Donald Fisk <···········@bt-sys.bt.spamblock.co.uk> wrote:
> Marc Wachowitz wrote:
> > [ISLisp] has roughly the same "size" as Scheme, but it's object oriented,
>
> PC Scheme (originally from Texas Instruments) was object-oriented.
> It came with SCOOPS. I don't know if SCOOPS is an extension or
> an integral part of Scheme.
Perhaps SCOOPS has (or could have) been implemented on top of Scheme, or
it was just an extension, but standard Scheme isn't what nowadays tends
to be considered object-oriented. There are indeed several implementations
of Scheme with different extensions for object-oriented programming.
Of course, depending what exactly you expect as object-orientation, and
the kind of optimizations you want to be realistic, it's already possible
to build one's own object system(s) in Scheme - depending on whom you'll
ask, this is seen either as a curse or a blessing. I think I remember a
footnote from the second edition of "Structure and Interpretation of
Computer Programs" (by Abelson, Sussman & Sussman) that in their view
there still isn't a clean view how generic operations should look like
in the general case (opposed to plain hierarchies, like Scheme's numeric
tower), and that may well be one reason why standard Scheme yet doesn't
have an object system in the usual sense of the word. Traditionally, the
Scheme gurus are supposed never to take something into RnRS on which they
can't reach a consensus that it's really the one and only pure solution ;-)
(Btw, pondering some issues of macros and modules, I think I'll go back
to Scheme rather than ISLisp as a foundation for my Lisp compiler, since
the primitive symbol-oriented macro approach of ISLisp just isn't good
enough for my purposes, whereas Scheme's hygienic macros can be extended
to work well for my interests [I can hear someone's sigh of relief ...])
-- Marc Wachowitz <··@ipx2.rz.uni-mannheim.de>