[Aaaahhhhhh - lightning strikes] RE: [Q] mixin pattern - ho
w to use an alternate Behavior for Class
Withers, Robert
rwithers at quallaby.com
Mon Jan 28 22:00:46 UTC 2002
> -----Original Message-----
> From: Tim Rowledge [mailto:tim at sumeru.stanford.edu]
> "Withers, Robert" <rwithers at quallaby.com> is claimed by the
> authorities to have written:
>
> > Nice try, tim. It is at least rotationally conservative,
> but you completely
> > left out the relativistic SU(5) group component to ensure
> that the Lorenz
> > transformation remains time conservative. Thus, no Taylor expansion
> > allowed. We gotta have that - how else would the duct tape
> work? I mean it
> > wouldn't be sticky then would it, since you just decided
> that no covalent
> > bonds can form? Ee << -Ew! hell, we wouldn't even be able
> to think about
> > it, with all our receptors being so non-receptive, as a
> result. Thank god
> > for covalent bonds - seeing as how Smalltalk relies on them
> so heavily.
> Actually..... there's always Van der Walls forces. A non-covalent (I
> think; no doubt somebody can correct me if I'm wrong) 'sticky' force
> that is used to exquisite effect by Geckos for example.
Yes, good point. :) Covalent bonding is perhaps not as useful as VdW in
Morphic for example, although the VM does depend on a certain fixed
molecular form, which is reminescent of covalent bonding (object->class->
{superclass, methodDict, format} ). It would be awefully nice to describe
that molecular form, to the VM, at the meta level. For instance, a
delegative model may have a different class pointer/structure/lookup
mechanism.
and so it goes.
Rob
More information about the Squeak-dev
mailing list
|