trust (was Re: Run Only)
trust (was Re: Run Only)
- Subject: trust (was Re: Run Only)
- From: Jeff Ganyard <email@hidden>
- Date: Tue, 19 Dec 2000 18:25:57 -0800
[snip]
Ultimately, it comes down to a question of verification and trust. For
any script you're going to run, you have to be able to answer:
1. Who's the author of this script, and how much do I trust them?
2. Did the claimed author *really* write it?
3. If so, is it still the same thing they sent out?
There are known techniques for answering these questions, but they either
aren't implemented in Mac OS yet, or require script authors to jump
through some very tough hoops. Once you can answer those three, you have
to decide what your policy is for various authors. Do you trust them
completely? Not at all? To do some things, but not others? The first
two are easy -- either you let the script run or you don't -- but the
third involves all sorts of tricky access control issues. It'll be a
while before we can seriously address all this.
With the proliferation of shareware and freeware over the past few
years, it's the same thing. AppleScript doesn't really lower the bar
or invite any more devious behavior that HyperCard, MicroBrew (I
loved that one!), RealBasic, Pascal or C/C++
If you do not trust the author, or the author is unknown don't run
the executable.
This should apply to everyone and all executables.
We are wonderfully excused from much malice on our platform of
choice, and I've never really seen an acceptable definitive reason as
to why other than, "mac people" love their machines and have little
desire to mess up someone else's machine. That's great and I am happy
to be part of such a community, but does that mean it will persist?
Some kind of reference to Genesis and the Tree of Knowledge and how
long can we expect to stay in the Garden of Eden seems apropos here
but you can fill in the details yourselves.
I've never understood the motivation behind creating virii and other
malicious code. I have certainly explored uses of a variety of
technologies that could be used in *bad* ways, but it was an
self-educating process and from it I learned how to build better
solutions for others.
I don't think it's the responsibility of the tool developer (in this
case Apple with regard to AppleScript but it could just as easily be
Metrowerks regarding CodeWarrior) to enforce responsibility on the
part of the tool users. I would rather see Apple educate it's
customers on "safe computing" practices. Or provide OS level support
for trust certificates, I've always thought, and frequently said, the
keychain could be great tool for developers and users to manage
issues of verification, licensing and trust, but that's a whole
different issue.
I'll get off the soapbox now...
jeff
- Follow-Ups:
- Save as
- From: Peter Mathiessen <email@hidden>
| References: | |
| | >Run Only (From: Rob Jorgensen <email@hidden>) |
| | >Re: Run Only (From: Chris Nebel <email@hidden>) |