Arithmetic with standard forms is performed by the functions12
|null(f)||test f = 0|
|minusf(f)||test f < 0 (formally)|
|domainp(f)||test if f is a domain element (e.g. number)|
|addf(f1,f2)||f1 + f2|
|quotf(f1,f2)||f1∕f2 if divisible, else nil|
|quotfx(f1,f2)||f1∕f2 divisiblity assumend, no check|
As standard forms are a superset of domain elements these functions also can be used for arithmetic with domain elements or for combining standard forms with integer numbers. They can even be used to compute with integers as long as these are regarded as “degenerate” polynomials.
For access of the components of a standard form there are functions
|mvar(f)||extract leading kernel|
|ldeg(f)||extract leading degree|
|lc(f)||extract leading coefficient|
|red(f)||extract reductum f|
These functions can be applied only if the form is not a domain element. The following example presents a typical function evaluating a standard form; it looks for the maximum power of a specified kernel y in a standard form f:
This function has a double recursion which is typically for functions traversing standard forms: if the top node is not trivial (domainp) or not a power of y the procedure branches to the coefficient form and to the reductum form - both can contain references to the kernel y. Kernels are unique in the system and therefore the equality test is performed by eq.
For the initial construction of a standard form a safe and efficient way is to build an algebraic expression first and convert it using simp (see below) or to use the arithmetic functions above too. For elementary conversions use
|*k2f(k)||convert unique kernel k to SF|
|numr simp(a)||convert algebraic expression a to SF|
|prepf(f)||convert a SF f to an algebraic expression|
Note that integers are already an SF and so need not be converted.
Although the are infix macros .** , .* , .+ to construct standard forms, they should be avoided unless you are completely sure what you are doing: they construct without any consistency checks, and wrong structures may lead to inconsistent results; such cases are difficult to debug.