Gems 16-20 from the Red Book of Maple Magic

Gem 16 - Frenet Vectors and Assumptions

The Frenet-Serret frame of orthonormal vectors along a curve are immediately obtained with the VectorCalculus command TNBFrame. For example, consider the calculations in Table 16.1.

Note that the TNBFrame command returns unit tangent, principal normal, and binormal vectors. The TangentVector, PrincipalNormal and Binormal commands in this same package return unit vectors only with the addition of the normalized option. Table 16.2 illustrates this for an expanding helix.

In an effort to verify orthogonality, compute the dot product of T with N.

What has happened? Associated with each vector in the Frenet frame is its point of contact (root point) on the curve. To see this, apply the GetRootPoint command, but first, modify the interface to show the tilde associated with assumed variables. (The showassumed parameter can have the value 0, 1, or 2, depending on whether no annotation, the tilde, or a proviso is displayed.) See Table 16.3.

Ordinarily, when the assume command is used to make a (permanent) assumption on a variable, a following tilde is displayed to indicate that this variable is no longer the same as the variable before the assumption was made. However, when the assuming functionality is used to make a (temporary) assumption on a variable in the course of a calculation, the assumption is removed before the result is returned. (For this reason, I always use "assuming" and never use the assume command.)

When a VectorCalculus Vector is created under the shadow of an "assuming" clause, that vector has an associated module that encodes the properties of the ambient vector space. When assuming exists, assumptions on terms inside this module cannot be "cleaned up" prior to output. Hence, both N and B have assumptions attached to t in their root points, but T does not, and the dot product fails because the root points are seen as different.

Consequently, for N or B, separate the calculation from the simplification, as shown in Table 16.4.

The principal normal now has a root point free of the entangling assumption, so the calculation  =  succeeds.

Gem 17 -  and  as Laplace Transform Operators

In some texts, the Laplace transform of the function  is denoted with a decorated  such as  or . Some texts use upper and lower case letters to distinguish between function and transform, with no standardization on which case denotes what. I have always used  for the transform because it is simplest to type.

Some texts use some form of script "L" as the transform operator. In Maple, the choice is between  and , neither one of which matches the character used in Ian Sneddon's The Use of Integral Transforms (McGraw-Hill, 1772). In some recent writing, I've used  (from the Script palette) for the transform operator, and  for the inverse. Table 17.1 shows how to set these symbols as operators in math mode.

Working in "exposed" math notation, the only challenge is remembering that  must be seen by Maple as a single symbol; hence, the need for setting it as an Atomic Identifier. But what about working in text mode, as when programming, or when putting the alias command behind an embedded component? Without using Maple's math mode, it is not immediately apparent how to implement the Atomic Identifier construct, or even set the symbol  from the Script palette. Table 17.2 summarizes how to discover the text-mode code for the requisite symbols.

For , simply enter it from the Script palette; for , set it as an Atomic Identifier. Then enter these strings into the alias command, as shown below. Of course, copy/paste avoids the tedium of reproducing these character strings.

alias(`ℒ` = inttrans[laplace], `#msup(mi("ℒ"),mrow(mo("&uminus0;"),mn("1")))` = inttrans[invlaplace]):

Gem 18 - Setting an Iterated Integral from the Expression Palette

At the moment, no palette in Maple has a template for setting a definite iterated integral. Constructing the iterated integral  by iterating the single definite integral can be expedited by recently discovered "magic."

When the upper- and lower-limit fields in the outer integral have been filled, and the tab key is pressed to move to the integrand field, another click on the definite-integral template will set this template. But it will autoselect the integrand field of this new integral. This requires backing up through the fields by pressing the Shift key while tabbing. I have always found this to be needless busywork, and because of it, have repeatedly asked for double- and triple-integral  templates for iterated definite integrals.

However, if after arriving at the integrand field in the first (outer) integral template, the Backspace key is pressed before the next integral template is inserted, the new template will arrive with the lower-limit field autoselected. This is because there is a hidden field-flag that is set to 3 when the integrand field of the outer integral is selected. When a new definite-integral template is inserted, it inherits the flag value of 3, and the new integrand field is autoselected. The intervening Backspace resets the flag to 1, so the new integral template arrives with the first (lower-limit) field autoselected.

I think I would still like to see double- and triple-integral templates added to the Expression palette, but meanwhile, my complaints about it aren't as intense as they used to be.

Gem 19 - Writing a Slider Value to a Label

Table 19.1 contains as Embedded Components, a Slider and a Label, with the slider value being written as the label's caption. The leading "" and the equal sign between the slider and the label are simply typed characters.

The code behind the Edit field of the Slider is Do(%Label0(caption) = sprintf("%.5f", Do(%Slider0)));

The sprintf command determines how the displayed value is to be formatted. The string "%.5f" indicates five places after the decimal. The value to be written is read from the Slider with the second Do command.

Gem 20 - Writing Text to a Math Container

Table 20.1 contains a simple configuration of embedded components, namely, two MathContainers, with a Button between. A function written into the left-hand container will have its derivative written to the right-hand container when the button is pressed. If no function is entered on the left, but the Derivative button is pressed anyway, the message "Please enter a function to be differentiated" is written to the MathContainer on the right.

This Gem shows how to write the text message to the MathContainer on the right.

If the Button is naively coded with Do(%MathContainer1 = diff(Do(%MathContainer0),x), then pressing the Derivative button with no entry on the left generates the error message shown in Figure 20.1.

Table 20.2 contains "defensive" code that catches the empty left-hand container.

An alternative to MathML[Export]("...") is "...".