Post Reply 
Bugs in Solve command.
03-04-2017, 05:39 AM
Post: #1
Bugs in Solve command.
Hello, I recently bought my HP prime and I am super happy with it, I think it is an extraordinary and complete and super fast machine. However it still has many bugs that I hope the following updates will be corrected. I have found the following bugs with the solve command:

1- The when I introduce solve(5x^2-2>=2x) and the CAS settings so: Exact and use radical with check and the options Complex and use i with not check the answer is
x>= 1/5 * (1+sqrt(11) and dont show the other answer x<= 1/5 * (1-sqrt(11).
However, if I adjust CAS settings with the check on Complex, then the answer is
x>= 1/5 * (1+sqrt(11) and 10>x. The first answer is incomplete and the second answer is incorrect.

2-With solve((x-a)*e^x=-x*(x-a),x) the calculator the HP Prime show undef, but in the other calculator of the competence, the nspire CX CAS the answer is e^x+x=0 or x=a in the exact mode, and x=a or x=-0.567143 in the approximate mode.
Find all posts by this user
Quote this message in a reply
03-05-2017, 07:31 AM
Post: #2
RE: Bugs in Solve command.
With default settings, I get what is expected list[x<=((-(sqrt(11))+1)/5),x>=((sqrt(11)+1)/5)] and list[-0.56714329041,a]
Find all posts by this user
Quote this message in a reply
03-06-2017, 03:21 AM
Post: #3
RE: Bugs in Solve command.
Hi, thanks for your reply Parisse. With the first equation yes it is correct now, after deleting the contents that had the variable "x" with the purge (x) command several problems were corrected. The first equation does give me both correct answers with the default settings. However with the second equation gives me an error message that says "Unable to isolate x in -x-e ^ x", how did you get the answers x = [- 0.56714329041, a]? With what CAS settings did you get the correct answer? Thank you very much.
Find all posts by this user
Quote this message in a reply
03-06-2017, 06:53 AM
Post: #4
RE: Bugs in Solve command.
exact, use sqrt, principal checked
complex, use i unckeched.
Find all posts by this user
Quote this message in a reply
03-06-2017, 10:12 AM
Post: #5
RE: Bugs in Solve command.
I suspect you came with an older version and should update. Install the connkit, run it, and then plug in there.

TW

Although I work for HP, the views and opinions I post here are my own.
Find all posts by this user
Quote this message in a reply
03-07-2017, 12:31 AM (This post was last modified: 03-07-2017 01:09 AM by math7.)
Post: #6
RE: Bugs in Solve command.
Not yet, the HP Prime still does not give the correct result, it's strange because with the first solve equation and the correct settings it gives the answer well, but the second equation keeps on appearing the "Unable to isolate x in x-e^x" message. And I have the 10637 version of the software on the calculator, I think it's the latest update. What can be the problem?

Settings:

Exact: NO check Complex: check
Radical: check use i: NO check
principal: check

But I have tried other combinations and nothing.
In the emulator I have the same result, the message "Unable to isolate...."
Find all posts by this user
Quote this message in a reply
03-07-2017, 05:24 AM (This post was last modified: 03-07-2017 05:26 AM by math7.)
Post: #7
RE: Bugs in Solve command.
It happens something strange, now with the emulator after using the command purge (x, a) and run the solve command with the second equation shows what the screenshots say.

It's frustrating because on the physical calculator I do the same thing and it does not give me the answer, I get the message "unable to isolate x in ......"

The answer given by the emulator is not the same as the one obtained by Mr. Parisse, I do not understand what happens or if it is a CAS error


Attached File(s) Thumbnail(s)
       
Find all posts by this user
Quote this message in a reply
03-07-2017, 05:28 AM (This post was last modified: 03-07-2017 05:30 AM by Han.)
Post: #8
RE: Bugs in Solve command.
You are using the Equation Library, which sets up its own variables. Press the [View] key and list the variables currently in use. If you see your x and a listed, then that is the reason you are not getting the desired results (i.e. variable conflict). You can also check using the AVars command.

Just type: DelAVars(AVars)

This deletes all the variables in the app (it is safe as the app will rebuild them when needed). Then try your solving again.

Graph 3D | QPI | SolveSys
Find all posts by this user
Quote this message in a reply
03-07-2017, 05:45 AM
Post: #9
RE: Bugs in Solve command.
Thanks Han, but it does not work yet. In the captures below it shows what comes out, I did as you said with the Delavars command (AVars) and then run the same command and now gives no response, it is very strange. I then enter the application function and exited it but when I run the same issue comes the same error message, I feel frustrated.


Attached File(s) Thumbnail(s)
           
Find all posts by this user
Quote this message in a reply
03-07-2017, 05:57 AM
Post: #10
RE: Bugs in Solve command.
(03-07-2017 05:45 AM)math7 Wrote:  Thanks Han, but it does not work yet. In the captures below it shows what comes out, I did as you said with the Delavars command (AVars) and then run the same command and now gives no response, it is very strange. I then enter the application function and exited it but when I run the same issue comes the same error message, I feel frustrated.

Have you used other programs that have modified your variables?

First, type x and a to see if there are numerical values stored in either one. If you see numerical values, then that means they have been created somewhere.

HVars; (this will list the current Home variables)
AVars; (this will list the variables in the current app)

restart; (this will clear out all existing CAS variables)
DelHVars(HVars); (this will delete all Home variables)
DelAVars(AVars); (this will delete all app variables within the current app)

The last thing to check is whether you have an programs that EXPORT the variables a and/or x. Again, you will know based on what typing a or typing x puts onto the screen.

Graph 3D | QPI | SolveSys
Find all posts by this user
Quote this message in a reply
03-07-2017, 06:20 AM
Post: #11
RE: Bugs in Solve command.
Nothing happens yet, I attach the screenshots. As you can see the values of x and a are null. There is also something strange about solving the equation with the variables "u" and "v" since the command solve the manual says that it is useful for any variable and you can see that with these variables the answer is an empty vector, I do not understand why either. But with the "x" and "y" variables, the error message appears. I'll try restoring with ON + Symb key. Could you tell me if you are given the correct answer to which CAS or HOME settings?


Attached File(s) Thumbnail(s)
       
Find all posts by this user
Quote this message in a reply
03-07-2017, 06:39 AM
Post: #12
RE: Bugs in Solve command.
I do this and nothing


Attached File(s) Thumbnail(s)
       
Find all posts by this user
Quote this message in a reply
03-07-2017, 11:26 AM
Post: #13
RE: Bugs in Solve command.
Mine provides correct results, so you perhaps could try to make a backup first, then erase everything (reset the device) and then try your equation again. If it works, you know its about settings and/or variable assignment, I usually do things like that with function running, Home settings: Complex checked, Cas: simplifx minimum, Checked: Exact, Use sqrt, and principal.
Arno
Find all posts by this user
Quote this message in a reply
03-07-2017, 01:31 PM
Post: #14
RE: Bugs in Solve command.
Curious, I get the same warning as math7 does "Unable to isolate...", in my physical calculator, with last version and x and a unassigned.
Find all posts by this user
Quote this message in a reply
03-07-2017, 02:49 PM
Post: #15
RE: Bugs in Solve command.
(03-07-2017 01:31 PM)akmon Wrote:  Curious, I get the same warning as math7 does "Unable to isolate...", in my physical calculator, with last version and x and a unassigned.

Yes, I get that warning, too, but after "Enter" it provides the correct result.
Arno
Find all posts by this user
Quote this message in a reply
03-07-2017, 03:48 PM (This post was last modified: 03-07-2017 03:54 PM by Han.)
Post: #16
RE: Bugs in Solve command.
I think we are getting differing results because we are using different versions. The current firmware seems to produce an error whereas newer firmware will produce the desired solutions.

Graph 3D | QPI | SolveSys
Find all posts by this user
Quote this message in a reply
03-07-2017, 04:58 PM
Post: #17
RE: Bugs in Solve command.
(03-07-2017 03:48 PM)Han Wrote:  I think we are getting differing results because we are using different versions. The current firmware seems to produce an error whereas newer firmware will produce the desired solutions.

Well, I´ve reseted the virtual calculator, to clear everything that disturbs the result, but the result is always the same, unable to isolate. Version 11226. Guys, how could you solve this with your virtual/ physical unit?
Find all posts by this user
Quote this message in a reply
03-07-2017, 05:59 PM
Post: #18
RE: Bugs in Solve command.
Yes to me it is still the same, I have tried different configurations, I hope that soon as Han says the update solve these problems, however the Prime virtual version 11226 gives me the wrong results. In a while I will delete all the applications that I have installed to the calculator and reset all the variables, to check if it is some application that causes the problem. And if that does not work, I do not know what else to do. But I think this should be resolved soon for the good of all users, because it is not nice to work insecure with the calculator and not know if the results you give are correct.
And I would love all those involved in the development of Prime especially CAS, like Tim Wessman, Parisse, and others to make it super reliable and more complete than the competition, because I have been a HP user for many years, more Or less since 1993 and I want to feel safe when using it when I make calculations.
Find all posts by this user
Quote this message in a reply
03-07-2017, 07:47 PM
Post: #19
RE: Bugs in Solve command.
(03-07-2017 05:59 PM)math7 Wrote:  Yes to me it is still the same, I have tried different configurations, I hope that soon as Han says the update solve these problems, however the Prime virtual version 11226 gives me the wrong results. In a while I will delete all the applications that I have installed to the calculator and reset all the variables, to check if it is some application that causes the problem. And if that does not work, I do not know what else to do. But I think this should be resolved soon for the good of all users, because it is not nice to work insecure with the calculator and not know if the results you give are correct.
And I would love all those involved in the development of Prime especially CAS, like Tim Wessman, Parisse, and others to make it super reliable and more complete than the competition, because I have been a HP user for many years, more Or less since 1993 and I want to feel safe when using it when I make calculations.

I do not think this is necessary. As mentioned in my earlier post, the current releases (both the hardware firmware and the slightly more recent virtual calculator) do in fact produce the error mentioned in your first post. I can confirm this on my own calculators and on my Rev. 11226 virtual calculator. Others have also mentioned replicating the error with a new/reset instance of the virtual calculator. Parisse's results suggest that the problem has been fixed in whatever current version is available (to the developers).

Graph 3D | QPI | SolveSys
Find all posts by this user
Quote this message in a reply
03-07-2017, 08:41 PM
Post: #20
RE: Bugs in Solve command.
I can confirm that I see:

solve((x-a)*e^x=-x*(x-a),x) => { -.5671..., a } in the latest development version (not any released one).

So there must have been a change there. Sorry for the confusion generated here.

TW

Although I work for HP, the views and opinions I post here are my own.
Find all posts by this user
Quote this message in a reply
Post Reply 




User(s) browsing this thread: 1 Guest(s)