Is a spare 35S a good idea or redundant?

[Matt Agajanian]

Hello all.

I was considering the possibility of purchasing a spare HP-35S. Since I have a 35S I bought a couple of years ago, I'm hoping HP has addresses the firmware glitches since my earlier purchase.

If the programming, calculating, and accuracy hiccups have been cured, I might just buy one. So, has HP smoothed out the bumps?

Thanks

[Maximilian Hohmann]

(03-18-2017 09:24 PM)Matt Agajanian Wrote:  So, has HP smoothed out the bumps?

No. But I never understood why this would be important. These are very special bugs which only occur in very specific circumstances that most users will never encounter. The worst thing that can happen is losing one's memory content. But then I would never input much data and programs into any device without a data connection anyway. I have a couple of these calculators for the sake of collecting and for me the worst thing about them is the short battery life. Not the bugs.

[Matt Agajanian]

(03-18-2017 10:07 PM)Maximilian Hohmann Wrote:  

(03-18-2017 09:24 PM)Matt Agajanian Wrote:  So, has HP smoothed out the bumps?

No. But I never understood why this would be important. These are very special bugs which only occur in very specific circumstances that most users will never encounter. The worst thing that can happen is losing one's memory content. But then I would never input much data and programs into any device without a data connection anyway. I have a couple of these calculators for the sake of collecting and for me the worst thing about them is the short battery life. Not the bugs.

Thanks for this perspective. I feel the same way because my appreciation for calculators goes in a different direction and I appreciate the mathematics, language/programming, and its beauty but, with the overt disappointment that some fervent engineers,scientists, and other mathematicians show about the 35S/33S bugs and flaws, it sounds as if these models are disastrous and monstrosities.

Thanks!

[rprosperi]

(03-18-2017 10:07 PM)Maximilian Hohmann Wrote:  

(03-18-2017 09:24 PM)Matt Agajanian Wrote:  So, has HP smoothed out the bumps?

These are very special bugs which only occur in very specific circumstances that most users will never encounter.

For most of the bugs, I agree, however the machines inability to generate a consistent checksum for an entered program means you can never be sure you've entered one correctly, even if just copying from an existing list. I can write a program and publish it here with the size and checksum, then you enter it into your machine and you see different checksum, even if entered 100% correctly. For me, if I don't have confidence in a machine, I won't use, and if your advice is just don't use it for programs, there are cheaper and better machines available.

[Matt Agajanian]

Yes, that is and has been quite a peculiar problem. Makes me wonder how and why the checksum algorithm is flawed in and of itself.

[Didier Lachieze]

(03-19-2017 01:26 AM)Matt Agajanian Wrote:  Makes me wonder how and why the checksum algorithm is flawed in and of itself.

HP 35s Checksum explained.

[Gerald H]

Concerning battery life, today

2017-03-19

I put in fresh batteries in a 35s, the exhausted batteries having been inserted in

2015-05-??.

The calculator has been well-used in the 22 months the batteries lasted.

Concerning speed, the 42S takes about double the time to run programmes compared with the 35s.

I have myself written more programmes than fit into the 42S whereas the 35s memory is practically empty (or infinite, as you prefer).

Keyboard on 35s definitely dodgy, on some still performing well after 7 years use.

[Logan]

Yeah, I have had keyboard problems too, after only three months of use and when going fast, sometimes it would skip entries. I really have to take my time and go back and press and check (never an issue with either of my 42s calculators). I also had random issues with vectors from time to time, which was important for me especially during the PE exam. I found the programming a bit restrictive because you only get 26 labels, etc. Once I finished the exam I immediately ditched it and went to the 42s. I do have a 32SII that I haven't sat down with yet.

[cjmcc]

(03-18-2017 11:03 PM)rprosperi Wrote:  ... For me, if I don't have confidence in a machine, I won't use, and if your advice is just don't use it for programs, there are cheaper and better machines available.

Which machines would you suggest instead for someone wanting the best possible HP/RPN experience at a decent price?

[Dave Britten]

(03-20-2017 12:14 PM)cjmcc Wrote:  

(03-18-2017 11:03 PM)rprosperi Wrote:  ... For me, if I don't have confidence in a machine, I won't use, and if your advice is just don't use it for programs, there are cheaper and better machines available.

Which machines would you suggest instead for someone wanting the best possible HP/RPN experience at a decent price?

If you're okay with (or prefer) RPL, you can get a used 48S pretty cheap.

[Paul Dale]

(03-20-2017 12:14 PM)cjmcc Wrote:  Which machines would you suggest instead for someone wanting the best possible HP/RPN experience at a decent price?

The WP 34S has the most features but the keyboard can be variable.
The HP 42S has a great keyboard and a more than adequate set of features but can be expensive.
The HP 32Sii has a great keyboard, decent features and is usually reasonably priced but it has limited memory.


Pauli

[rprosperi]

(03-20-2017 12:14 PM)cjmcc Wrote:  

(03-18-2017 11:03 PM)rprosperi Wrote:  ... For me, if I don't have confidence in a machine, I won't use, and if your advice is just don't use it for programs, there are cheaper and better machines available.

Which machines would you suggest instead for someone wanting the best possible HP/RPN experience at a decent price?

Hmmm.... "Decent" is a relative thing. If you mean in the range of a new 35S, about $65, there are not too many choices.

RPN: For scientific programmable machines on a limited budget, a used 32S is very good. A 32SII is better (if you need/like the enhancements such as fractions) but will always cost more. A 42S can't be beat, but will cost a lot, though the SwissMicros DM42, due out in a couple months, will do all a 42 can do, much faster, with a much larger LCD and with FAR higher precision and should cost in the area of $200.

RPL: If you are willing to invest the time to learn RPL, you can get a new 50g for less than $75, and though a 50g is far superior in every way, a 48GX still feels better, and retains more of the traditional HP 'way'. If you go for a 48GX, I highly recommend you find one with black LCD (which will cost up to $50 more) as it is MUCH easier on the eyes than the original, lower contrast blue. Also, a 48GX will seem quite slow once you've used a 50g, but it's less noticeable if you don't use 50g or Prime, etc. which respond virtually instantly for almost all operations.

Some folks prefer the color scheme and kbd layout of the 48SX over the 48GX, but this is (IMHO) due to those folks having learned the 48SX before trying the GX. There were no 48SX machines made with black LCDs but you can transplant a black LCD (e.g. from a 39G, which often sell for $15) into them. 48SX with black LCD is one of the nicest looking machines, again in my opinion.

Lastly, if you don't absolutely need scientific functions, then I'd recommend an HP-17BII, which you can often get for $20-30. This is a great RPN machine, and while it does not provide keystroke programming, it does include the HP Solve application, which is very flexible, and easy to use for equation solving. Don't get either of the 17BII+ versions as the solver in them isn't reliable.

[Dieter]

(03-18-2017 11:03 PM)rprosperi Wrote:  For most of the bugs, I agree, however the machines inability to generate a consistent checksum for an entered program means you can never be sure you've entered one correctly, even if just copying from an existing list. I can write a program and publish it here with the size and checksum, then you enter it into your machine and you see different checksum, even if entered 100% correctly.

To quote Maximilian Hohmann: "I never understood why this would be important". ;-)

Honestly: is this a real world problem? Do you have dozens of long printed program listings that have to be entered? And finally: what do you do with well-respected calculators like the 67, 41 or 42? They do not have a checksum feature at all! Are they useless if a program has to be entered manually? Are you mistrusting them as well because there is no checksum to compare? Come on...

Yes, the 35s has its weak points. But IMHO none of them is so important that it would spoil the picture of a well-made easy-to-use everyday RPN calculator. I have been using mine since 2007, and I really like it.

Dieter

[Dave Britten]

(03-21-2017 07:27 AM)Dieter Wrote:  Yes, the 35s has its weak points. But IMHO none of them is so important that it would spoil the picture of a well-made easy-to-use everyday RPN calculator.

Until you accidentally program an infinite loop and lose everything in memory. The first time that happened to me was also the last, as it was immediately relegated to the collection drawer. It's a shame, because the keyboard feels great. (Of course, working with different number bases kind of sucks.)

[Gerald H]

(03-21-2017 07:27 AM)Dieter Wrote:  

(03-18-2017 11:03 PM)rprosperi Wrote:  For most of the bugs, I agree, however the machines inability to generate a consistent checksum for an entered program means you can never be sure you've entered one correctly, even if just copying from an existing list. I can write a program and publish it here with the size and checksum, then you enter it into your machine and you see different checksum, even if entered 100% correctly.

To quote Maximilian Hohmann: "I never understood why this would be important". ;-)

Honestly: is this a real world problem? Do you have dozens of long printed program listings that have to be entered? And finally: what do you do with well-respected calculators like the 67, 41 or 42? They do not have a checksum feature at all! Are they useless if a program has to be entered manually? Are you mistrusting them as well because there is no checksum to compare? Come on...

Yes, the 35s has its weak points. But IMHO none of them is so important that it would spoil the picture of a well-made easy-to-use everyday RPN calculator. I have been using mine since 2007, and I really like it.

Dieter

However, the 42S does show the size of the programme, which is some sort of check on correctness.

[rprosperi]

(03-21-2017 07:27 AM)Dieter Wrote:  Honestly: is this a real world problem? Do you have dozens of long printed program listings that have to be entered? And finally: what do you do with well-respected calculators like the 67, 41 or 42? They do not have a checksum feature at all! Are they useless if a program has to be entered manually? Are you mistrusting them as well because there is no checksum to compare? Come on...

This is a fair comment and reasonable question which deserves a reply. The 67, 41 and 42 came from a known and verified team, with well established credentials for testing, quality and attention to detail. The 35S, though branded HP, from an engineering and quality point of view comes from some 2nd rate bulk manufacturer, with clearly limited budgets, skills and/or staff allocated for testing.

My point, which Dieter has correctly pointed-out that I did not make very clear, is that if a feature, whose sole purpose is to verify the integrity and correctness of a program, is itself fundamentally flawed, this tells us a lot about the overall quality of the design and/or testing that went into the machine.

I agree the 35S is an OK machine to use for simple daily non-programming use, but if you need to use advanced features or extensive programming, other machines are likely to provide you with more confidence and trusted results.

Like any tool, if I don't trust it, I don't want to use it when the work matters. YMMV.