[log in to unmask] quoted and then wrote:

>From:  [log in to unmask] (Beard, Frank)
>Sender:    [log in to unmask] (Team Ada: Ada Advocacy Issues (83 & 95))
>Reply-to:  [log in to unmask] (Beard, Frank)
>To:    [log in to unmask]
>Jim,
>
>Currently the two are not connected.  They are two separate
>networks.  E-mail can't be sent directly to the ship from the
>shore from non-encrypted E-mail systems.  The message
>would have to be copied (currently done by hand with pen
>and paper) from the unsecure E-mail system and
>re-transmitted over the encrypted ship network.
>
>Even if they were to connect over the "commercial" internet,
>they would still have NES encryption (for example) between
>the systems.  So the only way it could be received is from
>another authorized node with NES encryption.
>
>Now if someone were to write the E-mail to a disk and then
>re-enter it onto the encrypted system, then it might be
>possible.  The Navy would probably re-institute keel-hauling
>for the naval operator responsible.
>
>Some of the Navy systems were hit by the "I Love You"
>virus, but they were not the ship networks.  They were
>the land based internet E-mail systems.

Some US military (not necessarily Navy) classified systems
_were_ hit by recent Billybox "viruses" (actually worms).
When I asked I was told that the security policy only
prohibited "outgoing" transmission of data.  I know
that for previous protocol sets it was possible to use
a trusted "one-way gateway" for mail.  Thus the problem
becomes one familiar to those accustomed to designing
Ada-class software systems -- ensuring the specification
matches the real requirements.