RPN Mode: Possible to use advanced matrix operations?

[johan162]

Since I'm used to the older HP calculators I prefer the RPN in the normal calc (non-CAS).

However I can't seem to use some of the common matrix operations like eigenvalues etc. since that always gives me "Syntax Error". Switching to non-RPN mode they (of course) work just fine.

Is it really not possible to use those kind of "advanced" (really?) operations in RPN mode?

(That would make the RPN mode severely crippled and inconsistent since it is possible to, for example, compute the determinant in RPN mode - hopefully I've just missed something.)

[toml_12953]

(05-05-2017 07:45 AM)johan162 Wrote:  Since I'm used to the older HP calculators I prefer the RPN in the normal calc (non-CAS).

However I can't seem to use some of the common matrix operations like eigenvalues etc. since that always gives me "Syntax Error". Switching to non-RPN mode they (of course) work just fine.

Is it really not possible to use those kind of "advanced" (really?) operations in RPN mode?

(That would make the RPN mode severely crippled and inconsistent since it is possible to, for example, compute the determinant in RPN mode - hopefully I've just missed something.)

Put the matrix on the stack, then use DET(1)

The 1 in the parens tells the function how many parameters are to be used. All RPN functions work that way. You put the arguments on the stack, then tell the function how many arguments to use. You wouldn't believe how long it took me to remember that! The people here had to remind me several times.

Tom L

[johan162]

Thanks for your reply but I'm quite aware of RPN and stack, thats what I have been using for the last 30 years:-)

Calculating the determinant is one of the operations that actually happen to work (and is not in the second level menu) work as expected (i.e. put the mat on the stack and execute the operation). However, My question was regarding what HP consider advanced operations such as finding the Eigenvalues/Eigenvectors

If you put the matrix on the stack and try to execute, say, "EIGENVAL()" (menu 7->6->1) then that does not operate on the stack. However, if you try to copy the matrix from the stack as an operand to "EIGENVAL()" or use a matrix varible all you get is "Syntax error" _as long as you are in RPN mode_

So, as long as you are in the "Home" calculator view and have it set to RPN-mode I cannot for my life get things like the eigenvalues to work. Of course if you switch back to "Algebraic" (or Textbook) then operation like "EIGENVAL()" works as expected.

My guess is that since these operations are implemented in the CAS environment (even if you can use them from the Home) that will never work in the RPN mode.

Unfortunately this really makes the RPN mode very limited if that only supports a limited set of operation that on the surface seems inconsistent. For example, why should calculating the determinant work but not finding eigenvalues? I of course appreciate the difference in complexity in implementing these operations but from a high level point of view there is no real difference - just an inconsistency.

In the long run this really makes it necessary to have the calc in algebraic/textbook mode to avoid inconsistencies -ar at least as long as you expect to do heavy matrix calculations (which I happen to do a lot..)

[John Colvin]

(05-07-2017 02:37 PM)johan162 Wrote:  Thanks for your reply but I'm quite aware of RPN and stack, thats what I have been using for the last 30 years:-)

Calculating the determinant is one of the operations that actually happen to work (and is not in the second level menu) work as expected (i.e. put the mat on the stack and execute the operation). However, My question was regarding what HP consider advanced operations such as finding the Eigenvalues/Eigenvectors

If you put the matrix on the stack and try to execute, say, "EIGENVAL()" (menu 7->6->1) then that does not operate on the stack. However, if you try to copy the matrix from the stack as an operand to "EIGENVAL()" or use a matrix varible all you get is "Syntax error" _as long as you are in RPN mode_

So, as long as you are in the "Home" calculator view and have it set to RPN-mode I cannot for my life get things like the eigenvalues to work. Of course if you switch back to "Algebraic" (or Textbook) then operation like "EIGENVAL()" works as expected.

My guess is that since these operations are implemented in the CAS environment (even if you can use them from the Home) that will never work in the RPN mode.

Unfortunately this really makes the RPN mode very limited if that only supports a limited set of operation that on the surface seems inconsistent. For example, why should calculating the determinant work but not finding eigenvalues? I of course appreciate the difference in complexity in implementing these operations but from a high level point of view there is no real difference - just an inconsistency.

In the long run this really makes it necessary to have the calc in algebraic/textbook mode to avoid inconsistencies -ar at least as long as you expect to do heavy matrix calculations (which I happen to do a lot..)

If you do 7->6->1 and then type 1 and ENTER, you get the eigenvalues.

[toml_12953]

(05-07-2017 02:37 PM)johan162 Wrote:  Thanks for your reply but I'm quite aware of RPN and stack, thats what I have been using for the last 30 years:-)

Calculating the determinant is one of the operations that actually happen to work (and is not in the second level menu) work as expected (i.e. put the mat on the stack and execute the operation). However, My question was regarding what HP consider advanced operations such as finding the Eigenvalues/Eigenvectors

If you put the matrix on the stack and try to execute, say, "EIGENVAL()" (menu 7->6->1) then that does not operate on the stack. However, if you try to copy the matrix from the stack as an operand to "EIGENVAL()" or use a matrix varible all you get is "Syntax error" _as long as you are in RPN mode_

So, as long as you are in the "Home" calculator view and have it set to RPN-mode I cannot for my life get things like the eigenvalues to work. Of course if you switch back to "Algebraic" (or Textbook) then operation like "EIGENVAL()" works as expected.

My guess is that since these operations are implemented in the CAS environment (even if you can use them from the Home) that will never work in the RPN mode.

Unfortunately this really makes the RPN mode very limited if that only supports a limited set of operation that on the surface seems inconsistent. For example, why should calculating the determinant work but not finding eigenvalues? I of course appreciate the difference in complexity in implementing these operations but from a high level point of view there is no real difference - just an inconsistency.

In the long run this really makes it necessary to have the calc in algebraic/textbook mode to avoid inconsistencies -ar at least as long as you expect to do heavy matrix calculations (which I happen to do a lot..)

EIGENVAL(1) works for me when the matrix is on the stack. It works the same as DET(), or any function.

Tom L

[johan162]

Thanks!

I hadn't realized you could enter stack level as an argument for the CAS functions

Now if we could just get the firmware updated so that all functions with a single argument (or perhaps even two) would just automatically use the top of the stack we could save a key-stroke or two ...

Thanks again!

[Tim Wessman]

(05-08-2017 06:00 AM)johan162 Wrote:  Now if we could just get the firmware updated so that all functions with a single argument (or perhaps even two) would just automatically use the top of the stack we could save a key-stroke or two ...

The problem is that any CAS commands can take any number of arguments. Most of the matrix commands are actually from the CAS which is why those show there. Anything that is not a CAS command will behave as you describe.

In the CAS, "sin" is a valid input, as is sin(), or sin(1,2,3,4,5).