Document the current module system?

[NIIBE Yutaka]

Martin Grabmueller wrote:
 > I have just skimmed over the Reference Manual chapter about Guile
 > modules.  There are a lot of fixes and updates necessary.  Since I
 > suppose that Godot will not make it into the next release, I would
 > like to document it better (at least remove the errors), but I'd like
 > to avoid unnecessary work.  What do you think?

I think that it's very good idea.

 > Can someone point me to
 > a mail archive or documents which describe why the current module
 > system is `idiosyncratic' (as the manual says) and needs replacement
 > at all? (serious question)

For me, most annoying part (still remained) is the confusion of the
hierachy of module name and "Recursive Namespaces" of symbol.  The
hierachy of module name is good thing like file system with hierachy
(such as UNIX), it's just useful than flat space.

On the other hand, the idea of "Recursive Namespaces" is bogus, not
implemented, in fact.  It seems that its intention was supporting
namespace in namespace of symbol.  Note that it's *not* the namespace
of module name.  I guess that idea is like that: Symbols defined and
exported by the module (oop goops) are in the scope for the modules
under (oop goops), such as (oop goops stklos) automagically.

I think that it's questionable feature, much.  I think that user
should control the namespace explicitly.  It's dangerous to have
inherited symbol namespace which is implicitly-defined by module name.
At least, we should not combine that feature of recursive namespace
of symbol with the hierachy of module name.
--