"ts" <decoux / moulon.inra.fr> wrote in message
news:200105261501.f4QF1kQ07869 / orsay1.moulon.inra.fr...
> >>>>> "S" == Sean Russell <ser / maus.germane-software.com> writes:
>
> S> On Sat, May 26, 2001 at 08:12:01PM +0900, ts wrote:
>                                              ^^^^^^^^
>  No,
>
> >> ------------------------------------------------------------
> >> I personally don't prefer meta approach (a language written by
> >> itself).
> >> ------------------------------------------------------------
>
> S> Caveat: I don't think anyone proposing to self-implement Ruby is
expecting
> S> that the resulting code would replace the C code,
>
>  I really hope because see the author of this phrase ([ruby-talk:14893])
>
>
> Guy Decoux
>
>

Hey Guys (small pun intended),

I'm actually trying to challenge the logic and resolve behind a lot of
proposals I have been seeing to write a java version of the ruby
interpreter.  In a way, I think bringing to light the actual challenges of
doing so will lead to one of a few outcomes:

1)  More concentration will be focused on providing java mapping into the
standard ruby implementation.  (What I'd probably prefer to see.)

Or

2)  The challenges and drawbacks (#1 maintaining a second version) of
creating a usable implementation of ruby in java will be apparent and no one
will take it seriously.

Or

3)  The benefits (being able to distribute ruby functionality in corporate
applications via the existing java infrastructure) of a java version are
seen as so valuable that one or more good developers undertakes the project.
If this happens, the complexities should be understood up front.

Or

4)  A few of us hack around with it as an excercise like Sean Russel
mentions in his post.

I'm not particularly advocating any outcome, but I do feel that delving into
the potential complexities and choices is valuable.

Wayne