Yeah, I had this idea maybe ten years ago. A friend told me someone
(IRIX ?) has already done it. The idea of active objects in the filesystem
is interesting, and could make porting code from other systems easier. Of
course it could make debugging nearly impossible.
>
> That would be evaluated and expanded when the link is read (by the kernel,
> not the shell).
I would rather see kerneld pass a request to a daemon to handle it. I just
bumped into this Coda thing; maybe a good place to try the idea.
--Jim
" "
\ ____/| /
\ \ o.O| /
\ =(_)= /
\ U /
\ | /
\ #/
ACK! PHTHPHTH!
"I am not a crook!"