Producing symbolic derivative on 48G  Printable Version + HP Forums (https://archived.hpcalc.org/museumforum) + Forum: HP Museum Forums (https://archived.hpcalc.org/museumforum/forum1.html) + Forum: Old HP Forum Archives (https://archived.hpcalc.org/museumforum/forum2.html) + Thread: Producing symbolic derivative on 48G (/thread138561.html) 
Producing symbolic derivative on 48G  Roger King  07012008 I have a question about how to force the 48G to produce a symbolic derivative. Say I type in an expression like x/y, then y, and hit the derivative button. If there are no variables for x and y, the result is a symbolic solution, i.e x/y^2, which is what I want. However if there are variables for either x or y the numeric values will be inserted into the resulting expression. I am looking for a way to keep the symbolic result without using the derivative menu. By contrast on the 50G the result seems to be always be symbolic.
Thanks in advance for any help.
Re: Producing symbolic derivative on 48G  Allen  07012008 you can turn off the flag to automatically evaluate expressions. e.g clear flag (3). See Appendix D in the User manual.
Re: Producing symbolic derivative on 48G  Roger King  07022008 I did try setting and clearing flags 1 to 3 and 59 (the ones that looked like they might have some affect), but they do not seem to change the result: when evaluating the derivative on the 48G, the result will use the numeric values of stored variables.
Likewise on the 50G these flags seem to have no affect: the result is symbolic.
Re: Producing symbolic derivative on 48G  BruceH  07022008 Change into a different subdirectory which doesn't have those variables set?
Re: Producing symbolic derivative on 48G  Roger King  07022008 This sort of works as long as I can switch to another directory that does not have the variable in it, or above it since the calculator searches the current directory, and the directories above it for the variable. To include a bit more of the situation, the problem arose in a program I wrote for the 50G for propagation of errors. Actually everything works great on the 50G. Since the program was in user RPL I copied it over to a 48G and it seemed to work ok until I noticed some strange results, which turned out to be the problem at hand. There are some basic ways to work around the problem, such as being really careful about what variables are stored in the various directories, but I am looking for a bit cleaner solution that does not depend on the state of the calculator.
