▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
The HP 35s is finally announced! I've been using it for months and really like this new model.
On the HP 35s webpage found at www.hp.com/calculators, there are 55 learning modules for your download.
HPCC.org has graciously put up a link to my review of the HP 35s that appears in their special HP 35s Datafile issue which you can read/download on their website here:
http://www.hpcc.org/datafile/V26Special/the35s.pdf
Be sure to check out the review and all the other goodies in the special HP 35s issue of Datafile. Join now! :-)
Highlights of the HP 35s:
** Two index registers, regular variables I and J. Dedicated i register is gone.
** 801 indirect registers plus 32 other registers (A through Z and stat registers).
** Indirect registers dynamically allocated based on highest non-zero value containing indirect register.
** Each indirect register allocated uses 37 bytes of available program memory.
** Index values of 0 through 800 reference the indirect registers.
** Index values of -1 to -26 reference A through Z.
** Index values of -27 to -32 access stat registers.
** The index registers are no longer in the middle of the address space as on the 33s.
** 26 global labels with dynamic line number GTO and XEQ instructions.
** Transfer execution to any line number within any program.
** Add / delete a step and the GTO/XEQ instructions renumber themselves appropriately.
** They keep pointing where you want them to point.
** Programs "limited" to 999 lines per label now. A001 through A999, B001 through B999, etc.
** Since each program line uses at least 3 bytes, 1000 lines equals 3000 bytes. 26 labels x 3000 bytes is more than available memory.
** 20 subroutine levels now available.
** The stack can hold 4 full complex numbers.
** 2-D and 3-D vectors as a new data type.
** The stack can hold 4 vectors, 2-D or 3-D.
** No direct P->R or R->P conversions. Change display mode instead. Polar parts can be separated.
** Equations no longer limited to 255 characters.
** Ability for non-destructive editing of equations using left
cursor.
** New LOGIC functions work on base-N numbers. OR, NOR, NOT, AND, XOR, NAND
** Algebraic mode equation/stack "tricks" work in RPN mode providing very useful stack methods.
** A program line of REGTxREGZxREGYxREGX will find the product of all four stack registers and store the result in X. Only T would be lost.
** If you roll the stack down, the program line of REGTx2 will double the value previously in X without losing any stack levels.
** Learning module available allowing for 2400 real numbers to be stored/recalled in indirect registers using 3-D vectors. Stack is preserved by this program. Shows many techniques for creating, modifying values in vectors and the indirect registers.
** Datafile article available showing program for translated PPC ROM matrix utilities programs M1, M2, M3, M4, M5 along with BE and BX. All that is needed now to find the determinant of a 28x28 matrix or the inverse of an 18x18 matrix is the equivalent RRM style program. (Yes, I've been working on one but it isn't finished yet!)
** Algebraic mode now operates much more like an equation operating system than the HP 33s algebraic mode did.
** Key a long calculation in ALG mode and press ENTER. To go back and revise one part of that calculation, press the left cursor key, go back and make your change, then press ENTER again.
For more, read the review!
Gene
Edited: 12 July 2007, 5:43 p.m. after one or more responses were posted
▼
Posts: 23
Threads: 1
Joined: Jul 2007
Gene,
How about a comment on one of the most important issues of recent HP calcs:
The feel (response) of the keys.
Thanks,
Steve
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
The keys of the 35s look good, feel good and work good (ok, work well).
In the same way the HP50g keys work very well compared to the 49g+, the keys on the 35s do a good job too.
Posts: 349
Threads: 66
Joined: Apr 2007
Quote:
Gene,
How about a comment on one of the most important issues of recent HP calcs:
The feel (response) of the keys.
Thanks,
Steve
Having used a machine for a few months now, I can say that the keyboard feel is *very* good. No lost keystrokes have been experienced and one can be pretty confident that what was pressed is what one gets.
Jake Schwartz
Posts: 1,830
Threads: 113
Joined: Aug 2005
At HHC 2006, I told HP that I didn't want them to bring back the HP-42S. I wanted the machine that the 42S would have become, if they had continued development on RPN along those lines. I also said "I want to be surprised by the innovations in keystroke programming you come up with." Well, this machine isn't a 42S successor, but it certainly is a worthy successor to the HP32. And the unlimited line based GOTO and XEQ surprised me. It strikes me as a novel (for an RPN calculator) method of solving the HP33s deficiency in the number of labels. Prior HP calculators had line number GTO instructions, but didn't automatically renumber in the case of addition/deletion of lines A cynic might observe that we have now passed MBASIC and Appelesoft BASIC, since neither of those did automatic line numbering. But I appreciate the effective solution to a problem that this community identified as a serious shortcoming of the 33s.
But that's not the only surprising thing about the 35s. There are tons of extremely interesting features mentioned in Gene's post. Imaginary numbers/2d vectors occupying one stack position is great, but the addition of 3d vectors is outstanding. New logic possibilities will make my work easier when designing IP networks. "Stack tricks" sound intriguing. My bottom line is that HP has clearly delivered a product that innovates on top of the solid base of its heritage in scientific RPN calculators. Way to go, HP!
I want one now! 8)
Regards, Howard
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
Without the "dynamic" reunumbering, line number GTO and XEQ instructions would not be very helpful at all. The 12c uses line numbers for GTOs and it is very painful when you have to insert or delete a step. The 35s solution, while not as nice as unlimited labels :-) is a nice compromise, IMO.
▼
Posts: 1,830
Threads: 113
Joined: Aug 2005
Quote:
Without the "dynamic" reunumbering, line number GTO and XEQ instructions would not be very helpful at all.
I agree. The fact that it's there is the suprising thing.
Regards, Howard
Posts: 362
Threads: 30
Joined: Jul 2005
All this looks great, but the main reason (work in hex and conversions to DEC) I use a 33 or 32 over an RPL machine has gone!
What a pity.
I will get a 35 anyway.
Arnaud
▼
Posts: 521
Threads: 72
Joined: Jul 2007
Quote:
All this looks great, but the main reason (work in hex and conversions to DEC) I use a 33 or 32 over an RPL machine has gone!
What a pity.
I will get a 35 anyway.
Arnaud
Yeah, the base feature's surprisingly clumsy to use with this requirement. I kept thinking I was doing something wrong...
Posts: 4,587
Threads: 105
Joined: Jul 2005
Gene, after looking through your review article and some of HP's learning modules I agree with your conclusion. It's by far the most appealing RPN calculator emerged for 1.5 decades at least.
IMHO there is one thing, however, which ought to be improved *very* soon: the representation of theta in the dot matrix display is terrificly close to 8. There, the designer must have been *very* tired. A simple solution may be to make theta two dots smaller vertically.
The error in formula 2 of the formula solver learning module part 2 will be far easier to correct.
The unit conversions (though totally obsolete IMO as you know ;-) ) contain some surprising nice features. Programming technique sounds good. And we'll have simultaneous equations at least for up to 3 unknowns. Fine!
So I'll buy one asap. But I will not drop my 42s until HP will launch a successor for this calc, too, including all the features discussed in this forum several times. So please, HP, you know where to look...
Edited: 12 July 2007, 12:06 p.m.
Posts: 1,153
Threads: 94
Joined: Mar 2006
In your article, I see the missing function on the beta unit's keyboard is the "->gal" function.
But the reciprocal ("->guy") isn't there either.
;-)
(Seriously, it's simply amazing that H-P sees fit to produce calculators at all, let alone one as promising as this.)
Posts: 1,619
Threads: 147
Joined: May 2006
Perhaps I missed something, but is there any type of I/O?
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
No I/O.
Keep in mind that this is a revision of the 32s, 32s2 and 33s line, not the 42s line.
As such, there are operating system limitations as to what can be done.
Wish there WERE I/O, but there isn't.
▼
Posts: 124
Threads: 11
Joined: Jul 2007
Thats too bad. I consider the lack of backup capability a major flaw in the design. I can't imagine anyone doing serious work with a device than has no backup.
I can already invision the calc imbedded in the dry wall above my desk the first time I get a "memory lost" condition. I think I will save my time, money, and patience.
I learn the hard way, but I learn.
▼
Posts: 2,448
Threads: 90
Joined: Jul 2005
Lots of serious stuff has been done with the 15c and it has no back-up.
Really, no I/O is no big deal for calculating. If you want backup, use a computer.
:-)
Just having a new, well-made RPN with a readable keyboard, readable display, and yes, even the ENTER key in the right place, is every reason for me--even if it weren't programmable at all, I'd get one at this price.
▼
Posts: 1,830
Threads: 113
Joined: Aug 2005
Lack of I/O hasn't prevented me from buying one. I intend to buy one more, in fact. But any device I leave the results of serious work on ought to be able to save those results on external media. I'm guessing that the intersection of certification hassles, design complexity and production cost accounts for the lack of I/O on this machine. But I can't pretend I'm not disappointed in that. Over all, of course, I'm delighted.
Regards, Howard
Posts: 124
Threads: 11
Joined: Jul 2007
My first post about I/O sounds a little harsh because I am a little frustrated about the lack of I/O (actually backup), but all comments are intended as constructive critisism.
"Lots of serious stuff has been done with the 15c and it has no back-up."
Agreed,but you may be comparing apples and oranges. I am not an expert on the 15C but AFAIK it does not have a reset button or a reset procedure. The 33 (the 35s predessor) reset procedure is on page A4 and B2 of the user manual. I have had to use it twice so far. I don't imagine too many 15C users have had that experience. Otherwise, I am afraid that less serious stuff would have been produced.
These reset tools seem to exist almost universally across all the computing devices I have bought in the last 20 years regardless of the manufacturer. I accept them because I have to. To me it is a tacit admission by the manufacturer that 1 of 3 things is true
1. This device has become so complex that that we are unable to test all possible permutations. (I suspect this is the case)
2. We are unwilling to spend the finances to insure that all possible permuations have been tested. (This may be the case)
3. We know from first hand experience that this device will crash and this is the only way out. (I hope this is not the case)
"Really, no I/O is no big deal for calculating. If you want backup, use a computer.
:-)"
Its no big deal if you have a device as "bullet proof" and reliable as a 15C. The proliferation of the reset capabilty may imply that this type of manufacturing is no longer possible. I am willing to accept that (Do I have any choice?) if I can protect my intellectual property. I would argue that if it requires a reset capability then it requires a back up capability
"Just having a new, well-made RPN with a readable keyboard, readable display, and yes, even the ENTER key in the right place, is every reason for me--even if it weren't programmable at all, I'd get one at this price."
I think you reflect the opinon of just about every enthusiast on this forum. I want this calculator to be a success in the maket place but I find it puzzling that a programmable device has no back up or documentation capability.
▼
Posts: 1,619
Threads: 147
Joined: May 2006
The 15C does have a reset, however the memory is so limited that you can backup your 15C on paper (I have) and restore it during your lunch hour with time to spare for lunch.
▼
Posts: 124
Threads: 11
Joined: Jul 2007
I guess I am just getting ahead of the curve again. I am complaining about backing up the 35s with pencil and paper and neither myself nor probably anyone else has done it yet. Might take the better part of the afternoon. Anyone for a long lunch break.
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
Well, of course good programmers write and debug their programs on paper first...so you should always have a paper copy filed.
Big :-)
Seriously, I too wish for I/O. But I'll certainly take this much improved model!
Posts: 2,761
Threads: 100
Joined: Jul 2005
It's a pity apparently the trig bug hasn't been addressed thus far.
My HP-67 (1976) doesn't have this kind of bug, but it has the sin-1 and cos-1 bugs. However, they are well documented in the addendum to the Owner's Handbook (these would eventually be fixed). Can we expect something like that from today's HP?
Gerson.
Posts: 1,368
Threads: 212
Joined: Dec 2006
The use of 26 alpha labels as global labels and dynamically renumbered line-based GTO/XEQ within programs is not what I was hoping for (I really would've preferred the system in the 41 and 42S), but it is a hell of good compromise.
At first I thought 26 global labels weren't much for my purposes, but practically speaking I don't think I have anymore than a dozen or so user programs in my 41CX or 42S at any one time. Also, I am mindful that in the 33S my programming style gobbles up the local labels like crazy. Now, I can port those fast programs and use the line addressing.
Sounds like a real gem for only 60USD!!!
Les
Posts: 111
Threads: 22
Joined: Jul 2007
Whoo-hoo! The 35s is on the way...Many of it's characteristics and elements are competing for my attention. Particularly, the 3-space vector, 0.98 reversion to classic HP looks, and the new programming features (memory, architecture, and "tricks")
This is great news, I look forward to getting my article.
I'm likely to stalk the local Fry's to pick up my copy. Online purchases just don't have enough bang. Much more fun to court the calc on the shelf in person.
ECL
▼
Posts: 156
Threads: 19
Joined: Jun 2007
When Fry's has them in stock, please let us know!
Posts: 1,392
Threads: 142
Joined: Jun 2007
Gene, from the picture in your review, I don't see IP and FP. Did I miss them or are they in some menu?
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
The yellow shifted INTG menu has the functions IP and FP along with INTdivide, Remainder and INT.
▼
Posts: 1,392
Threads: 142
Joined: Jun 2007
Thanks Gene. I was thinking that the INTG key would probably have those functions, but couldn't tell without further documentation.
Posts: 4,587
Threads: 105
Joined: Jul 2005
Quote:
The yellow shifted INTG menu has the functions IP and FP along with INTdivide, Remainder and INT.
Gene, what does the last function of your list?
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
Greatest integer. Different from IntegerPart (IP).
Greatest Integer
▼
Posts: 406
Threads: 47
Joined: Jul 2005
Gene,
Does the 35s have a remainder function (RMDR) as does the 33s or the modulo (MOD) on the 42s? Or even the remainder function (RMD) of the 16C in integer mode?
tm
▼
Posts: 1,545
Threads: 168
Joined: Jul 2005
|