Eckhard Lehmann wrote:
>  Leave Tcl as it is, without OO in the core. There is no value in
>  adding something to the core that exists in many extensions for ages
>  now. People have found their way over the years with one of the
>  extension.

Ah, the "Oh no, don't do anything" viewpoint. The year 2000 called; it
wants that argument back. :-)

>  Instead add lexical scoping and closures to Tcl. These are features
>  that can not be provided by extensions - But once they are available,
>  you have OO implicitely and can create your own OO in a few hours.
>  This would be more powerful and "future-defining" for the language.

It's also blisteringly difficult. Lexical scoping is really tricky and
also makes writing debuggers a lot more difficult. Closures are even
worse as they require a completely different execution engine that will
significantly impact upon extensions. (The problem is that you have to
do cunning games with the stack.)

Now, it is actually the case that Miguel is looking into doing changes
that will enable this sort of thing somewhere down the road. But I do
not think that closures will make 8.6! Limited lexical scoping might - I
can see ways to make it work for some things - but to be honest it
probably won't either.

And just because there's closures and lexical scoping doesn't mean that
the OO system built on top will be really elegant. Just look at Perl5
for example! Or maybe not. :-)

>  Consider Scheme: it has NO builtin OO system, but is doing quite well
>  in terms of popularity.

It is? I must have blinked and missed it.

Donal.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Tcl-Core mailing list
Tcl-Core@[...].net
https://lists.sourceforge.net/lists/listinfo/tcl-core