cancel
Showing results for
Did you mean:
cancel
Showing results for
Did you mean:
cancel
Showing results for
Did you mean:

11-Garnet

## symbolic solver - possible bug.xmcd

I have a description of the problem in the file.  Basically I cannot interpret the symbolic solution.  I already know the answer but if I didn't the out put doesn't help much

6 REPLIES 6
3-Visitor
(To:Joeboy)

If this is the same problem as you've posted on the Collab, then you need to delete the constraint [CTRL}0, since it probably forces the symbolic processor to consider setting the derivative to zero.  If you delete the constraint, the symbolic processor produces the expected result

11-Garnet
(To:Joeboy)

Yes, that works but the user has to know to do that, which is my point.  The fact that I set the expression equal to zero does not change the derivative expression.  The solver should give a correct result or issue some sort of warning/error instructing the user what to do.  The current result is useless unless you already know the answer.

1-Newbie
(To:Joeboy)

You can differentiate an algebraic expression.  What you have written is an equality, a boolean expression.  That cannot be differentiated.  While you may think of the expression as an equation, Mathcad does not see it as an equation except is very special contexts (solve blocks), and will not distribute the differentiation over the equality.  Also note that mathematically, since the expression on the left has been identically equated to zero, its derivative is also properly zero.

11-Garnet
(To:Joeboy)

That is all well and good; however a less than expert user of Mathcad would not know that, which is my point. Furthermore if I change the RHS to a real number I get similarly useless output.  Finally, if I use the symbolics toolbar the expression is differentiated "correctly" regardless of the RHS value. The fact that the value of the derivative may be zero should have no bearing on the symbolic expression for the derivative.

19-Tanzanite
(To:Joeboy)

The problem is that while you may know what you want, the symbolic solver doesn't. It only knows what you give it. Unfortunately, computers are still rather dumb 🙂

11-Garnet
(To:Joeboy)

Ya know, its a good thing computers can't sue for libel or slander or we would all be in trouble.

Announcements
Check out the latest