[self-interest] Re: The Effort on Self
gordon at cichon.de
Fri Nov 19 11:36:54 UTC 1999
> And we've carefully do distinguish the
> * language - that Smalltalk-style, round and square bracket thing
> * semantics - prototype based, slot accessing, dynamic inheritance
> * environment - kansas gui, outliners, morphic
> because I'm sure different people understand different things when they
> talk about Self.
In fact, Sunlabs Self is a big thing. (Again, I think what makes it
different from Mozilla is the license). And there is a large portion
of code written in C++ which is less nice than Self but that was the
way to get that compiled quickly and to show that's possible to
generate efficient assembler code from Self source at all.
And there is a lot of things in it that go much beyond a simple
Self language implementation. And some of them are of course
tight to a special machine architecture or operating system.
As it is a fundamental Smalltalk philosophy, and Jecel pointed it
out recently, too, a good OO system should be implemented in itSelf.
(what a nice word game again... :-)
Sunlabs Self already has a GUI, and it has a parser generator. I
think, these things are very helpful for evolving Self, and they
are very difficult to develop standalone, and I assume impossible
to develop in Self if you don't have Self running already.
So, it may be more difficult to port Sunlab Self than to reimplement
your onw Self from scratch, but at least when it comes down to a
programming environment, you are going to have a hard time writing
Self code with vi.
My proposal for whatever way to diversify Self afterwards, we
could use Sunlab Self to bootstrap it.
More information about the Self-interest