▼
Posts: 177
Threads: 12
Joined: Jan 1970
All, I have found another difference between the HP 33s and HP 32sii. While using the seed and random functions I noticed that the two handle a seed of zero differently. With a seed of zero the 32sii consistently generates 0.03136 but the 33s actually produces a random number between 0 and 1.
Example:
32sii: 0 enter, seed, random: 0.03136, 0.03136, 0.03136, etc.
33s: 0 enter, seed, random: 0.86939, 0.74239, 0.34987, etc.
Consulting the 33s User's Manual I found the following on page 4-15:
"The RANDOM function uses a seed to generate a random number. Each random number generated becomes the seed for the next random number. Therefore, a sequence of random numbers can be repeated by starting with the same seed. You can store a new seed with the SEED function. If memory is cleared, the seed is reset to zero."
Up to this point the description of the seed and random functions in the 32sii and 33s User's Manuals are identical. But then the 33s adds the following sentence:
"A seed of zero will result in the calculator generating its own seed."
I am not sure how many users will be affected by this difference but I thought it to be important to post this to the forum. The important issue is that when using the 33s (and any calculator) the user should carefully read the manual to make sure that he/she understands how each function is designed to be used and what are the limits of the function, etc. This is especially important if you are a current or former 32sii user as the 33s is not turning out to be a 1 for 1 replacement. The good news is that this particular difference is documented in the manual although it would have been nice to have had it highlighted.
Regards,
John
▼
Posts: 5
Threads: 0
Joined: Jan 1970
As much as I love my 32SII, the 33S is fast catching up in that department. What you've just reported and the fact that it has a built-in constants library and a little extra wiggle room (raam?) in the PRGM mode has raised my esteem of that model. Good discovery, John. But to be a wiseguy, I don't think it's really too possible to have a true random number generator on any electronic device, save having a dedicated sensor port to catch gamma ray bursts or neutrino fluxes from space.(.. and even then, I wonder!)
▼
Posts: 1,193
Threads: 43
Joined: Jul 2005
There are devices based on the thermal noise which appears on any electronic device. Their output are as random as it gets.
On a simpler approach, if you have a free-running counter inside your device, and the counter is not closely tied to the processor, and if you just take a sample with such a frequency as to allow for many counter cycles between samples, you can consider it random enough.
Pseudorandom sequences can be good enough or very bad, and anyone interested on them should read Donald Knuth's "The Art of Computer Programming" Vol 2, "Seminumerical Algorithms"
Just my 0.02$
▼
Posts: 5
Threads: 0
Joined: Jan 1970
Thermal noise! I should have known that! What could be easier, as it is, as you said, an accompanying, natural facet of anything, not just electronic devices.
Posts: 172
Threads: 13
Joined: Jul 2005
In many cases, you don't want a true random number generator. A PRNG with a user-settable seed allows for reproducible results, which is useful when testing software.
Posts: 177
Threads: 12
Joined: Jan 1970
If HP will rework the LCD and clean up some of the functions the 33s will definitely go from acceptable to good in my opinion.
Regards,
John
▼
Posts: 1,792
Threads: 62
Joined: Jan 2005
John stated,
Quote:
If HP will rework the LCD and clean up some of the functions the 33s will definitely go from acceptable to good in my opinion.
My opinion is, it'll take more than that. My diatribe on this matter is in the recent Forum Archives.
-- KS
Edited: 10 Oct 2004, 4:05 a.m.
▼
Posts: 177
Threads: 12
Joined: Jan 1970
Hi Karl. Implementing the observations and suggestions in your note would then take the 33s from good to great.
Regards,
John
▼
Posts: 801
Threads: 36
Joined: Nov 2005
Forgive the cynicism, but if (or when) the 33S does have these features implemented, thereby making it great, we won't be able to buy it anymore for $50-65 USD.
But, if I recally, the 32S and 32SII were both around $70 USD (I got mine during a special at MicroCenter for $60) and the 20S was also about $60-70... the 34C was $125-250, if memory serves!
Again, we all have different task needs/pain tolerance levels, but I think that for its current price, the 33S is actually a steal! (Consider, after all we've said [in the negative sense], it is at least better in mechanical quality and basic features than the $10 drugstore scientific or the $7 "$0.99 store" scientific calc... which are only algebraic!)
▼
Posts: 1
Threads: 0
Joined: Jan 1970
"Forgive the cynicism, but if (or when) the 33S does have these features implemented, thereby making it great, we won't be able to buy it anymore for $50-65 USD"
@#?$ Why? Almost all the feature improvements suggested are software-only, and as it is now the HP-33S actually has far more RAM than one could realistically ever use, so reducing that and getting a larger ROM instead could keep the manufacturing price the same.
You do bring up a good point that the HP-33S is cheaper than the older machines... mine was $50, whereas I seem to recall my 32SII was $65 -- which I did consider a little pricey at the time. Still, I'd happily pay up to about $100 for a REALLY GOOD small programmable scientific calculator.
Has any one else noticed that the 33S is the only non-graphing programmable calculator left on the market today? Pretty sad, IMO.
---Joel Kolstad
▼
Posts: 801
Threads: 36
Joined: Nov 2005
"... Has any one else noticed that the 33S is the only non-graphing programmable calculator left on the market today? Pretty sad, IMO.
---Joel Kolstad"
Thanks for your comments, Joel. To your unquoted comments first- you're right, as I was not aware until just a couple of days ago by reading in this forum that these improvements are software or firmware modifications.
And as to your quoted comment, I agree, if it's true; doesn't TI still make a programmable non-graphing scientific... or Casio? I haven't checked into their wares for a while now, so I don't really know.
Anyhow, I too, personally find the lighter mass and smaller form factor of a scientific to be MUCH more convenient to realistically use on a day to day basis and the "added bonus" (actually, in my opinion often a necessity) of powerful programmability makes calculator of this kind a joy to use.
Posts: 57
Threads: 3
Joined: Feb 2007
Well, Ed Look says TI doesn't have a non-graphic programmable, Casio at least has one the fx-3950p (3650p solar) but it changes from country to country. There are places where it is not sold or maybe it has a different name.
▼
Posts: 801
Threads: 36
Joined: Nov 2005
Oh, no, Julian!
I merely said I did not know whether TI made one. The last ones I remember are the TI-57 or 59 models. Since then, I've really only been interested in HP (RPN) models.
I'm sure it varies from country to country, according to demand. But, it is good for these companies to all have a little competition... good for us!
Posts: 363
Threads: 22
Joined: Jul 2007
<< Has any one else noticed that the 33S is the only non-graphing programmable calculator left on the market today? Pretty sad, IMO. >>
The harsh realities of the programmable calculator market in the U.S. in the 21st Century are as follows:
1. The average professional today does not perceive any need for a programmable calculator (either graphing or non-graphing). He has a PC on his desktop at work, another PC on his desktop at home, and a laptop for use on the road. He probably does keep a calculator in a desk drawer, but only for quick and simple calculations; PCs are used for all serious number crunching. The spreadsheet has replaced the programmable calculator in the professional market, just as the calculator replaced the slide rule.
2. In contrast, the average high school student does not have a PC on his desktop at school. So the programmable calculator still rules in the educational market. This market is large, lucrative, and almost completely owned by TI (Casio is also a player, HP is trying to be one). However, students want graphing calculators: they expect their calculators to have menus, to display mathematical formulas, to do symbolic math, to plot graphs, and to play games. A non-graphing programmable, like the 41C or 42S, is as obsolete as a DOS computer as far as the education market is concerned.
3. There is only one remaining niche for non-graphing programmable calculators: NCEES licensing exams. At present, PCs, PDAs, and graphing calculators are not allowed on the Engineer-in-Training exam, the Professional Engineering exam, or the corresponding exams for Land Surveyors. The 33S is the most powerful permissible model; it is therefore quite popular among NCEES exam candidates. Some vendors are actually marketing 33S exam software (is anyone else selling calculator software these days?). However, the NCEES exam market for the 33S is trivial in size compared to the high school market for graphing calculators.
Edited: 25 Oct 2004, 1:04 a.m.
Posts: 673
Threads: 20
Joined: Oct 2008
Yes, it is a sad state of calculator affairs. The responses above are very accurate of what corporate America and others perceive the market to be like and therefore do not bother to develop a powerful pocket calculator. Instead, Hp invested in an older Hp32sii design that would not violate the NCEES exam policies that were going to be in effect when Hp released the Hp33s. Any additional feature (ANY!!!) would have gotten the Hp33s banned from said exam. Therefore the Hp33s was released as you know it. Not a great calculator by any means, but the best that can be sold and used in the NCEES exam.
Are there better pocket calculators?? The Hp17Bii+ is probably better, if YOU DON'T neeed Trig (nearly every one of us does! ). If you can get hold of an Hp42s (and $$$ isn't a big deal), you will get a significant step up in power and features. But you won't get real I/O that an Hp41c series offers. Hp could have released an Hp42sx, but did not want to rob from their precious 28/48s line.
Sadly, the corporate leaders feel that PC's are what you should be using, if you need more than a four function calculator (and there is some truth in this for tracking and accountability). But pure R&D or napkin engineering suffers from this as, I sometimes don't have my PC or I am not going to drag this anchor around with me. I hate to drag out my Hp48 as I feel it is to big and bulky.
RANT OFF.
▼
Posts: 363
Threads: 22
Joined: Jul 2007
NCEES has never clearly explained how they determine which calculators are "acceptable" and which are "banned". But they recently made an interesting move: they put a genuine graphing calculator, the TI-81, on the "acceptable" list.
You can see it on their list here.
Why is the TI-81 OK? It is an old discontinued model, but the manual and other details are available on the Web. It seems to have rudimentary text handling capability; it can store long text strings for display in programs. The key difference between the TI-81 and the "banned" models seems to be that the TI-81 has absolutely no means of I/O. No IR port, no cable port.
The acceptability of the TI-81 implies that NCEES might accept more advanced calculators than the HP-33S, if I/O was crippled. For example, a remodeled 41C or 42S might be OK, as long as the text handling was primitive and there was no IR or cable port.
There would actually be a small, but real, niche for such a device among NCEES exam candidates. On the other hand, I don't think it would sell at all in the education market, which is by far the most important for calculator sales.
Edited: 25 Oct 2004, 2:04 p.m.
|