Baran

Comments inline:

On 28/07/2014 17:39, "baran_H" <[email protected]> wrote:

>Hello,
>
>i have been always gready to find an interesting use for Semantic Web. Now
>i want to present you something where i don't know exactly how original it
>is:
>
>You have an app to tip a very secret message in the dataset of a given
>ontology, 'air force' for example. The app asks you after a 'controlled
>input', to which ontology should i transform it? You choose football for
>example. The app suggests a text about football, you read and accept it or
>you change the text how you like it without changing special colored
>expressions. Then you send your 'football text' to your friend. No
>observation of the world can check that it is an encrypted message, but
>your friend can read the secret message when he/she has the same app.

You've basically just described a RDF based steganography technique
(http://en.wikipedia.org/wiki/Steganography) I.e. the art or practise of
concealing a message within another message.


One problem with steganography is that anyone with the app can read the
message because unlike cryptography there isn't a shared key involved.  So
if your malicious eavesdropper has the app then they too can read the
original message.

>
>For the case that both computers are already infected: Both have a second
>computer inline with a selfcontrolled embedded connection.

I'm not a security expert but that above statement doesn't fill me with
confidence.  If a machine is compromised then it is compromised, even if
the malicious eavesdropper can't read the original message they can likely
get a copy of the app at which point they can decode the message anyway.

>
>This is basically mapping ontologies to each other and the practical
>programming of this staff is finally trivial, would i say, i thought a lot
>about it, also for the case of a second comp in line...

True, though I would say trivial typically means insecure.  There is a
reason information security is hard and even the experts can get it
horribly wrong e.g. OpenSSL and Heartbleed

Rob

>
>Now you can ask, what Jena have to do with this: Speaking from experience:
>This is the best place in the web to put such a posting, would i say. And
>i have a somewhat dubious hint, that such concepts are practiced already.
>Everything is already thought anyway, but i am very curious about comments
>of interested or better informed users...
>
>thanks, baran.




Reply via email to