Re: How does xotcl compare to ruby?

GN wrote:
I have certain doubts about this (concerning the XOTcl part). How much
of the XOTcl regression test does oo run?

None. How could it? It doesn't define anything called Object or Class.
The real question is what fraction of this basic incompatability could
be dealt with through defining some classes and some interface code.

I am pretty sure that itcl
can be implemented on top of XOTcl (did a proof of concept
implementation a few years ago), and i am pretty sure that snit can be
implemented on top of xotcl (most likely with some gain in execution
speed, if there is some will).

Snit's very likely to go directly on top of my code.

On the other hand, if the itcl community is happy with itcl, why to
force this into a new implementation? The existing Tcl extensions won't
stop working, only because new code is implemented.

True. Mind you, one of my motivations is to clear up the mess of ad hoc
OO systems that are springing up in various pure-script extensions
(starting with the core's own http package!) Tcllib has many different
such things; they're a mess and they won't base on top of any non-core
OO system because of redistribution requirements. Like this, at least
*some* of the swamp can be drained.

Donal.
.