At Mon, 01 May 2006 14:34:26 -0400, "Jonathan S. Shapiro" <[EMAIL PROTECTED]> wrote: > I suspect that the discussion of use cases (coming next) will be > revealing to both of us. Let us defer the discussion of moral rights in > digital information until then. What I can say on this subject now is > this: > > It appears to me that there are cases where disclosure is > morally and ethically mandatory. I do not (personally) believe > that the encoding of information in digital form should make > disclosure mandatory (not even to legitimately interested > parties) but there are *other* cases where it is. > > Simultaneously, there are cases where it is morally and ethically > required NOT to disclose, and for this non-disclosure to be > computationally enforced. > > It is obviously not possible to satisfy both of these objectives > simultaneously using mechanical means. > > My belief is that the first situation is extremely rare, and in those > cases that I recognize it is something that would legitimately justify > the cost and difficulty of forensic analysis. I do see that the "digital > information should be free" argument may lead to a much larger space of > cases. > > My belief is that the second situation happens legitimately in > computational systems all the time, and that support for encapsulation > is therefore pragmatically necessary. I will attempt to show examples in > my next message.
I feel we have not yet reached a point where we can clearly pin-point our differences in a way that we both agree on, but we are probably getting closer. I do agree that there are cases where disclosure is mandatory, and that these cases are probably independent of the medium or form. I also agree that there are cases where privacy is required, and that sometimes this non-disclosure needs to be computationally enforced. So, this categorization alone can not tell us where we differ. Because you staked out what you believe, let me follow in turn so we have some background and perspective on where we are standing. I do believe that the need for disclosure is ubiquitous, and not rare at all. So there is a difference between you and me right there. Examples for me are fair use, public domain, parody, in fact any cultural progress, whistleblowing, subpoenas of evidence, and emergencies where the authorized party is incapable of responding (this example applies to medical records in particular). These are the obvious ones, but I feel that there are many more. I do not think that disk forensics are a sufficient mechanism to address these problems. This issue will become increasingly important, should more and more social mechanisms be replaced by technical mechanisms, which, if successful, will be excessively overreaching in their protection (this is my prediction). Now, at this point, one may say that we should then work on improving the mechanisms to allow the good uses of disclosure in a safe manner. However, it seems to me that this is to some extent impossible, and to another extent strategical suicide. On the other hand, at the privacy side, I think that I do not need encapsulation (in the sense of the constructor mechanism) to support the cases where I believe computationally-enforced privacy is important. Beside the obvious but very important case where the machine owner and user are identical, there are cryptographic protocols that are designed to overcome the boundary of the local machine. We currently do not have any agenda to discuss the need for disclosure. However, it will likely come up when the use cases for encapsulation are discussed, because one way to show that encapsulation is wrong is to show that in the particular scenario unathorized disclosure can be important. We do have an agenda to discuss encapsulation and identification. So, let's progress with that! Thank you, Marcus _______________________________________________ L4-hurd mailing list [email protected] http://lists.gnu.org/mailman/listinfo/l4-hurd
