Sujet : Re: Lispification of Forth
De : minforth (at) *nospam* gmx.net (minforth)
Groupes : comp.lang.forthDate : 28. May 2025, 10:03:11
Autres entêtes
Organisation : novaBBS
Message-ID : <3182ec14697011f4dda6a722e2f95020@www.novabbs.com>
References : 1 2 3
User-Agent : Rocksolid Light
Of course you know better. In my fantasy, xt's have no
relation to physical addresses, only to table slots.
Therefore words such as ' or , or deferred words can work
with - let's call them - hashed-xts in the same way as
with 'normal' xts.
The inner interpreter, of course, has to resolve the
hashed-xts in order to retrieve real memory addresses.
This is a time penalty, but we are not talking about
speed here, but about great flexibility.
IIRC in HolonForth it was also possible to access and
modify code and even sources on the fly, and were
immediately active after commitment. I don't know how
Wolf Weijgaard implementd that, perhaps he was once
inspired by Smalltalk.
--