To initialize Guile, you can use one of several functions. The first,
scm_with_guile, is the most portable way to initialize Guile. It
will initialize Guile when necessary and then call a function that you
can specify. Multiple threads can call scm_with_guile
concurrently and it can also be called more than once in a given thread.
The global state of Guile will survive from one call of
scm_with_guile to the next. Your function is called from within
scm_with_guile since the garbage collector of Guile needs to know
where the stack of each thread is.
A second function, scm_init_guile, initializes Guile for the
current thread. When it returns, you can use the Guile API in the
current thread. This function employs some non-portable magic to learn
about stack bounds and might thus not be available on all platforms.
One common way to use Guile is to write a set of C functions which
perform some useful task, make them callable from Scheme, and then link
the program with Guile. This yields a Scheme interpreter just like
guile, but augmented with extra functions for some specific
application — a special-purpose scripting language.
In this situation, the application should probably process its
command-line arguments in the same manner as the stock Guile
interpreter. To make that straightforward, Guile provides the
scm_boot_guile and scm_shell function.