Should I replace my TI-89 Swiss Army Knife on the table with the HP Prime?

[Tom Flatterhand]

I sporadically observe the development of the HP Prime and have read some of the tests that appeared in the beginning. There were also written about some quirks. Now I got an HP Prime as a birthday present and would like to know how HP-41C, HP-48 and TI-89 owners rate the HP Prime?

Have initial quirks been eliminated over the years?

How good is the CAS compared to the Derive built into the TI-89?

How have you come to terms with the three-part nature (home mode, CAS mode, apps)?

Is it true that the TI-Nspire™ CX CAS and the HP Prime have the same CAS (Giac/Xcas)?

Maybe someone of you knows about the mentioned calculators and may answer me? I would be very happy.

[rawi]

(07-25-2022 07:19 AM)Tom Flatterhand Wrote:  I sporadically observe the development of the HP Prime and have read some of the tests that appeared in the beginning. There were also written about some quirks. Now I got an HP Prime as a birthday present and would like to know how HP-41C, HP-48 and TI-89 owners rate the HP Prime?

Have initial quirks been eliminated over the years?

How good is the CAS compared to the Derive built into the TI-89?

How have you come to terms with the three-part nature (home mode, CAS mode, apps)?

Is it true that the TI-Nspire™ CX CAS and the HP Prime have the same CAS (Giac/Xcas)?

Maybe someone of you knows about the mentioned calculators and may answer me? i would be very happy.

Sorry, of course I wanted to add some hints.

I use the HP Prime, the TI 89 Titanium and the TI NSpire CX II CAS.

My impressions are:
- The Prime has by far the best hardware. Even very complicated functions are shown in an instant. The memory is huge. The screen the biggest of all three with a good resolution. And it is the only one with touch screen.
The shortcomings are:
The switch between CAS and numeric is annoying. TI does not need this.
The manual is the worst of all three. The index is a bad joke. It is really difficult to find the information needed and sometimes the information is wrong.
- The TI 89 Titanium has by far the worst hardware: Rather slow, small memory, black and white screen with rather poor resolution.
But: It is very reliable, I did not find any bugs.
The manual is the best of all three. All commands are explained in alphabetical order. To find them faster there is a grouped index of them. Rather complete additional index. And it is printed. The others only have pdf-manuals. So it is easy to find all relevant information.
- The TI Nspire CX II CAS: Hardware is between the TI 89 Titanium and the HP Prime, but much nearer to the Prime than to the TI 89 Titanium. It is fast, but slower than the HP Prime, it has a color screen with a good resolution. The memory is much bigger than the memory of the TI 89, but only about 40% of the memory of the prime. The software is reliable, the manual not as easy to use as the manual of the TI 89 Titanium, but it is OK.
The TI Nspire has in my view one big disadvantage: There is no way to download data from a PC to a matrix. In statistics very often the data are needed in a matrix. You can download to the spreadsheet App, but you cannot copy data from a spreadsheet to a matrix. With the TI 89 Titanium and the HP Prime this is no problem. So if you have data on a PC and you want to use them on a TI Nspire in a matrix the only way is typing them.

Concerning the PC applications that come with the calculators there are as well significant differences:
- TI 89 is worst. You can downlod and upload data, upload screens, but there is no complete replica of the software. There are special apps which emulate the TI 89 on a PC but they are not from TI.
- HP Prime is in the middle. It is a comlete replica of the software and you can upload and downlod data. The Pro-version comes with some additional costs.
- TI Nspire in my view is best (with the exception of downloading data) because it is the only one which has a mode with higher resolution. So the bigger screen of the PC can be used. And it is included in the purchase of the calculator without additional costs.

[ijabbott]

(07-25-2022 07:19 AM)Tom Flatterhand Wrote:  Is it true that the TI-Nspire™ CX CAS and the HP Prime have the same CAS (Giac/Xcas)?

No, the Nspire CAS is based on Derive. However there is an app called KhiCAS based on Giac/Xcas (and also written by Bernard Parisse) that can be installed on Nspire using Ndless.

[Nigel (UK)]

(07-25-2022 07:19 AM)Tom Flatterhand Wrote:  I sporadically observe the development of the HP Prime and have read some of the tests that appeared in the beginning. There were also written about some quirks. Now I got an HP Prime as a birthday present and would like to know how HP-41C, HP-48 and TI-89 owners rate the HP Prime?

Have initial quirks been eliminated over the years?

How good is the CAS compared to the Derive built into the TI-89?

How have you come to terms with the three-part nature (home mode, CAS mode, apps)?

Is it true that the TI-Nspire™ CX CAS and the HP Prime have the same CAS (Giac/Xcas)?

Maybe someone of you knows about the mentioned calculators and may answer me? i would be very happy.

I used to love my TI-89. I would still love it, if my eyesight allowed! I really like the dedicated X, Y, Z, and T keys, along with the simplicity, consistency, and reliability of the CAS. The TI documentation is also of a high standard.

However, I now prefer to use a HP Prime. It isn't just the clearer display (which is *far* better); it is that the calculator can do so much more. Graph plotting is superb (I particularly enjoy the advanced graphic app) and the touch screen is integrated well into almost everything. The CAS is certainly capable of more than the CAS on the TI-89. HP provided good basic documentation for the CAS; I have found that looking at the documentation for Xcas often reveals extra commands.

It isn't perfect. Since its introduction, bugs have been fixed but new features have been added, introducing more bugs. It isn't bug-free, and it looks now as though existing bugs may never be fixed (I hope I'm wrong about this). However, I think that most of the bugs are in commands used mainly in writing programs - I find few problems in my typical use.

There are annoyances. As an example, "Ans" (last answer) doesn't work as I would wish - if the last answer is x^2, then Ans | x=2 ought to return 4, but it returns x^2 as Ans isn't evaluated until after the substitution. (I mention this because I used to use "|" so much on the TI-89!) There are other (rare) situations in which commands appear to work, but the answer is wrong. I don't think that this ever happened with the TI-89, at least not for me.

For me, the benefits easily outweigh the problems. Try it. I think you'll enjoy it!

Nigel (UK)

[Tom Flatterhand]

@ijabbott: Thanks for this info!

@Nigel: This is exactly the information that interests me. Especially the trustworthiness of the calculator is important for me. Since I'm not a mathematician and I'm not good at math, I use the TI-89, or now the HP Prime, as my personal mathematician, which helps me, for example, to solve or check my children's math problems. But even when I am dealing with mathematical problems myself, I need a trustworthy instance to test my thoughts. It is important that I can trust the results of the calculator.

>It isn't bug-free, and it looks now as though existing bugs may never be fixed

Can you give me an example?

[ijabbott]

(07-25-2022 06:40 PM)Nigel (UK) Wrote:  I used to love my TI-89. I would still love it, if my eyesight allowed!

The Voyage 200 solves that problem, but could be annoying due to using different key sequences than the 89.

[Nigel (UK)]

(07-26-2022 05:36 AM)Tom Flatterhand Wrote:  @ijabbott: Thanks for this info!

@Nigel: This is exactly the information that interests me. Especially the trustworthiness of the calculator is important for me. Since I'm not a mathematician and I'm not good at math, I use the TI-89, or now the HP Prime, as my personal mathematician, which helps me, for example, to solve or check my children's math problems. But even when I am dealing with mathematical problems myself, I need a trustworthy instance to test my thoughts. It is important that I can trust the results of the calculator.

>It isn't bug-free, and it looks now as though existing bugs may never be fixed

Can you give me an example?

Here's a couple.

(1) In Home mode, try 10.1^600/10.1^599. The correct answer is 10.1, but the calculator returns 1. Both numerator and denominator overflow and evaluate to 9.99999999999e499. This caught me out a few times with expressions that included factorials of large numbers not throwing an error, but evaluating to an incorrect answer. I think others have noticed this too. In CAS mode there's no problem - both top and bottom evaluate to \(\infty\) and the result is undefined.

(2) Go to CAS mode, and assign 10 to x. solve(x*3/2=3, x) and solve(x*3/2=3., x) return {2} and {2.} respectively. No problem. (Decimal numbers are approximate, just as on the TI.)

Now try solve(x^(3/2)=3, x) and solve(x^(3/2)=3.,x). The first returns {\(9^{1/3}\)}, which is fine, but the second returns {0.94868...}, which is wrong. If x is purged first (purge(x)) then the correct answer is given. So it's sometimes necessary to purge x before solving, and sometimes not. But instead of an error an incorrect answer is silently given.

I'm using software version 20210505. My Windows emulator has version 20210609 and has the same problems; it says that no update is available, so I'm assuming that there is no fix out there.

Nigel (UK)

[Tom Flatterhand]

Hello Nigel,

I have software version 2.2.14603 (2021 12 02) installed.

Ok, I have followed all the examples and I also get the wrong results, just as you described. That's a pity, the TI-89 gets the right thing in both cases! Since i don't always know if the variable used has already been created and given a value, this is unfortunately very dangerous and a strong argument against using the HP Prime.

Thomas

[Tom Flatterhand]

@rawi: Thanks for the compact summary of the comparison of the 3 calculators!

[ijabbott]

(07-30-2022 05:24 PM)RobinhoodPandey Wrote:  HP Prime is far the best cas and many bugs have been fixed but ndless doesn't work on ti nspire cx2 cas,due to latest O.S but khi cas on ti npire is more powerful than on hp prime.

Ndless is supported on CX II, but not (currently) for the latest OS version 5.4. I'm sure that's just the usual cat-and-mouse game between TI and Ndless!

[Bill Triplett]

Quote:Since i don't always know if the variable used has already been created and given a value, this is unfortunately very dangerous and a strong argument against using the HP Prime.

I dearly love my HP Prime. I love and use the simulators on the PC, and on my Android phone. Using a smartphone with a big display and a stylus, it really is a surgical tool.

That said, there is no reason to skip keeping a TI on hand.

I noticed with the TI-92+, TI-89, and V200 it is possible to simply type in ln(1+i) and immediately we see symbolically that the answer is ln(2)/2 + (π/4)*i.

That funny character that looks like a cubist lowercase "n" in the text is the Greek letter pi.

With the Prime, I enter ln(1+i) and the result comes back ln(1+i). This same thing happens if I hit Simplify, Expand, Collect, or Factor.

The Prime refuses to acknowledge that this simple input can be converted into something else. Perhaps there is a Settings secret handshake that could cure this, but I gave up.

Perhaps someone here could show me a cure, other than switching on my backup system.

[parisse]

Try evalc(ln(1+i))

[toml_12953]

(07-25-2022 04:47 PM)ijabbott Wrote:  

(07-25-2022 07:19 AM)Tom Flatterhand Wrote:  Is it true that the TI-Nspire™ CX CAS and the HP Prime have the same CAS (Giac/Xcas)?

No, the Nspire CAS is based on Derive. However there is an app called KhiCAS based on Giac/Xcas (and also written by Bernard Parisse) that can be installed on Nspire using Ndless.

Ndless doesn't work on the latest TI OS version 5.4 (yet)

[jte]

(07-26-2022 11:40 AM)Nigel (UK) Wrote:  ⋮
(1) In Home mode, try 10.1^600/10.1^599. The correct answer is 10.1, but the calculator returns 1. Both numerator and denominator overflow and evaluate to 9.99999999999e499. This caught me out a few times with expressions that included factorials of large numbers not throwing an error, but evaluating to an incorrect answer. I think others have noticed this too. In CAS mode there’s no problem - both top and bottom evaluate to \(\infty\) and the result is undefined.
⋮

This has indeed been noticed. An enhancement request (to report numerical overflow) is in the bug tracker (ticket 52). Addressing this (once it gets near the “top of the work items”) will take some time as it involves careful review and adjustment of the numerical routines (to incorporate flags for numerical events, such as overflow, at a minimum). (And adding automated tests etc.)

Within the CAS numerical environment, overflow can lead to things like 10.1^306/10.1^307 being (silently) quite far from 1/10.1. (Or 10.1^300/10.1^400 being far from 1/10.1^100; A/B with A staying within the range of finite floating-point values and B bubbling over.)