"Justin T. Gibbs" writes:
> > > is documentation on how a fetch command for NetBSD's package system
> > > has to work somewhere available?
> >
> >What do you mean?  In NetBSD, this is just done w/ ftp's command line
> >interface, i.e.:
> >
> >	ftp ftp://ftp.foo.com/pub/foo/foo.tar.gz
> 
> This won't work for the packages that are only distributed via HTTP
> unless the NetBSD project caches them somewhere that has ftp access.

NetBSD's ftp client supports retrieval of files via http on the
command line. This feature was implemented because:
	a) it is useful functionality, and solves the problem you highlight
	b) the code is small, simple, trivial

In my view (*), adding functionality which results in a small amount
of `bloat' but provides a very useful feature (and negates the need to
have yet another program to do command line URL fetching) is worth it.

(*) I'm biased with respect to ftp, by the way.