[Rd] Re: [Omega-devel] StatDataML Description element

[Paul Gilbert]

Torsten

>The idea was not encryption but having something that we can be sure
that
>the dataset has not been manipulated.

This is called authentication. You can add a checksum to the data
format, or something like that, but that just helps stop accidental
manipulation. I think it is useful and may be the way to go, but don't
think of it as any kind of certification or authentication. If you
really want authentication, then the whole format (data, description and
all)  needs to be encrypted by the source with their private key, and it
can then be unencrypted by anyone with the source's public key. The
encryption does not prevent anyone from reading it, since anyone can get
the source's public key. It just prevents manipulation. To do this
properly you then need to be sure that you can get the true source's
public key (not some bogus key provided by whoever is providing you with
manipulated data). That is where Public Key Infrastructure (PKI) is
necessary.

Needless to say, this gets a bit messy, but there are some important
points. The key cannot be part of the message (data format), otherwise
you cannot unlock the key. That is, the authentication has to be part of
what I loosely call the "communications protocol."  I believe these
services are provided by CORBA. If you really want authentication then
you should be looking at the communications protocol, and that is
separate from the data format.

Paul Gilbert

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._