Yes, the physical data link, in this case the RS232 (start/stop fashion)
has nothing to do with the "transparency itself", but how your software
would use that interface.  For example, a lot of software consider the
character hexa "0D"h (decimal 13) as carriage return, so it would return
the cursor or the print head to the left position. When you send data
0Dh to a video card memory it would be just there, it would not return
any cursor to the left of your screen, but who does that is some
software that would intercept it, understand as a "carriage return" and
execute all the necessary steps to insure a "carriage return" on your
screen.  It is a particular situation when that specific software would
use that specific byte for that purpose.  When you transfer an
executable or binary file via Internet or when you copy that file to a
diskette, no special action would be done when transferring the
character "0D"h... so you could say it is being transferred in
"transparent" mode.

In other situations, when a data link use some special characters to
control the data link, as for example, XON and XOFF, or any other
character that makes sense to the protocol software to execute some
specific action, and you need to send a program via that same datalink,
and tihs program contains those special control characters, you don't
want the protocol to understand them as real "control" since they are
just part of a program, then this characters or the whole program needs
to be transferred under "transparent mode", so the other side protocol
just treat them as "regular data characters".

Wagner

Gerhard Fiedler wrote:
> transparency is not a characteristic of a data link, it is a characteristic
> of a relationship between a targeted application protocol and a link. one
> link may be transparent for one protocol but not for another, depending on
> the requirements of the protocol.
>
> ge