[DGD]Initializing problem

[Par Winzell]

 > > On a more serious note, I think it could be a very good idea to
 > > supply minimal implementations of ~System/initd, ~System/sys/telnetd
 > > and ~System/obj/user with the kernel library. This will give new
 > > developers a good place to start hacking at. Any comments on this?
 > 
 > I totally concur. I'll use Erwin's to get started (a million
 > thanks for them!) but if they could be right in the kernel
 > lib that would help a lot.


As good an idea as this may seem, it has been my experience that if
you include any kind of minimal anything, "for example/documentation
purposes", then every single damn library developed will forever be
constrained by your example implementation. By contrast, if you just
document your library (and if there's anything we should be begging
Monsieur Croes for, it's to finish the documentation ;) and put it
out there, people have to go through the truly brain-crunching ordeal
of figuring it out for themselves.

Example mudlibs, however skeletal, create a starting condition vector
that seem to have an unreasonable psychological hold over subsequent
developers. It's a shame to write such a generic kernel library only
to see everybody use it in a completely predictable manner. The point
of writing something so generic and releasing it to the world is that
the world has the power to surprise you with the unexpected. At least
in theory.

That said, perhaps putting Nino's skeletal frame on a permanent site
somewhere and including a link to it in a kernel library README would
be permissable. That might avoid some of the sheep-like behaviour in
mudlib implementors.

Zell


List config page:  http://list.imaginary.com/mailman/listinfo/dgd