Next: Implementation notes.
Up: Introduction.
Previous: Warnings.
Erable and alg48.
Erable is partially derived from the alg48 package. The
arithmetic functions of Erable are derived from those of alg48.
Erable and alg48 have some other common features like
simplifications, partial fraction
expansion or rational integration. The main differences are:
Erable is most of the time slower (about 2 times I would say,
this is only a mean, sometimes alg48 is 3 times faster than
Erable, sometimes Erable is 1 to 2 faster than alg48).
Why?
Because Erable handles complexes and square roots natively:
e.g. Erable simplifies expressions like
,
,
.
Hence arithmetic operations in Erable must be generic, that's
why they are slower.
On the other hand, Erable treats matrices as global
objects for simplifications, as alg48 simplifies matrices
element by element, hence Erable may be faster if matrices are
involved.
Erable accepts strings embedded in symbolics, this
means that if you EXPAnd
(5x+12)16 with Erable
you'll get the exact answer. You may also do arithmetic operations
on strings representing integers (e.g. try
"123456789123456789" DUP MULT)
Erable handles numerical data (reals which are not integers
and complexes with non integral real and/or imaginary parts)
Erable has a partial implementation of the Risch
integration algorithm: it handles most of the common integrals, not
only rational fractions.
alg48 (version 4) implements the complete factorization
algorithm over the rationals, Erable finds only first order
factors of the square-free factorization and then, for 1-variable polynomials,
calls the numeric solver if necessary and tries to rebuild 2nd order
factor. Hence, alg48 factors the expanded form of
(x4+x3+1)(x4+x+1) but not x4+1 and Erable does not
factor the first example but factors the second one.
- The main specific feature of
alg48 is the Gröbner base
computation. The main specifics features of Erable are
eigenvalues and eigenvectors of matrices, differential
equations (first order: linear, Bernouilli, homogenous; linear
with constant coefficients), limits and Taylor series,
quadratic forms, permutations, variable substitutions.
If you have enough memory, do like me: keep both on your calc and
choose the right instruction!
Next: Implementation notes.
Up: Introduction.
Previous: Warnings.
Bernard Parisse
1998-07-31