Re: Re[2]: import and LuaBinaryModules

[Thatcher Ulrich <tu@...>]

On Mar 06, 2003 at 12:14 +0800, Peter Hill wrote:
> Thatcher Ulrich:
> > In a nutshell, the objection I have to these gymnastics is that they
> > don't solve the underlying problem, which is in order to avoid name
> > clashes, modules need globally unique names.  When you say:
> >
> > import "a"
> >
> > The "a" needs to uniquely specify the desired module.  If someone else
> > in the world implements a different module named "a", you still have a
> > name clash, regardless of whether you can put the results of import()
> > in an arbitrary table.  In my opinion, it is much simpler to just
> > adopt the convention that:
> >
> > import "a.b.c"
> >
> > puts its public interface in the global table a.b.c, and then try to
> > impose some organization on module names.  If the user wants to alias
> > it, they do:
> >
> > local_alias = a.b.c
> 
> Ah, but as far as:
>     import "a.b.c"
> goes the name clashs problem only occurs with *other modules*, which
> can be handled by whatever module naming and/or versioning
> convention may be decided upon.

Yes -- exactly!  The problem must be solved by packagers and module
maintainers.  Once the problem is solved there, it is solved for
everybody; there is no problem in Lua's global table, if the module
names are unambiguous.  This is a big piece of the value that a
distribution like luacheia should provide.

Java and Perl, successful languages with enormous class/module
libraries, handle it this way.  I don't see the point of trying to
innovate here.

-- 
Thatcher Ulrich
http://tulrich.com