About calculators in general [LWN.net]

About calculators in general

Posted Oct 5, 2017 5:43 UTC (Thu) by bjlucier (subscriber, #5281)
Parent article: The NumWorks graphing calculator

Once, when teaching introductory linear algebra, my students asked, of course, "Why do I have to learn this, my calculator can do it!"

This happened to be during the lecture on ill-posed systems, so I gave them a series of examples where there was no solution, multiple solutions, etc.

On each problem, the sole HP calculator in the classroom basically gave the right answer, and the TI calculators gave nonsense.

To which one of the TI-owning students exclaimed "It's not fair that he's allowed to use an HP calculator on the exam!"

Vel Kahan spent many years consulting for HP, from the very first HP calculator, helping them to get the right answers.

I hope that open-source calculators can replicate that success.

to post comments

About calculators in general

Posted Oct 5, 2017 16:14 UTC (Thu) by smitty_one_each (subscriber, #28989) [Link] (10 responses)

Inviting the question: can this thing be configured for RPN? Otherwise, it's dead to me.

About calculators in general

Posted Oct 6, 2017 7:06 UTC (Fri) by madhatter (subscriber, #4665) [Link] (2 responses)

Traction getting be to seem not does but, for asked been has It: https://github.com/numworks/epsilon/issues/35

Excuse me; I meant it has been asked for, but does not seem to be getting traction (sorry, that was childish of me).

About calculators in general

Posted Oct 6, 2017 9:46 UTC (Fri) by smitty_one_each (subscriber, #28989) [Link]

I saw what you did there immediately.

About calculators in general

Posted Oct 6, 2017 15:25 UTC (Fri) by nybble41 (subscriber, #55106) [Link]

Did you mean:

> it asked-for has-been it traction getting seems not but

Just reversing the order of the terms does not change an infix expression to RPN. However, if you diagram the sentence and then write out the resulting tree in postfix order you should get something approximating "RPN English". (Complicated, of course, by implied/missing clauses, conjugation, and tenses. No doubt a specialist in English grammar could produce a more precise translation.)

About calculators in general

Posted Oct 6, 2017 7:44 UTC (Fri) by jem (subscriber, #24231) [Link] (6 responses)

Even if the firmware could be updated to support RPN, the keyboard doesn't look very usable for RPN entry. The HP Prime might be a better choice if you want RPN, or the HP 35s.

About RPN...

Posted Oct 9, 2017 9:25 UTC (Mon) by Herve5 (guest, #115399) [Link] (5 responses)

Or, alas, just RPN Calculator on Android, which besides not having real keys is otherwise close to perfect.

If NumWorks did propose an RPN version (with associated keys) I would buy it immediately indeed...

Android RPN Calculator

Posted Oct 9, 2017 13:44 UTC (Mon) by rschroev (subscriber, #4164) [Link] (4 responses)

There are at least two apps in the Play Store by that name, and a boatload of apps with similar names. I'd like to see if I can get used to RPN; which app would you recommend?

Android RPN Calculator

Posted Oct 9, 2017 13:48 UTC (Mon) by corbet (editor, #1) [Link] (2 responses)

I have "RpnCalc" by Edward Falk. It's not like having my old HP 41C back, but it is a good emulation of the RPN calculator experience. I'd recommend it.

Android RPN Calculator

Posted Oct 9, 2017 14:08 UTC (Mon) by rschroev (subscriber, #4164) [Link]

Thanks, I'll give it a try

Android RPN Calculator

Posted Oct 10, 2017 8:59 UTC (Tue) by Herve5 (guest, #115399) [Link]

same for me...

Android RPN Calculator

Posted Oct 10, 2017 9:46 UTC (Tue) by jem (subscriber, #24231) [Link]

One alternative is Free42, which emulates an HP-42. Then there's an app version of the HP Prime, which was mentioned earlier.

About calculators in general - HP vs. TI

Posted Oct 8, 2017 13:17 UTC (Sun) by pr1268 (guest, #24648) [Link]

Interesting anecdote about the lone HP user in a sea of TI calculators... That being said, I feel a need to vent a little frustration with my new (less than 2 year-old) HP 50g:

My TI-89 Titanium's display gave up after 10 years, so I got the '50g from Amazon.com at a "fire sale" price (about $70 USD packaged with a 2GB SD card); I'm unsure why it's so expensive now.

Despite lacking RPN (which I prefer), the '89 was a superior calculator, IMO. For example:

The '89 has the Π (product) function. I have yet to find this on the '50g.
Whereas Mr. Yates' TI-84 took "over three seconds" to integrate the normal distribution for -5 to 5, my shiny new '50g takes 13 seconds to (numerically) evaluate the same. At least Mr. Yates' experience with the patched NumWorks device sounds promising.
The '89 correctly evaluated the infinite sum of 2^-2n from 1 to ∞ as 1/3, but the 50g gives "1/(e^2*ln(2)-1)" (also correct but unintuitive). I have to do LNCOLLECT (through a menu) just to get "1/(2²-1)" and then EVAL to (finally) get 1/3.
Supporting both algebraic and RPN entry totally complicates the '50g's manuals (three large PDF's totaling 1765 pages!).

In summary, I'm unimpressed with the HP 50g. I thought that HP made calculators (graphing or not) superior to all other brands, but not this time.

Still, I'm intrigued by Mr. Goyet's product. Good luck to him and his company. And thanks to Mr. Yates for this article.