but to sell it well to young and old people in order to proof that calculator market is not only to TI... This way, we will get the opportunity to have other new models from HP...

Remember, when the HP-41 was made, this model also "look" like other hardware that were made at that time; the color scheme of electronic hardware in 70' and 80' and even "hardware fashion" were present in the HP-67, 15C or HP-41 so why not with the HP-33S.

But the "fashions" to which machines like the 41C and 15C conformed were for technical instruments, not consumer electronics. I never saw an HP calculator designed to look like a transistor radio or a Princess phone. There was a clear distinction in appearance between electronic devices meant to appeal to everyday consumers and devices that only a scientist or engineer would need (or want). When you look at a 33S does it make you think of white lab coats and test tubes and spectroscopes and blackboards filled with arcane equations, or any of the other things associated with "Math geeks" or "Science nerds"? No, it looks like it belongs with designer clothing and matching purses and expensive watches. It's designed to appeal to people who just want to fit in with everyone else, people who don't want to be associated with the "Math geek" or "Science nerd" stereotype, and that's why I dislike it so much.

I just have to add a story here.

I'm taking a statistics class at a university. Last night, the prof had us do some calculating. Most of the class didn't have calculators at all. But a couple of people actually pulled out cell phones and started using them. (I guess they have a "calculator mode" built in.)

But I was ahead of them all with my trusty HP-25 and got my answers first. Knowing what I was doing helped, too.

I don't have a cell phone yet. I'm holding out for a rotary-dial model built into my shoe. :)

- Michael

I'm holding out for a rotary-dial model built into my shoe. :)

Too late. In the words of Maxwell Smart, "missed by this much." Or, would you believe a wristwatch phone à la Dick Tracy?


Could you check if SOLVE and Integral give the same result as on a 32 and if they as well are 20% faster.




I'm a Certified STOMT ("Shirt-Tail Optics Maintenance Technician") -- my glasses show evidence of it, and now my HP-33s does, too.

If I'd just got some of those optics handy wipes, or whatever, I'd have avoided grinding a bit of grit into the center of the display cover.

In this the 49G+ is definitely superior (or at least, more Bozo-resistant): it has a non-glare textured glass screen and withstands such abuse.

Anyway, words to the wise.

Fortunately, it's a currently manufactured item, and I can get another if so motivated. (And that is the most significant aspect underlying this thread: there is a decent RPN scientific keystroke programmable available once again!)

Edited: 20 Feb 2004, 11:06 a.m.




I agree and why pay $75?????

Hmmm. No photo and a 'new' ebay user, please post here if (oops, I mean when) you get it!

I wrote seller asking him to email me a picture of his HP33s. O rea;;y doubt it if I get any picture from him.

"We're sorry to say that your order has been delayed due to
a delay in our fulfillment process. Unfortunately, we
were unable to ship the item(s) you ordered below by the
date we originally gave you. We are working with our
suppliers to fill your order as soon as possible, but we
are currently unable to provide a revised shipment date
for your order."

Am I the only one REALLY expecting this calculator for the NCEES FE test?

from: http://www.acsm.net/policy15.html
"When NSPS President John Fenn and I visited with NCEES in December, we were shown a prototype of the calculator that Hewlett Packard has developed to comply with NCEES Exam Policy 15. I do not know if the HP-33 shown above as one that can be pre-ordered from Wal-Mart is the finished version of that prototype."

Now, why is it out of stock? What about the 39 series? I've heard that there was also a new model comming...does anybody know any other calculator similar to the 33s (not above $200 like retail prices of the 32s)?

Any suggestion?

I was expecting it for the FLS. However I am now banking my hopes on the Casio 115 ms plus. Its a fine machine.


Casio ms 115: $14? Is it the one?

If you want an NCEES-compliant RPN scientific calculator, then you currently have just two options: (1) wait for the 33S, or (2) get an older discontinued HP model (like the 32SII) on eBay or elsewhere, probably for $200-$300. Sorry.

My *guess* is that the 33S will become readily available in March. If so, then you would still have a few weeks to get comfortable with it before the April exams. Walmart.com apparently got one of the very first shipments, but it seems that they didn't get enough to fill all preorders yet.

The other alternative is to get used to a non-RPN scientific, which are readily available from Casio, TI, or even (ironically enough) HP. The HP-9G, 9S, and 30S are all inexpensive and NCEES-approved.

Just out of curiosity, [since I am not in a profession which is in any way related to the test], I was wondering if NCEES would find the Datexx DS-833 acceptable. It is, essentially, a more powerful version of the HP-9G. As far as I can tell if the HP-9G doesn't violate the policy, this model wouldn't either. It would probably be the most powerful non-RPN model currently available that would meet the NCEES requirements.

I picked up one of these at a CVS pharmacy and as of a week or so ago, they were heavily discounting them here in Phoenix. They are clearly from the same series as the HP-9G.
They have a rudimentary algebraic history stack and don't have the HP-9G bugs that I referenced in an earlier posting.

NCEES has two lists of "banned" and "approved" calculators, neither of which purports to be complete. The Datexx 833 is not on either list, so its status is ambiguous.

From what I can tell, the 833 probably is NCEES-compliant. NCEES doesn't like calculators that (1) can be used to store text strings, and (2) can output such text strings via serial or IR ports. Doesn't appear that the 833 is capable of either (1) or (2).

In practice, though, it would be risky to bring an 833 to an NCEES exam. Since it is not on either NCEES list, it would be up to the local exam proctors to determine if it was acceptable. Exam proctors are not normally calculator experts, and they might reject the 833 simply because it looks very much like a banned model. Practically all other graphing calculators have been banned, and the 833 obviously fits the general graphing calculator description. So it could be guilt by association.

Anyone planning to use an 833 in an NCEES exam would be well advised to get written confirmation of its acceptability from NCEES in advance.

The NCEES calculator policy has led to confusion regarding the acceptability of many models. NCEES will probably resolve this in the future by becoming even more restrictive: you will only be allowed to use calculators from a relatively short "approved" list. All other models will be banned, even if they are technically compliant, in order to make life simpler for the proctors.

So it is quite possible that many older HP models, and uncommon models like the Datexx, will in effect be banned in the future.

I have a 33s due to arrive tomorrow from Walmart. If there is someone here who really needs it for an exam, I would be willing to pass it along (unopened) at my cost and wait until they are more generally available. I have to take my Ph.D. candidacy exam soon, so I appreciate your predicament (unfortunately a calc won't help me in molecular biology).

I suspect however that Walmart will soon begin to ship the remainder of the backorders, since I actually ordered mine fairly recently.

If interested, email me at brunell at cdinc dot com.



Edited: 18 Feb 2004, 10:11 a.m.

I just got an e-mail from WalMart today saying that they shipped mine out. The UPS website verifies that it was shipped from Cirritos, CA today. I only pre-ordered mine last Saturday (14 Feb) so apparently they still have a few left. I'm sure they're filling their pre-orders in a FIFO manner which explains the delay. What makes me wonder just what's going on is the fact that every other seller that I've found have reported availablity anywhere from mid-March to 21 April. Maybe WalMart bought the entire first production run?

Virgil Taylor

'My I.Q. test came back negative'

I heard that WalMart has received the HP33s on a priority basis.

I just checked out my status at walmart and it appears that it has been shipped. Anyways, thank you very much for your offering!

Possibly Related Threads...
Thread Author Replies Views Last Post
  Flash Flood Warning: 9/16/2013 (One Week from HHC13) Eddie W. Shore 8 182 09-17-2013, 09:20 PM
Last Post: Craig Ruff
  HHC 2013: One Week To Go Eddie W. Shore 2 95 09-13-2013, 05:32 PM
Last Post: Craig Ruff
  Auction Scam of the Week? Frido Bohn 4 108 06-03-2013, 01:45 PM
Last Post: Mike Morrow
  SandMath routine of the week: Inverse Gamma Function Ángel Martin 39 640 03-24-2013, 08:19 AM
Last Post: peacecalc
  41CL Routine of the Week - DLD48 Ángel Martin 1 62 02-15-2013, 08:30 PM
Last Post: Michael Lopez
  41-CL Routine of the week: MMUSWAP Ángel Martin 3 106 11-09-2012, 01:39 PM
Last Post: Luiz C. Vieira (Brazil)
  41CL Routine of the week: "YEDIT" Ángel Martin 4 114 05-17-2012, 04:42 AM
Last Post: Ángel Martin
  [WP34S] Beginner's positive thoughts - 1 week in. Chris Tvergard 7 180 05-14-2012, 03:35 PM
Last Post: Walter B
  Re: [WP34S] Beginner's positive thoughts - 1 week in. Chris Tvergard 2 87 05-13-2012, 06:50 PM
Last Post: Chris Tvergard
  The week of repairing calculators David Ramsey 6 161 08-25-2011, 12:15 PM
Last Post: David Ramsey

Forum Jump: