> Another question: have you EVER written a program for which you did not know
> the running time?
Dozens and dozens. They run until the user tells them to stop.
> Because any person who tries to ship a program (and we're
> assuming here that the software WILL be shipped to a great many users)
> without knowing if there are (undesired) infinite loops in the code is
> simply irresponsible.
I've got lot of programs running right this second which contain infinite
loops: Linux, FreeNet, Enlightenment, xterm, bash, mutt, Mozilla, emacs,
gimp, X11, syslogd, telnet, VMware, Microsoft Word. The only way any of
those programs will terminate is if i interrupt them. The only difference is
that i hit 'x' or Alt-Q instead of Ctrl-C.
qsort is one of these programs too. The only difference is that it doesn't
look for Ctrl-C, it looks for Ctrl-D.
> code. At least, not *rigorous* proofs. But I CAN give you upper bounds,
What's the upper bound for my xclock process?
> As for the automatic downloading of software; god help us when our computers
> start downloading code without our permission.
But that's the context of our entire discussion -- it's in regard to JF's
proposal. With regard to FreeNet, one might have said, "God help us when our
computers start downloading files without our permission."
> But if we want them to do that, then we sure better put restrictions on
> what can run; and I, for one, would be very happy to ban web servers and
> other infinite loops from my machine
What i'm trying to tell you is that it is impossible to ban them without
banning every single program that i mention above. You'd be limiting
yourself to 1950s-style "text stream in, text stream out" non-interactive
batch processing jobs.
--
Mike Schiraldi
Verisign Applied Research
_______________________________________________
freenet-tech mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/tech
